Heartland of Cities

oi.uchicago.edu

The University of Chicago Pressgratefully acknowledges a subvention from the Oriental Institute

in partial support of the costs of production of this volume.

oi.uchicago.edu

Heartland of CitiesSurveys of Ancient Settlement and Land Use

on the Central Floodplainof the Euphrates

Robert McC. Adams

The University of Chicago PressChicago and London

oi.uchicago.edu

The University of Chicago Press, Chicago 60637The University of Chicago Press, Ltd., London

© 1981 by The University of ChicagoAll rights reserved. Published 1981Printed in the United States of America85 84 83 82 81 5 4 3 2 1

ROBERT McC. ADAMS is Harold H. Swift DistinguishedProfessor of Anthropology in the Oriental Institute and theDepartments of Anthropology and Near Eastern Languagesand Civilizations at the University of Chicago. He is theauthor of several books, including Land behind Baghdad andThe Uruk Countryside, both published by the University ofChicago Press.

Library of Congress Cataloging in Publication Data

Adams, Robert McCormick, 1926-Heartland of cities.

Bibliography: p.Includes index.1. Babylonia-Antiquities. 2. Babylonia-Civiliza-

tion. 3. Cities and towns, Ancient-Iraq-Babylonia.4. Land use-Iraq-Babylonia. 5. Irrigation-Iraq-Babylonia. I. Title.DS70.A32 935 80-13995ISBN 0-226-00544-5

oi.uchicago.edu

To the Memory ofFuad Safar

oi.uchicago.edu

oi.uchicago.edu

Contents

List of Figures ixList of Tables xi

Foreword Thorkild Jacobsen xiii

Preface xvii

1 WATER, LAND, AND LIFE 1

River and Alluvium 3Local Climate and Vegetation 11An Overview of Major Watercourse Succession

Patterns 14Effects of Human Agencies 19Features of a Premodern Euphrates Landscape 22

2 THE RECOVERY OF ANCIENT SETTLEMENT AND IRRIGATION PATTERNS 27

Maps, Air Photographs, and Their Interpretation 28Boundaries of Survey Coverage 33Reconnaissance Sequence, Procedures, and

Effectiveness 37Site Survey, Collections, and Dating 43Reinterpretation of Lower Diyala Canal Patterns 46Major Limitations of This Survey as a Source for

Settlement History 47

3 URBAN ORIGINS (UBAID-EARLY DYNASTIC I PERIODS) 52

Onset of Alluvial Settlement 54Mass Appearance of Sedentary Cultivators Around

Regionally Differentiated Hierarchies of UrbanCenters 60

The First Urban Climax 81Appendix A: The Survey Data Base: Uruk-Early

Dynastic I Sites, Collections, and ChronologicalIndicators 94

Appendix B: A Growth Simulation ProgramRobert G. Hassert 127

Vii

oi.uchicago.edu

Contents

4 INTEGRATION AND FRAGMENTATION UNDER SUCCESSIVE, CONTENDING DYNASTIES(EARLY DYNASTIC II-MIDDLE BABYLONIAN PERIODS) 130

Urban Hierarchies and Continuity 136Upper and Lower Limits of Settlement, and

Associated Patterns of Land Use 141The Shifting Network of Watercourses and

Settlements 155Appendix: The Survey Data Base: Late Early

Dynastic-Middle Babylonian Sites andChronological Indicators 170

5 CULMINATION AND COLLAPSE OF AN AGRARIAN BASE AND URBAN SUPERSTRUCTURE(NEO-BABYLONIAN-LATE ISLAMIC PERIODS) 175

An Overview of Local Demographic and SettlementTrends 177

Changing Agricultural and Fiscal Management 185Appendix: The Survey Data Base: Neo-Babylonian-

Late Islamic Sites and Chronological Indicators 228

6 CONCLUSION 242

Ancient Mesopotamia as an Irrigation Society 243Ancient Mesopotamia as an Urban Society 248

7 GENERAL SITE CATALOG 253

APPENDIX The Southern Margins of Sumer: ArchaeologicalHenry T. Wright Survey of the Area of Eridu and Ur 295

Notes 347

References 353

Index 361

viiirill

oi.uchicago.edu

Figures

1. Variations in Euphrates and Tigris RiverDischarge

2. Periodicity in High and Low Precipitation atThree Stations in South Central Iraq

3. Contemporary Tigris and Euphrates MeandersCompared with Ancient Meander Traces

4. Settlement and Cultivation in the Fara Region,ca. 1900

5. An Ottoman View of the Plain below Baghdad,A.D. 1848

6. Major Ancient Levees Identifiable inLANDSAT Imagery

7. Sample One-Kilometer Squares for Restudy ofSurvey Coverage

8. Archaeological Sites and Traces of AncientWatercourses on the Lower Diyala Plain

9. Pre- and Protohistoric Sites on theMesopotamian Plain

10. Pre- and Protohistoric Sites on the Lower DiyalaPlain

11. Fourth Millennium Channel Succession Northand East of Nippur

12. Early-Middle Uruk Period Settlement Patterns

13. Late Uruk Period Settlement Patterns

14. Schematic Representation of Arteries of Supplyand Settlement Distributions in the Nippur-Adab Region during the Early-Middle UrukPeriod

15. Distribution of Uruk Site Areas by Region and4 Subperiod

16. Classification of Settlement Size by Region in13 the Uruk Period

17. Settlement Rank Size by Region in the Uruk17 Period

18. Jemdet Nasr Period Settlement Patterns24 19. Distribution of Early Dynastic I Site Areas by

Region25

20. Early Dynastic I Period Settlement Rank Size

21. Early Dynastic I Period Settlement Patterns34

22. Simulation of Early-Middle Uruk PeriodS Cultivated Areas

4023. Simulation of Late Uruk Period Cultivated

4 Areas4824. Simulation of Early Dynastic I Period

Cultivated Areas56

25. Settlement Numbers and Hierarchy of Size inEarlier Historic Periods

5726. Percentage of Wheat-Barley Harvest Completed

at Successive Time Intervals in Five Iraqi62 Provinces

64 27. Sites and Watercourses of the Third Millennium

65 B.C. on the Mesopotamian Plain

28. Sites and Watercourses of the Late Third,Second, and Early First Millennia B.C. onMesopotamian Plain

68 29. Late Early Dynastic Period Settlement Patterns

ix

71

73

74

83

84

85

89

91

92

93

139

146

156

157

161

oi.uchicago.edu

30. Akkadian Period Settlement Patterns

31. Ur III-Isin-Larsa Period Settlement Patterns

32. The Third Millennium Euphrates at AncientAdab, Also Showing Modern (ca. 1962) DuneFormations

33. Old Babylonian Period Settlement Patterns

34. Cassite Period Settlement Patterns

35. Middle Babylonian Period Settlement Patterns

36. Simulation of Sasanian Cultivated Area withinthe Intensively Surveyed Region

37. Size Distribution of Sasanian Sites

38. Rank-Size Distribution of Sasanian Settlements

39. Neo-Babylonian-Achaemenian PeriodSettlement Patterns

40. Late First Millennium B.c.-Early FirstMillennium A.D. Sites on the MesopotamianPlain

41. Neo-Babylonian-Achaemenian Watercoursesand Settlement Patterns on the Lower DiyalaPlain

42. Seleucid-Parthian Watercourses and SettlementPatterns on the Lower Diyala Plain

43. Seleucid-Parthian Period Settlement Patterns inthe Intensively Surveyed Region

Figures

162 44. Sasanian Sites and Watercourses on the

163 Mesopotamian Plain

45. Sasanian Period Settlement Patterns in theIntensively Surveyed Region

164 46. Sasanian Watercourses and Settlement Patterns

166 on the Lower Diyala Plain16647. Islamic Sites and Watercourses on the

167Mesopotamiian Plain

169 48. Early Islamic Period Settlement Patterns in theIntensively Surveyed Region

181 49. Middle Islamic Period Settlement Patterns in182 the Intensively Surveyed Region

183 50. Late Islamic Period Settlement Patterns in theIntensively Surveyed Region

189 51. Early Islamic Watercourses and SettlementPatterns on the Lower Diyala Plain

52. Samarran-Late 'Abbasid Watercourses and191 Settlement Patterns on the Lower Diyala Plain

53. Ilkhanid Watercourses and Settlement Patternson the Lower Diyala Plain

193

195

198

Inside Rear Cover: Base Map of Intensively SurveyedRegion on the Central Euphrates Floodplain

x

207

208

212

219

220

221

222

224

226

227

oi.uchicago.edu

Tables

1. Climatic Data from Three Stations in SouthCentral Iraq

2. Results of Resurvey of Sample Square Kilometers

3. Gross Regional Trends in Uruk PeriodSettlement

4. Urban and Nonurban Settlement by Region inSuccessive Uruk Subperiods

5. Possible Indicators of Status and Specializationin Uruk Period Surface Collections

6. Changing Characteristics of Late Prehistoricand Protohistoric Settlement Enclaves

7. Periods of Occupation, Estimated Areas, andDiagnostic Surface Materials for Late Prehis-toric and Protohistoric Sites 95-

8. Presence-Absence Tabulation of Probable Urukand Jemdet Nasr Traits in Nonintensive Collec-tions at Selected Sites

9. Types and Frequencies in Intensive SurfaceCollections at Selected Uruk-Jemdet NasrPeriod Sites

12. Declining Proportion of Urban Settlement in12 the Third and Second Millennia B.c.

41 13. Totals of Assumed Site Areas in Hectares, bySize Category and Historic Period

69 14. Periods of Occupation and Size Categories ofSites from the Third, Second, and Early First

S Millennia B.C. 177515. Numbers of Sites by Size Category and Period

79 of Occupation, ca. 700 B.C.-A.D. 200

16. Comparative Data on Lower Diyala Settlement90 Distributions, Middle Babylonian-Parthian

Periods

-116

17. Numbers and Occupied Areas of Sasanian Sitesby Size Category

18. Numbers and Periods of Occupation of IslamicSites by Size Category

117 19. A Euphrates-Diyala Comparison of Trends inSettlement

118

10. Provisional Grouping of Intensive Collectionsat Uruk Period Sites by Chronological Subperiod 123

11. Rank Ordering of Uruk Period Sites in Termsof Typological-Chronological Criteria Listed inTable 10. 125

20. Neo-Babylonian-Parthian Sites: Size Categoriesand Periods of Occupation

21. Estimated Occupational Area in Hectares ofSasanian Sites

22. Islamic Sites: Size Categories and Periods ofOccupation

138

142

22-73

178

179

179

184

194

230

233

235

xi

oi.uchicago.edu

oi.uchicago.edu

Foreword

This century has seen great discoveries in the field of an-,cient Mesopotamian archaeology. The royal tombs ofUr with their gold treasure come immediately to mind,and many other spectacular finds run a close second.With time, though, the less sensational surveys by RobertMcC. Adams that form the basis of this book will beseen, it seems safe to predict, to equal or perhaps evensurpass them in fundamental importance; for they giveus for the first time the original geographic setting withinwhich the ancient Mesopotamian history evolved, tracethe ancient watercourses on which all communicationmoved and along which all settlement ranged itself, andshow that their courses were totally different from thoseof present-day rivers and canals. They likewise throwlight on the rise of the earliest cities in human history;and their data, bearing on the shifting density of popula-tion through the millennia, raise questions and suggestanswers about the basic factors that shaped the country'sfortunes-the determinants for major trends of its history.

The achievement the surveys represent has not been at-tained easily; the demanding treks through uncharteddeserts in burning sun, in rain, or in dust storms de-manded devotion, physical stamina, and persistence inno common degree. To this add the pressure of otherdemanding duties and the wearisome, constant difficul-ties of obtaining permission to work owing to shiftingpolitical orientations and military considerations. ThatAdams did succeed in surveying almost all of the alluviumis a tribute to sheer single-mindedness and unflaggingperserverance. It is also a stroke of great luck for ancientstudies, for the expanding of cultivation into formerdesert areas-commendable as that is in itself-is rapidlymaking archaeological reconnaissance of this type im-possible.

The fieldwork on which this study is most immediatelybased was undertaken by Adams between November

1968 and December 1975. It continues and completes hisearlier survey work, such as the survey of ancient Akkadundertaken with Vaughn Crawford in 1956/57 and hissurvey of the Diyala region undertaken as part of theDiyala Basin Archaeological Project in 1957/58. Drawingon this earlier work and supplemented by a survey ofthe region around Ur by Henry T. Wright, this volumethus covers almost all of the area of ancient Babyloniaor, in earlier terms, of Sumer and Akkad. Only the Lagasharea in the southeast now waits for coverage with thetechniques Adams has developed.

When Adams began work in 1956, the method ofceramic surface survey was still very new. Two pioneeringsurveys had been undertaken, which had convincinglydemonstrated the potential of the survey method as ameans of recovering the river and canal network of suc-cessive periods, and therewith their settlement patterns.But, partial and incomplete as they were, they urgentlycalled for large-scale, systematic coverage of the countryas a whole. The basic technique then used was collectingand dating the surface sherds on the ancient sites of aregion, then plotting the dated sites on period maps. In-terpretation proceeded from the premise that in a semiaridcountry like ancient Mesopotamia settlement would havebeen possible only where water was available-alongrivers and canals. Where the settlements of a periodshowed linear patterns, it could be assumed that the linesreflected the watercourses upon which the settlements de-pended. Further information offered by the data includedthe delineation of settled areas in a region as contrastedwith swamps or desert wilderness, and evidence of occa-sional wholesale abandonment of once inhabited tracts.Sizes of the sites visited were recorded, but mainly with aview to identifying the larger ones with cities knownfrom textual evidence.

Adams developed and refined these techniques sig-

xiii

oi.uchicago.edu

Foreword

nificantly by using aerial photographs, which often showin great detail the actual shifting courses of ancient canalsthat could be dated, but only approximately located, bydata from the ground survey. He also tightened the cov-erage in the search for ancient sites to a narrow grid sothat he might not miss even the faintest traces of oc-cupancy, "pot drops," in the area under investigation. Ininterpretation, he shifted emphasis to data on the size ofthe sites studied. He is fascinated by the contrast betweencity and country and so has become keenly aware thatsurvey data may well miss out on very important factorsin the population picture: the nomadic herders and theirsemisettled brothers, not necessarily groups to be con-sidered aliens and enemies of the settled population. Afurther aspect of the city-county contrast has his atten-tion: movement into the city of people from surroundingvillages, either because of the promise of economic better-ment or because the village, in times of war or unrestrictedbanditry, was too vulnerable, so that its inhabitants hadto seek safety behind the walls of a fortified major city.Conversely, in times of relative peace and internal order,the trend would reverse itself and people would movefrom the city into the country to be near the fields onwhich the city relied for its living.

In developing his own method, Adams has kept abreastof recent methodological thinking. And in exploring thepossibilities of deriving the maximum of significant in-formation from the data provided by an archaeologicalsurvey he shows a refreshing and most timely caution.Over and over he weighs proposed new methods of in-terpretation to see what the evidence will actually sustainand warns against assuming for it a degree of precisionthat, by the nature of the case, it cannot have. Uncertaintyabout how completely smaller sites have been noted andrecorded, even in the most careful survey, makes himhesitant to apply analytical approaches such as "centralplace theory" and Thyssen's polygon technique. He alsojudiciously rejects recent redefinitions of the term "state"that jettison its sociolegal core-"monopoly of violence"-in order willy-nilly to make it recognizable in termsof potsherds gathered. Thus the book offers an instructivecritique of method by a veteran in the field.

Most basically, perhaps, village and city-their originsand ends-present themselves to Adams as primary ques-tions. And so the basic theme of his study is an extensive,detailed, and incisive inquiry into "the forces responsiblefor precocious early growth and those that later con-tributed to catastrophic decline and outright abandon-ment" of these forms.

Overall, Adams's data show from the beginning ofsettlement down to medieval times a curve with threedistinct peaks of population density separated by lows:the Late Uruk period, the Ur III period, and the Neo-Babylonian and Parthian periods. As reason for thesefluctuations Adams proposes the inherent vulnerability

of irrigation agriculture, with its dependence-the moreso the more extensive it is-on a stable centralized ad-ministration. Stability, he suggests, demands a morevaried economic basis, for example, one that wouldallow an alternative when for one reason or other agri-culture faltered. This, of course, implies that the countrycould then support a smaller population than it couldwith exclusive reliance on intensive irrigation agriculture.In developing this thesis Adams takes the reader on afascinating journey through time that, without slightingthe particulars of history, yet keeps the focus steady onthe general forces behind history, shaping it. He speaksas an anthropologist.

And yet there is no lack of insights and problems toengage and challenge the philologist and the historian.Adams's argument that the earliest cities of the Ubaidand Uruk periods, such as Eridu and Uruk, with theirastoundingly elaborate and monumental architecture, arebest understood economically as "central places"-thatis, as centers for pilgrimage to religious festivals and forexchange of goods, and so drawing support widely fromboth settled and nomadic populations-fits remarkablywell with the apparent meaning of many of the oldest citynames. They suggest terms for tribal storehouses ofnomadic or seminomadic groups in which the tribe's valu-ables, especially its religious emblems, were kept. Again,the "heroic age" in Sumer-Early Dynastic-was, likeheroic ages anywhere, a period of unchecked raiding andfeuding. Thus its character is strikingly confirmed byAdams's findings that outlying villages were abandonedwholesale at that time, with people seeking safety behindcity walls, so that the large cities grew larger. For theAkkade period there are some curious findings. Thevillages become repopulated, most likely because of in-ternal controls by means of a chain of police posts meantto guard the safety of the major trade routes. Yet in thesouth many larger cities dwindle or are abandoned out-right. Here, perhaps, historical causes are responsible.Umma, for instance, dwindled in area from more than400 hectares to somewhere between 200 and 40 hectares,while the substantial city of Umm-el-Aqarib to the southof it-possibly ancient Ki-dingir-was totally abandoned.Here one remembers that Rimush of Akkade boasts thatin battle with Umma and Ki-dingir he laid low 8,900men, took 3,000 men captive, and led out 3,000 men tobe massacred. A loss such as this-some 12,500 workers-may go some way toward explaining the cities' decline,for it clearly would have seriously affected their abilityto keep up the extensive irrigation works on which theydepended, as well as to carry on general agriculturalwork. An so one could go on. The political and economicdominance of Isin and of Larsa in the period called afterthem is strikingly reflected in the contemporary extensivecanal works around them, and for period after periodthe survey data enrich or change our traditional picture.

xiv

oi.uchicago.edu

Foreword

Surprisingly, at the time of the Islamic conquest the surveydata show extensive destruction of populous settlementsthat is quite unexpected since it was passed over in silenceby the written historical tradition.

To sum up: A great and lasting achievement with awealth of new insights awaits the reader.

Thorkild Jacobsen

XV

oi.uchicago.edu

oi.uchicago.edu

Preface

Much of the central floodplain of the ancient Euphratesnow lies beyond the frontiers of cultivation, a region ofempty desolation. Tangled dunes, long disused canallevees, and the rubble-strewn mounds of former settle-ments contribute only low, featureless relief. Vegetationis sparse, and in many areas it is almost wholly absent.Rough, wind-eroded land surfaces and periodicallyflooded depressions form an irregular patchwork in alldirections, discouraging any but the most committedtraveler. To suggest the immediate impact of human lifethere is only a rare tent, its mirage floating just over thehorizon; the occasional ruined scarps of mud-walledtribal watchtowers dating back a century or more; some-times a small knot of women collecting dead scrub forfirewood; and at long intervals a distant file of camelsor a scattering of sheep and goats with their youngherdsman. The bustling commercial towns of modernIraq lie out of reckoning, hugging the modern rivercourses and their major effluent canals. Towns today, asalways, are concealed as one approaches them on theground by dense surrounding belts of palm groves. Butsometimes, from a high dune on still, early mornings, onecan detect them even from the remote desert as faint, spikyclusterings of electric transmission pylons, brick factorychimneys, and water towers. Just so wayfarers once musthave taken their bearings on the turreted walls and zig-gurats of much more ancient urban centers.

Students whose interests are confined to the modernMiddle East, and in fact most modern Iraqis, have littlereason to venture into this parched wilderness. Eventhe tribally organized, seminomadic groups that foughtover it until the time of the First World War, both witheach other and with tax-collecting, punitive expeditionssent out by the Ottoman authorities, have mostly settledin better-watered adjacent areas. Yet at one time herelay the core, the heartland, of the oldest urban, literate

civilization in the world. Both sides of this yawning con-trast equally demand explanation: the forces responsiblefor precocious early growth, and those that later con-tributed to catastrophic decline and outright abandon-ment.

Nothing quite so ambitious, or foolhardy, as a com-prehensive explanation can yet be offered for either therising or the falling portion of the curve. To a degree notwidely enough understood, the study of the immenselylong and rich past of Mesopotamia is still in a relativelyearly state of development. Campaigns of excavation, al-though some of them are justly famous, have been almostentirely limited to a handful of the major cities of anti-quity. In nearly all of those, only a minute portion of theassociated debris has yet been sampled, even though manyyears and work crews frequently numbering in the hun-dreds have been devoted to the effort. Many thousandsof other sites are now known from archaeological recon-naissance like that which furnishes the primary empiricalbasis for this study. Most are much smaller, and it isespecially the fully rural as well as moderate-sized com-munities that are grossly underrepresented in what weknow at present. In other words, both archaeological andarchival sources for the most part provide little more thannarrow beams of light with which somehow we seek toilluminate an immense dark room. To speak of generalexplanation in anything other than a very loose, informalsense when even the most prominent, enduring contourswithin that room remain so indistinct and subject to dis-pute would be to misapply the basic precepts of the searchfor historical causality.

This study is concerned with certain major features ofthe infrastructure of Mesopotamian civilization, princi-pally its patterns of agricultural land use and the hier-archical array of communities in which people lived. Thelarger cycles of growth and decline that were mentioned

xvii

oi.uchicago.edu

Preface

earlier are perhaps best mirrored in these features,whether or not the same features supplied the principalenergizing forces that produced the cycles. Land use andsettlement have always involved the livelihood and geo-graphic clustering or dispersal of virtually the wholepopulation. By plotting changes in these variables throughtime, we can at least hope to arrive at a few aggregativeindexes of economic well-being-far cruder, but also witha far deeper time perspective than those toward whichlarge-scale statistical compilations of social and economicindicators are now directed. My intent is to trace whatis known of the changing character of this infrastructureover a span approaching six millennia from a wide, neces-sarily somewhat eclectic range of published textual andarchaeological sources as well as from the findings ofarchaeological surface reconnaissance.

Fifteen years ago, while working in his characteristi-cally incisive fashion over the manuscript of an earlierstudy along similar lines, Benno Landsberger was im-patient for work to be directed at the region that finallyis reported on here. The lower plains of the Diyala weremarginal and hence unrepresentative, he maintained, andthe discussion of settlement patterns on them had thedefect of seeking "to define a dialect before the paradigmof the heartland is known." He was probably right aboutwhat would have been the optimal order of precedencein fieldwork, though considerations other than archaeo-logical preference frequently were the determining ones.We now know that the northern Tigris-Euphrates allu-vium and its Diyala counterpart later came to overshadowcentral and southern Babylonia, but until late in the firstmillennium B.C. there was nothing in the north that re-motely approached the interwoven continuity and masseddemographic strength of the cities of the south. Whiledoubtless of interest in its own right, the Diyala regionwas, as Landsberger saw, unrepresentative of the primaryprocesses by which urban civilization first came intoexistence.

My reply to Landsberger at the time conceded less thanthis: "What is important... is not the degree of deviationof this or any other region from some undefined 'norm'but the encouragement of the study of general historicaltrends in the differing regional contexts in which theywere manifest" (Adams 1965, p. ix). Perhaps I was lessconfident than-he that his implicit distinction between a"key" area and neighboring marginal or dependent ones(Palerm and Wolf 1957, p. 29) would find such strongsupport in settlement pattern data when an opportunityfinally came to survey his heartland. But perhaps also Iwas hesitant even then about circumscribing creativeprocesses within sharply defined but somewhat arbitraryboundaries. To identify the zone of the greatest, longest-lived cities as the key, after all, was in a way only torestate the preoccupation of most Assyriologists withcities. In meeting the day-to-day demands of field recon-

naissance in the desert and taking the academic stanceof an anthropologist, it was natural that my concernsfocused primarily on agricultural infrastructures in thecountryside.

Since that rather metaphorical exchange, much hashappened to highlight certain distortions or anachronisticelements in both positions. In his case, one consequenceof the further development of ecologically oriented studieshas been to undermine his assumption that patterns ofsettlement and land use could be analogized to a linguisticparadigm. In most respects other than vocabulary, lan-guages apparently can be considered closed systems. Dy-namics of change within the realm of language aloneaccount fairly satisfactorily for the great bulk of observedchanges in linguistic structure and usage. Modes of sub-sistence and settlement, on the other hand, seem moreand more clearly to be open, externally determined sys-tems. They are products of shifting, converging socialand natural circumstances rather than outgrowths ofpossibilities inherent in earlier arrangements displayingan unfolding internal momentum of their own.

Similarly in my own case, recent years have witnessedwhat might be called an explosion in the relative im-portance of regions formerly disregarded (by specialistsin the great river valley civilizations of the Fertile Cres-cent) as largely peripheral. I then accepted the geographi-cal framework of the Mesopotamian alluvium as a "na-ural" boundary within which to describe all the crucialprocesses associated with the growth of a civilization,and to expect ultimately to find explanations for them.Perhaps the most significant progress that has been madeover the past two decades or so has consisted in tracingcomplementary developments in vastly different as well asdistant regions: up the Euphrates into Syria and Anatolia,far out on the Iranian plateau, and even down the Gulftoward partners in maritime trade possibly as remote asthe Indus Valley. What seems increasingly clear, in otherwords, is that my reliance on bounded regions as unitsof analysis differed from Landsberger's only in beingslightly more encompassing. This was then, and increas-ingly becomes, an unnecessarily limited approach. Com-prehensive explanations, when and if we ever reach them,will still involve important factors around which it hasbeen possible and convenient to draw boundaries. Butthey will also involve other factors that can be understoodonly within an indefinitely widening series of interac-tional contexts.

Archaeological reconnaissance, or at least this variantof it, is a very small-scale undertaking. Collaborators inother disciplines could have contributed much to the find-ings reported here, and I have had preliminary discussionsof that possibility with a number of individuals withwhom it would have been a pleasure to work. The deci-sive obstacle, in the end, has been that the schedulingof this type of fieldwork is bound to be somewhat erratic

xviii

oi.uchicago.edu

Preface

since official approval is less easily assured than in thecase of excavations. As is detailed more fully in chapter2, there were lengthy interruptions between its inceptionin November 1968 and its completion in December 1975.But, if the sole responsibility for the accuracy of the fielddata is therefore my own, my indebtedness to colleaguesin many other respects is correspondingly greater.

Thanks are due, first, to the Directorate General ofAntiquities both for its continuing official support andfor informal acts of cooperation and kindness on the partof many of its officials. Dr. Behnam Abu-Soof, head ofthe Inspectorate of Surveys within the directorate, re-peatedly provided valued counsel and assistance. Twomembers of his staff accompanied me for different periodsin 1973 and 1975 as representatives of the directorate,Sabah Jassim al-Shukri and Abdul Qader al-Shaykhli.Riyadh al-Qayssi acted in a similar capacity during theinitial season of reconnaissance in 1968 and Abdul SalaamSim'an in December 1975. Their tactful handling of manyofficial aspects of the undertaking and their unflaggingcooperation in the fieldwork were much appreciated.

It is with a heavy burden of personal loss that I mustalso make a last acknowledgment of the special con-tributions of Fuad Safar, inspector general of excavationsin the directorate. His support for, and seminal con-tributions to, programs of research aimed at understand-ing the historical geography of Mesopotamia probablyhave been decisive in whatever success they have had.Only those who had an opportunity to tap his encyclo-pedic knowledge, or to benefit from the subtlety andpenetration he brought to bear upon the research prob-lems of others, in spite of his own very burdensomeofficial responsibilities, will fully appreciate the magni-tude of the loss that Middle Eastern archaeology suf-fered with his untimely death in January 1978. Thededication of this book to his memory takes cognizanceof the fact that without his contributions it would besubstantially weaker and less complete and that withouthis encouragement it might never have been undertakenat all.

Within Iraq, other assistance to the project came froma number of sources. The British School of Archaeology,later the British Expedition to Iraq, hospitably provideda glace for us to stay while in Baghdad, a useful library,and well-informed, congenial company. Thanks areowing to a succession of directors there-David Oates,Diana Kirkbride Helbaek, and J. Nicholas Postgate. Dr.Jiirgen Schmidt, director of the Warka Expedition andthe Baghdad Division of the German ArchaeologicalInstitute, made his Warka headquarters available as afield base for part of the work in 1975. While there, I wasjoined for several days of reconnaissance by one of hisassociates, Dr. Barbara Finster. Her well-informed viewson Sasanian and Islamic archaeology have influenced thediscussion of these periods in the appendix to chapter 5.

Dr. Douglas Kennedy, of the Centre Nationale deRecherche Scientifique, participated briefly in the 1968reconnaissance that was based at the Oriental Institute'sfield headquarters at Nippur. It is likely that a muchfuller, less elastically dated picture of the Hellenisticperiod would have emerged had we been able to con-tinue joint work in 1969 as we then planned. Finally, thedriver-mechanic for the project during the greater partof its existence was Jabbar Nasr Shoja. His willing andresponsible aid under a variety of difficult conditions,and his exceptional familiarity with a wide region cen-tering on the town of 'Afak, were among the project'smost useful field assets.

Several readers of an earlier draft have contributedmuch of their own specialized knowledge to the finalform of the manuscript. They include two of my OrientalInstitute colleagues, Professors John A. Brinkman andIgnace J. Gelb. It should also be said that all the facultymembers housed on the third floor of the institute con-stitute an unparalleled resource in the field of Assyriologyupon which I have drawn informally, repeatedly, andheavily. Professor Henry T. Wright, to whom I am furtherindebted for the account of his survey of the Ur regionthat forms an important appendix to this study, providedan exceptionally thoughtful and painstaking series ofcomments on the entire manuscript. Dr. Gregory A. John-son not only read and extensively criticized chapter 3,but actively collaborated, in person for a time in July1977 and by correspondence, in the analysis leading toits preparation. Treatment of the earlier periods dealtwith in chapter 5 was revised on the basis of valuablesuggestions made by Professor Joachim Oelsner and Dr.Matthew Stolper, and the numerous penetrating com-ments of Dr. Michael G. Morony (who also supplied someof the references to classical Islamic sources) occasionedseveral important modifications of the discussion of thelater periods. The general conclusions to the study com-posing chapter 6 were written during a period of residencein the German Democratic Republic in November-December 1978, under an exchange agreement betweenthe Akademie der Wissenschaften there and the UnitedStates National Academy of Sciences. To a considerabledegree they were shaped by discussions with members ofthe staff of the Zentralinstitut fiir Alte Geschichte undArchaiologie, and especially with Professor Horst Klengel.It should naturally be understood that none of these in-dividuals is in any way accountable either for specificerrors remaining in the final text or for the general ap-proach it takes.

Another less specific, but no less important, set of in-fluences on the form and content of this study are thosethat contributed to my earlier studies in a similar vein(Adams 1965, Adams and Nissen 1972). Prefatory ac-knowledgments made there apply here once again, there-fore, for in spite of its broadened scope this volume will

xix

oi.uchicago.edu

Preface

be recognized as consequent upon a trajectory that wasgenerally set before the latest phases of field activitybegan.. In this respect I should particularly like to expressmy indebtedness to Professors Thorkild Jacobsen andHans J. Nissen. The first is responsible for having initiatedsystematic archaeological surface reconnaissance inMesopotamia with pioneering insight as to its scholarlypotential, and for having directly stimulated my ownlater work on the theme. To the second is owing an in-definable but substantial share of the many methodologi-cal and interpretive refinements that we hammered outtogether during the course of the Warka survey, im-proving the application of that insight without alteringits fundamental character.

Valuable contributions of a somewhat more specializednature have been made by a number of other scholars.Professor Karl W. Butzer helped correct several referencesto fluvatile and geomorphic processes and in particularaided in the understanding of the ancient, meandering riverchannel north of Nippur that is described in chapter 3.Another colleague, Professor Richard L. Chambers, trans-lated terms appearing on the redrawn Ottoman map infigure 5. Dr. Robert G. Hassert, of the staff of the Uni-versity of Chicago Computation Center, designed thealgorithm used in the land-use simulations shown infigures 21-23 and 35 and has provided a brief descriptionof it in an appendix to chapter 3. Much of the basic view-point taken with regard to the locational analysis ofancient settlements in the same chapter stems from adiscussion of the potentialities and limitations of my fielddata with Professor Brian J. L. Berry. Professor A. ColinRenfrew and Mr. John Dixon kindly provided the an-alyses of obsidian samples from site 1072 that are referredto as part of the description of that site in chapter 7, andDr. Stephen R. Lintner identified a marine shell from thesame site. And while a degree of "antiurban bias" may beinescapable when one undertakes to rectify the prevailingneglect of the countryside by most archaeologists andhistorians, Professor Aage Westenholz has properly in-duced me to soften a few of the more orotund expressionsof it.

The many problems of Arabic transcription are treatedherein with consistent casualness. Common words andnames are rendered throughout in their most popularlyrecognized English form. Vowel length is only occasion-ally indicated for the less common names, especiallygeographical ones. The names given for archaeologicalsites in chapter 7 in most cases seek to render the spokendialect of local informants, but some have also beendrawn eclectically from a variety of English and Arabicmaps of various dates that may or may not offer a morestandardized version. Compounding the unreliability ofany such compilation, therefore, are the different periodsof currency of its components. As discussions to be foundin chapter 5 and under site 1389 in chapter 7 attest, exceptfor the largest, longest-lived sites, geographical nomen-clature often appears to have been made rather transientby the advent of unsettled conditions. The only recourseof those who may wish to visit the overwhelmingly largerproportion of the archaeological sites on which this studyis based accordingly is an assured knowledge of fairlyprecise compass navigation in country very modestly en-dowed with stable landforms or permanently recognizablelandmarks.

It remains, finally, to take grateful note of the variousforms of institutional and financial assistance withoutwhich this project would not have been possible. Sup-port for the 1968 reconnaissance was provided in partby the Baghdad School of the American Schools ofOriental Research, of which I was resident director forthat year. The major 1975 campaign was funded pri-marily by a grant from the National Science Foundation(NSF-SOC-74-12491), which also substantially defrayedthe direct and released-time costs of the preparation ofthis volume. And the Oriental Institute not only has under-written some of the costs of fieldwork and the prepara-tion and publication of the manuscript, but has consist-ently provided the unique scholarly setting in whichindependent, long-continuing field undertakings are ac-cepted as a primary responsibility both of individualsand of the institution itself.

XX

oi.uchicago.edu

1Water, Land,

and Life

The Euphrates River of this study is a source of waterand sediment, change and permanency in a harshly aridsetting whose other natural resources are notably few.It makes fertile a green ribbon of cultivated fields andorchards, interspersed by the desolate ruins or bustlingmodern descendants of the world's most ancient towns.Shifting its course in greater or lesser degree as it riseseach spring, it periodically contradicts its promise ofsecurity and contentment with the hardship of drought-induced low water or the devastation of floods. But inthe long run, overriding all such vicissitudes, it has pro-vided the only possible foundation for an immense col-umn of human achievement that has risen laboriously ina pivotal region over hundreds of generations. It is, inshort, a brown, sinuous, pulsing artery that carries thegift of life.

The Euphrates course is generally regarded as begin-ning at the confluence of the Kara Su and the Marat Suin south central Anatolia, whence it descends steeplythrough an alternating succession of rich valleys and diffi-cult gorges until it reaches the elevated Syrian plain atSamsat. Entrenching itself in a broad, steep-sided valley,the river meanders across this plain for more than 1,400kilometers to the head of the lower Mesopotamian alluvi-um at Hit. In this long reach, the slope of its bed isabruptly reduced, and midway across it the Euphrates re-ceives its last important tributary, the Khabur. At Hit,still more than 700 kilometers from its mouth, the low-water elevation of the Euphrates is barely over 50 metersabove sea level. Thence it enters the Mesopotamian al-luvium, periodically bifurcating and rejoining in a natu-rally anastomosing pattern while also dividing the greaterpart of its flow into the hundreds of dendritic arms ofirrigation canals. At the lower end of the plain it mergeswith the waters of the Haur al-Hammar, a permanent,marshlike lake, then joins the Tigris to form the Shatt

1

al-Arab and reaches the Arabo-Persian Gulf southeast ofBasra.

The Euphrates, as thus sketched, is not one of the greatrivers of the world in length, volume of flow, or size ofwatershed. As a river that is preponderantly a source forirrigation, perhaps the volume of water it carries is themost useful variable with which to illustrate its compara-tive position. In rough orders of magnitude, its averageflow is only one-third that of the Rhine or Nile, or amere tenth that of the Danube or Volga. Not to speakof giants like the Amazon or Congo, it is altogetherdwarfed by relatively large rivers like the Mississippi andand Yangtze, which are more than twenty-five times itssize. It is even overshadowed in average volume of annualflow by its "twin" river, the Tigris, approximately in theproportions of five to three (Ubell 1971, p. 3). Perhaps itsclosest comparison in volume is with the Colorado River,which, like the Euphrates, also has its headwaters inrugged mountain country, traverses a broad semiarid toarid belt, and is sharply reduced by human use before itreaches the sea.

But it cannot be overstressed that there are few if anyother streams, regardless of size, that have played socentral and long-continuing a role in human history. Theformative processes leading to the world's first urbancivilization cannot be understood except as a creativeadaptation to the priceless resource of Euphrates water.Vigorous later traditions in political economy, religion,administration, literature, and art continued to build onthe foundation of an assured food supply that the Eu-phrates made possible. Even the land itself, the alluvialplain of southern Mesopotamia, is in large part composedof silt that the river carried down.

This study is about the greater part of the heartlandin which Sumerian civilization arose during the fourth andearly third millennia B.c., a small but central portion of

oi.uchicago.edu

Water, Land, and Life

the Euphrates floodplain. Hence there is no need for amore extensive description of the Euphrates itself, al-though we must deal with aspects of the behavior of theriver that have influenced patterns of human action. Asis set forward more fully in chapter 2, my concern is notonly focused on a pivotal region but largely confined toit by the limits of an extensive but still definitely boundedarchaeological survey that has furnished the primary data.And I must concede that the boundaries of that survey,however convenient and necessary they were for the de-sign of an intelligible and systematic program of research,have essentially no correspondence to traditional bound-aries of human action or relationship other than in thepresent and very recent past.

Thus this is not in any sense a study of the role of theentire Euphrates in history. Probably the proper geo-graphic unit of analysis for that quite different undertak-ing would be its entire watershed. A watershed is onlyanother topographic unit, of course, and need never havecoincided with a historical and cultural unit of any sig-nificance. But in this case there are many leads to a re-current unity within the Euphrates basin that deserve tobe explored, even if that far-ranging task cannot be ade-quately undertaken here.

There were, to be sure, extended periods when deepsociopolitical divisions lay athwart the Euphrates in Syriaor Anatolia, far above the northwestern extremities of thealluvium at the lower end of its valley. That was almostcontinuously the case during the interminable rivalriesbetween the Parthian, Sasanian, 'Umayyad, and early'Abbasid empires, on the one hand, and the Romans andtheir Byzantine successors on the other. But even duringthese lengthy intervals the frontier was less often a fixeddemarcation of actual movements than a shifting gradient.Caravans as well as shallow-draft riverine commerce fol-lowed the line of the Euphrates when possible, from thecities of the Levant to Charax Spasinou, Ctesiphon, Bagh-dad, and the Arabo-Persian Gulf, as they continued to dointo the nineteenth century. Armies, with their irregulars,supply trains, and camp-followers, and on occasion theirprisoners and the masses of population they forced intoexile, did likewise. Seminomadic tribes alternately bar-tered their support to, and preyed on, both caravans andarmies. Even when a durable military demarcation was theobjective of both sides, therefore, the middle valley of theEuphrates was probably always closer to being a long andtenuous but effective bridge than an impermeable barrier.

There were other times when it is difficult to discernformal barriers of any kind, and when intercommunica-tion was so close as to suggest a degree of cultural unity.The first of these occurred in late prehistoric times, in thelatter part of the fourth millennium, with a site like Ha-buba Kabira on the middle Euphrates (Heinrich et al.1969-73) mirroring many features of Late Uruk periodoccupations at better-known sites like the ancient city of

Uruk in southern Mesopotamia. Similarly during the thirdand early second millennia, archaeological finds from TellBrak on the Khabur River and textual as well as archaeo-logical finds from ancient Mari on the Euphrates in Syriaattest to a close and at least partly dependent relationshipwith southern Mesopotamia. The extraordinary recentfinds at ancient Ibla, on the other hand, demonstrate ahigh degree of linguistic as well as political autonomyeven in a context of far-flung and intense commercialrelations (Gelb 1977). During these and other similar in-tervals, it appears, the valley of the Euphrates was a vital,heavily traveled artery of interregional contact betweenMesopotamia and the world around the Mediterranean.Perhaps the topographic limits of its watershed are nomore meaningful a framework in which to consider theseintervals of rapid, wide-ranging, and yet obviously closeinteraction, therefore, than the geographic boundariesof the alluvium in general, and of this study in par-ticular, are for a different, more narrowly defined set ofproblems.

Southern Mesopotamia was a land of cities. It be-came one precociously, before the end of the fourthmillennium B.C. Urban traditions remained strong andvirtually continuous through vicissitudes of conquest, in-ternal upheaval accompanied by widespread economicbreakdown, and massive linguistic and population re-placement. The symbolic and material content of civiliza-tion obviously changed, but its cultural ambience re-mained tied to cities. How firmly the occupants of thelower Mesopotamian plain ever recognized that alluvialterrain as a special object of attachment is uncertain, buttheir enduring loyalty to familiar associations and lo-calities within it-to cities-is not a matter of doubt.Here we are concerned with the material conditions thatmust have played an important part in originating andsustaining these roots of attachment. And it is impossibleto escape the conviction that irrigation agriculture-orthe comparative security, population density and stability,and social differentiation and complexity that it induced-was at the very heart of these material conditions.

Leo Oppenheim (1950) has drawn an evocative con-trast between the rootless, wrathful storm gods of thelands around the Mesopotamian perimeter, which de-pended on rain agriculture, and the irrigation zone godswhom city folk sustained in temples in their midst. Thepoint is not to argue that the arrow of causality must bedirected unilaterally toward religion as a dependentgrowth or epiphenomenon, however, but to suggest thatat the root of any civilization there probably has to be acongruence between modes of agricultural productionand sociocultural (surely including religious) institutionsmore generally. To sketch that congruence or harmony indetail is the task of many specialists, principally in thelinguistically and topically diverse genres of textual ma-terial. Here we deal primarily with the other side of the

2

oi.uchicago.edu

Water, Land, and Life

equation-with the adaptive base for settled, urban lifethat was provided by the Euphrates landscape.

RIVER AND ALLUVIUM

From a relatively early time, well back into the thirdmillennium B.C., the name most generally given to theEuphrates in lowland Mesopotamia was the "SipparRiver." Disagreements persist over other names that oc-casionally were applied locally or in a primarily literarycontext. Also obscuring the matter of nomenclature aremany ancient texts that speak only of "the river," sinceits name was obvious to the intended readers. As a generalname applying to the whole system of branches, it ap-pears at present that we can do no better than what waspronounced as "Buranunu" or "Purattu" in the laterSumerian and Akkadian vocabularies (Adams and Nissen1972, p. 44).

Sippar, in turn, was the name of an ancient, long-lived,but fairly modest town on the Euphrates. An antediluviandynasty there appears in the Kinglist, but at least in fullyhistoric times it was never a dynastic capital. Accordingly,it is unlikely ever to have been a dominant political oreconomic power outside its immediate sustaining region.Its distinction, accounting for the application of its nameto the whole network of watercourses serving many moreimportant cities, must be at least in part that it lay at theuppermost extremity of the alluvium, closest of all thetraditional towns to the point at which the Euphratesdebouched onto the lower plain from its broadly incisedmiddle valley. Possibly this suggests a long-standing,traditional awareness of the alluvial plain as a distinctivezone, with Sippar the point of its beginning. I shall followthat assumed usage, dealing with the river hereafter onlyas it emerges onto the alluvium itself.

Figure 1 summarizes the major variations in modernEuphrates flow as recorded at Hit, slightly upstream ofancient Sippar but indistinguishable from it in these char-acteristics. Average monthly measurements over a thirty-five-year period are recorded, as well as monthly maximaand minima. Since water is needed for irrigation at no lessthan monthly intervals during the primary winter growingseason, it is these monthly figures rather than the annualtotals that have the most critical effect upon the fortunesof the cultivator. A succession of late fall and early wintermonths of unusually low water, coupled, as wouldnormally happen, with less than average early rainfall,can seriously cut back harvests even if the spring floodsare well above the normal level. Clearly, the Euphrates isa :somewhat capricious and undependable provider ofthe water that is vital for irrigation, at least for thoseinferior in politicomilitary power or not otherwise ad-vantageously situated to satisfy their own needs withoutregard to competing claims.

The Tigris, unlike the Euphrates, has a number of left-

bank tributaries along its entire middle course tha: stemfrom catchment basins on the lower flanks of the Zagros.Rainfall in these basins accounts for a large proportion ofits flow, and disastrous flooding on the plains below canfollow heavy, widespread precipitation at any time duringthe winter rainy season. The "normal" but highly variableseasonal flood, added to by a component of melted snow,comes in April. The Euphrates receives only the modestcontributions of the Balikh and Khabur as left-bank trib-utaries. A higher proportion of its flow consequentlyderives from the more elevated interior of the Anatolianplateau. Hence the Euphrates flood comes with later-melting snows, in early May, as figure 1 shows. Highwater at that time is essentially too late to affect the Mayand June harvest-unless it inundates the mature cropswith a destructive flood. Even in good years, in short, thetiming of the arrival of high water in both the Tigris andthe Euphrates is poorly synchronized with the needs ofcultivators on the alluvium.

Present patterns of Euphrates water utilization ob-viously can be described in greater detail than those forany historic period. While we shall see that in some re-spects they differ greatly from those of the more ancientpast, and that there were also striking divergences amongthe latter, it is reasonable to regard the basic features ofthe river regime as constituting a relatively constant setof conditions to which ancient as well as modern agri-culturalists would have had to adapt in essentially similarways. Hence it may be useful to trace in fuller detailsome of the contemporary constraints that the Euphratesimposes on human life.

The advent of efficient pumps, cheap fuel, and modernexcavating machinery has greatly reduced the real costto the farmer of securing a supply of irrigation water.Forward deliveries throughout the growing season cannotbe assured in view of the vagaries of the Euphrates flow,however, and the total supply remains in active conten-tion among a growing number of potential users. Hencethere is a strong predisposition to apply excessive irriga-tion water whenever conditions permit. Present use hasbeen estimated at 13,300 cubic meters per hectare peryear, equivalent to a uniform depth of 1.33 meters onall cultivated land, and not surprisingly this is grosslyin excess of crop requirements (Ubell 1971, p. 9). Con-sidering winter-grown cereals alone, still the heavily pre-ponderant form of agricultural produce and for all earlierperiods the absolutely decisive one, requirements for con-sumptive use may be more reasonably approximated at0.55 meters. With roughly half again this amount neededto cover evaporation, seepage, waste runoff, and otherlosses, this amounts to a gross diversion requirement ofabout 0.83 meters, or 8,300 cubic meters per hectare (FAOMediterranean Development Project 1959, p. III-2). Undermodern conditions it may be realistic to plan for a re-duction in the large coefficient of loss, but a gross diver-

3

oi.uchicago.edu

Water, Land, and Life

Monthly mann disrhhnrn (m 3 I /ee-

EuphratesRiverat Hit

TigrisRiver atFatha

(after K.Ubell 1971)

Fig. 1. Variations in Euphrates and Tigris River discharge.

sion of at least this general magnitude probably may beassumed as the historical requirement for attaining aver-age yields under traditional conditions of agriculture.'

The modern pattern involves increasing attention toregulators and structures that will control water distribu-tion and utilization and that ultimately will permit cen-tral managerial decisions to place a brake on excessiveuse. Upstream water storage facilities also are being in-troduced as rapidly as their heavy capital and technicalrequirements permit. These will encourage a shift towardincreased cultivation of cash crops during the summerlow-water season and will help to smooth irregularitiesof seasonal flow so as to permit an enlargement of culti-vated hectarage. Because factors of production have notbeen equally and simultaneously available, however, themodernization of Iraqi agriculture heretofore has in-volved only a limited shift away from its traditional sub-sistence orientation and only very moderate increases inthe total arable area. And water withdrawals, thoughthey have risen steeply to the point where periodic short-ages are a critical factor in further agricultural develop-ment, still involve very limited use of agregate seasonal

4

flow. As late as 1959, gross agricultural diversions (forboth Tigris and Euphrates) were estimated at as low as16.6 billion cubic meters per year, less than a fifth ofIraq's potential surface-water resources (FAO Mediter-ranean Development Project 1959, pp. III-1-2; Ubell1971, pp. 3-4).

These modern conditions, connected though they arewith stimuli toward rapid development, have led to arapid intensification of problems of salinity. As much as70 to 85 percent of total land under irrigation is said tosuffer from the effects of salinization (FAO 1959, p. III-12). The physicochemical processes involved, as well astheir relationship to traditional agricultural practices,may be succinctly described as follows:

Whenever the water table rises to 1 /2 metres from thesurface, capillary action is sufficient to carry the salt tothe surface, where the water evaporates and leaves a layerof salt. This applies especially to the sub-soil of manyparts of the Tigris and Euphrates valleys. Soils with 0.2percent soluble salts in a surface of 15 centimetres mayhave more than 1.0 percent in the second metre and so

oi.uchicago.edu

Water, Land, and Life

are threatened by excess salinity; 0.2 percent in a surfaceof 30 centimetres is detrimental to most crops, while per-centages as high as 1.0 will prevent the growth of all ex-cept the most salt-tolerant crop species. If the sub-soilwater contains as much as 0.5 percent salt, and the watertable is within 1 /2 metres of the surface for a few monthseach year, the area is likely to go out of production infive years or less. All areas where the water table ap-proaches that critical level during any part of the yearwill need a drainage system if production is to continue.

The methods used by the Iraqi farmers have oftenbeen criticised; it has been said that they are wastefulof water and not suitable for high yields and productivity.It is true that the wild flooding most common in theirrigation of grain fields does not ensure equitable dis-tribution and efficient use of the water. But as the landwas not levelled, this was the only practicable method.Moreover, the excessive irrigation practised with thissystem in most cases is justifiable in saline soils as ameans of pushing the salt out of the surface and the rootzone, so as to establish a good stand and carry the cropthrough to maturity. In the absence of drains, the fieldwas left in worse condition at the end. The dangers ofsalination were enhanced by the rise of the water table.But the farmer had learned to move to new land andnot to come back until natural forces and weeds weregiven the time to lower the ground water table and driedthe soil to a point where a crop-usually poorer than thepreceding one-could be grown again. This has beenthe adaptation of the Iraqi farmers to adversities in theirenvironment with which they could not cope otherwise.They have thus succeeded in surviving this environment. . having learned "to live with the salt in the land."[FAO 1959, pp. 111-13, 24]

It should be apparent from the foregoing discussionthat salinization is a generic problem of Iraqi agriculture.Among contributing factors are high evapotranspirationrates caused especially by extreme summer temperatures,limited surface runoff owing to very low alluvial gradients,and even more limited lateral movement of the alwayshighly saline groundwater as a result of fine sedimentsize and poor soil structure. But it is also clear that thechallenge of "living with" salt is variable, not constant.Its sometimes severe effects can be traced into the ancientpast, but total, irretrievable losses of productivity are notsuggested by either the historical or the archaeologicalrecord-even in the absence of the massive drainageprograms on which modern agricultural developmentincreasingly relies (Jacobsen and Adams 1958; Adams1965, pp. 17-18). The pervasive modern problem, inother words, is not to be understood as the direct, in-eluctable outgrowth of natural forces. Salinization isindeed a recurrent, widespread condition on the lowerMesopotamian plain, but its recent, rapid intensificationand heavy economic impact are also products of the un-precedented technical and institutional means with which

modern agriculturalists are seeking to adapt to the peren-nial uncertainties of Euphrates water.

SAs I noted earlier, the monthly flows recorded infigure 1 approach more closely than any annual totalsthe conditions to which farmers of any period wouldhave had to adapt. In the absence of any possible provi-sion for extensive water storage before very recent times,it was minimum reliable flows at certain critical juncturesduring the winter growing season that set limits on theextent of cultivation-and hence indirectly on the popu-lation of the alluvium. We shall see presently that or-ganizational, not technical, means were found to trans-cend those limits during certain times of unification,stability, and heightened rural investment like the Sasan-ian period, and that at such times demographic levelsresponded accordingly. But for by far the greater part ofthe plain's history of human occupancy the limits wereforbiddingly real, and establishing even very roughlywhat they were is correspondingly important for an un-derstanding of that history.

A perduring part of Mesopotamian agricultural prac-tice is the application of a limited number, of cycles ofirrigation water during the winter growing season.Smaller, more frequent applications would impose theunreasonable requirement that water supplies be main-tained permanently at adequate levels in all componentsof the extensive canal systems and would also greatlyincrease losses through evaporation and seepage. Thispractice of well-defined, short intervals of watering isfirst identified for us already in the Sumerian Georgica,or "farmer's almanac," known from second-millenniumcopies but surely reflecting practices that were alreadyold in the third millennium. The Georgica speaks of theneed for a preliminary watering before the first plowingand seeding in the fall, then describes four iterative ap-plications before the harvest between late April andeIrly June (Salonen 1968, pp. 202-12). Four or five water-ings also were specified by Ibn Wahshiyya, writing ofestate management no later than the tenth century A.D.(El-Samarraie 1972, p. 62), and are reported as still typi-cal today (Adams 1965, p. 16). If we take into accountthe winter cereal requirements for consumptive use thatwere indicated earlier, some 0.55 meters of water overthe cultivated surface, it is apparent that the averageamount of water supplied in a single monthly wateringshould be slightly more than 10 centimeters. An onlyinsignificantly smaller figure (9.6 cm) can be establishedfrom an eleventh-century Islamic text providing calcula-tions of areas of winter crops that could be irrigatedwith norias and other irrigation lifting devices (Cahen1947-48, p. 130). Working with this amount as a cus-tomary rule of thumb, further use can be made of themonthly Euphrates flows that are recorded in figure 1.

What is crucial is that fall and early winter flowssharply restrict the amount of land that can be irrigated.

S

oi.uchicago.edu

Water, Land, and Life

Let us concentrate upon the recorded averages, whilerecognizing that half of any given sequence of years prob-ably would not have reached these amounts and that abetter approximation of the reasonably reliable flowslimiting agriculture are likely to have been only abouttwo-thirds as much. The following tabulation recordsboth the approximate monthly flows of the winter grow-ing season and the total areas that could be irrigated ifthe entire Euphrates flow were diverted to this purpose:

Month

OctoberNovemberDecemberJanuaryFebruaryMarchApril

Flow in Cumecs(m'/second)

350450600700800

1,2002,100

Area(km2)6,0487,776

10,36812,09613,82420,73636,288

The figures in the right-hand column take into accountwinter crop requirements of 0.55 meters and losses ofone-third of gross diversions through evaporation, seep-age, and so forth, but they ignore the aforementioneddifference between reasonably secure minimum flowsand recorded averages. Even more plainly unrealistic isthe assumption that the entire early flow could be di-verted, leaving the lower course of the river absolutelydry. Further, the right-hand column disregards compet-ing domestic requirements and the heavier consumptionof other forms of land use such as vegetable gardens andorchards. What these calculations do establish, there-fore, is an upper limit of surely much less than 12,000square kilometers-probably on the order of 8,000 squarekilometers or even less-that could be adequately irri-gated in any given season with Euphrates water even ifthe time of initial plowing and sowing was extended fromOctober until as late as the end of January. (The totalarable area, cultivated on the basis on alternate years infallow, was of course twice as large.)

Why, one may ask, would the time of sowing not havebeen extended later than January in order to take ad-vantage of the Euphrates' normal late spring maximum?Some variation in sowing season was indeed explicitlyprescribed by Ibn Wahshiyya as a means of permittingat least part of the crop to be salvaged in the event ofblight, drought, or flood (El-Samarraie 1972, p. 60). Butthe crucial contrary consideration is that later sowingwould also delay the harvest, and that postponing thelatter until after the advent of hot summer weather en-tails rapidly mounting losses from crop diseases, insects,and other pests. In the traditional agricultural regime,therefore, rising population and an extension of the agri-cultural frontiers impose irrigation requirements that

can be met only if a growing proportion of cultivatorspostpone sowing, with, as a further consequence, unre-corded but by all accounts very heavy losses in agri-cultural output.2

But what of the Tigris? Why should that even moreimpressive source of water be excluded from these calcu-lations? The Euphrates watershed was, to be sure, theprimary region of ancient settlement. Why were all but

>two or three of the known historic towns of any im-portance before Hellenistic times distributed alongbranches of the Euphrates rather than the Tigris if thetiming of the flood on the latter was certainly not in-ferior to that on the former and perhaps was slightlymore favorable? Two factors contribute to an answer.

> First, the greater size of the Tigris was more of a dangerthan an attraction to societies with limited technicalmeans. More dependent on rainfall in its watershed, ittherefore also floods more rapidly and destructively afterwinter and spring storms. The 1954 Tigris flood-theworst of modern times, although well below the theoreti-cal maximum that hydrographers can conceptualize fromrainfall characteristics-brought a raging, uncontrollablecrest of 16,000 cubic meters per second (cumecs) belowBaghdad at the confluence of the last left-bank tributary,the Diyala. The greatest modern flood on the Euphrates,by contrast, was the 1929 crest of 5,200 cumecs (FAOMediterranean Development Project 1959, p. II-5; Wolfel1962, p. 164). The latter, too, was far beyond the capa-bilities of existing bunds; periodic catastrophic floodsare common to both rivers. But the difference betweennormal and extreme was less, even taking into accountthe lesser average flow of the Euphrates, and in thatsense the Euphrates has always been more manageable.Within the constraints of ancient technology, even enor-mous investments of human labor in bunds and otherprotective works along the Tigris provided only very

\qualified security. Towns founded in its vicinity had tobe sited at such a distance from the river that some ofits advantages were lost, or else they were periodicallyexposed to inundations so severe that the only safetylay in general abandonment.

- A second factor is immediately apparent to an observertoday, although the equivalence of conditions in thefairly remote past is perhaps problematic. The Tigris to-day enters the alluvium as a single great stream im-mediately below the modern town of Samarra, and itremains one for 400 kilometers more. The Shatt al-Gharraf is a major right-bank effluent at Kut, to be sure,but at least under modern conditions its successful opera-tion depends on the existence of a barrage across theTigris. Nothing suggests that any such weir lay withinthe earlier limits of human capability. For most of thisdistance, then, the Tigris was and is not only more turbu-lent and unpredictable as a potential source of irrigationwater, but also more deeply entrenched. Usable canal off-

6

oi.uchicago.edu

Water, Land, and Life

takes could be cut through its banks if the canals them-selves were extended far enough down the backslope ofthe Tigris levee to provide a water level higher than theadjacent fields to be irrigated. But such offtakes weredirectly exposed to the worst the Tigris had to offer.Built only at a heavy cost in labor, they could be sud-denly swept away or submerged beneath a deep blanketof silt. In flood, the Tigris bed load may reach twentythousand parts of silt per million, five times that of theNile and more than three times the highest level knownfor the Euphrates (Cressey 1960, p. 144). Insofar as themodern Tigris regime is an accurate index to the ancientone, then, that river must long have seemed both toodifficult and too unpromising to tame for irrigation. Inthe absence of other constraints, a less risky decisionfor a Sumerian or Babylonian ruler with human resourcesadequate to divert a part of the Tigris was to use theminstead for subjugating his smaller Euphrates neighbors.

It must also be observed, however, that the right bankof the Tigris was eventually canalized fairly extensively.The base map of the survey makes this immediately ap-parent. Although some of the ancient levees shown inthat area might once have been the tails of Euphratescanals having their origins far to the northwest, otherscan only be explained as drawing their supplies directlyfrom the Tigris. From the clarity and continuity of canaltraces on the air photographs as well as from the detailsof ancient settlement uncovered by the survey, it is clearthat all or virtually all of this activity dates to Hellenisticand later times-principally Sasanian and Early Islamic.There are occasional references to earlier canals emanat-ing from the Tigris (see below, pp. 134, 159), but nothingsuggests that they continued to function over long periodsor achieved real economic importance. The few townslike ancient Akshak/Upi that definitely were situatedalong that river lacked political importance as well andprobably should be regarded as maintaining their ownsmall, autonomous irrigation enclaves. Only at a fairlylate date, then, did the inhabitants of the alluvium under-take to utilize the Tigris on an extensive scale. That newcapacity is appropriately symbolized by the choice of aTigris site first for the Hellenistic city of Seleucia, laterfor the Parthian and Sasanian capital at Ctesiphon, andfinally for the 'Abbasid founding of Baghdad.

These developments portend more than the overcom-ing of some localized barrier to settlement. A character-istic of the vast canal system that was introduced by nolater than the Sasanian period was that provision wasmade to supply almost all of its component elementsnot merely from the northwest, along the main gradientof all Euphrates canals, but from the northeast. Surelythe explanation for what would otherwise constitute aninexplicable case of overdesign of the system at giganticcost is that Tigris water was periodically supplied toareas that formerly had had to depend exclusively upon

the Euphrates. And therein lay a way to escape the re-strictions of Euphrates flow during the early winter grow-ing season. Tigris water could provide a vital supplementat a time when it still posed little danger of serious flood-ing and when its silt load was still relatively moderate;then, as the season advanced, the headworks could beclosed off and protected insofar as possible and the entireburden of irrigation shifted to the now greatly enlargedEuphrates. It was a scheme of extraordinary comprehen-siveness, entailing the artificial reshaping of the relation-ship between major rivers and their many effluents andan unparalleled degree of direct state intervention in day-to-day irrigation management. These impressive advancesultimately proved to be accompanied by a correspondinggrowth of new and unprecedented risks, but their im-mediate effect was to support a proliferation of new citiesand a dramatic rise in population. By early medievaltimes, then, agriculture and settlement on the lower Meso-potamian plain are no longer to be thought of as con-sequences of the deployment of the Euphrates alone butmust be seen as mushrooming outgrowths of what hadbecome a single, interdependent system integrating thewaters of both the Euphrates and the Tigris.

Save for this impressive but transitory achievement,the concern of this study is focused on the Euphrates asthe primary influence on historical patterns of Mesopo-tamian settlement. Several features of its regime needsome reference here, in addition to the effects of varia-bility in its monthly and annual flow that I have alreadydescribed. Its course, in the first place, is typically ananastomosing one for a considerable part of its traverseacross the alluvium to the head of the Gulf. This impliesa natural pattern of multiple channels, separating andrejoining, rather than the single incised one that theTigris maintains until its lowermost reaches. Since theflow is divided among a number of smaller, less danger-ous channels, correspondingly smaller, more easily ini-tiated and managed systems of canal irrigation are suffi-cient. For an anastomosing pattern to exist, moreover, weare necessarily dealing with an aggrading stream thatregularly overtops its banks and maintains an elevatedbed on a natural levee made up of sediments carrieddown by the steam itself, either as a moving bed loador in suspension. The general process may be brieflydescribed as follows:

Deposition on levees occurs when a stream overtops itsbanks. The velocity is checked, so that not all of theprevious load can be transported, and sediment is de-posited adjacent to the banks. The coarsest debris is laiddown close to the channel and the finer material furtherdown the levee at a greater distance from the stream.Deposition rate is at a maximum close to the channeland declines down the levee, giving the slope into thefloodbasin. When a stream is not in flood, its levees are

7

oi.uchicago.edu

Water, Land, and Life

attacked by atmospheric agencies and in particular areeroded by rainfall. [Allen 1965, p. 121]

This setting has obvious advantages for technicallyill-equipped irrigators. The backslope of the levee, lead-ing away from the stream and into an adjoining basinor depression, provides natural drainage and is highlysuitable for cultivation. Yet at least the upper surfaceof the stream itself is generally elevated in relation tothis land surface. Fairly short, shallow cuts in its banksare sufficient, therefore, to bring water out onto thebackslopes at or above the land's level, establishing "com-mand" of it for irrigation. There is, to be sure, a moretroublesome corollary to these advantages. Anastomosingstream channels are not naturally fixed but are given tomovement, both through continuous processes of lateralchannel-cutting that accompany aggradation and throughmore extensive course substitutions during periodic flood-ing. The attendant cost of irrigating these areas, as weshall see, is a continuing and not always successful battleto maintain a gradual, manageable rate of channel changethat minimizes destruction of canal headworks alongthe banks and that prevents towns from being suddenlyisolated from their water supply.

The history of the Euphrates floodplain includes ex-amples both of abrupt, permanent channel replacementand of slower movements marked by repeated restora-tions. Sudden avulsions are seemingly less characteristic,even though the water level exceeds the elevation ofthe backslope depressions during the winter and springportion of the hydrological cycle. Bank failure duringtimes of high water is accompanied by a sudden decreasein the velocity of flow and a consequent tendency forsediments to be deposited, filling the gap. Hence mostescape crevices are healed when the river stage falls. Insome cases, however, a distributary is formed that canmaintain a channel throughout the year. If its courselength and gradient offer a comparative advantage, sucha distributary can gradually enlarge (primarily throughthe erosive processes associated with a succession of highannual floods) to become a permanent diversion (Schumm1977, pp. 304-5).

A second morphological feature, applying in this caseto the Tigris as well as to the Euphrates, is that the chan-nels generally exhibit meander patterns. Meander geom-etry is exceedingly complex and for the most part neednot concern us here. Minor stream sinuousities are re-lated to variations in bed load and current, and the latterin turn initiate erosional and depositional variations onopposing banks. Individual meander loops develop grad-ually out of these, affected by many factors includingstream volume, slope, and bank material. Ultimatelymeanders develop to a point where a cutoff occurs moreor less directly across the neck of the loop, leaving theold channel as an oxbow lake that gradually fills with

sediment, and the whole process begins again. These con-tinuing, localized movements also widen the stream leveeinto a much broader and more regular meander-belt leveethat encompasses the extremes of amplitude of the entiresequence of meanders.

Meander loops on other rivers, in addition to theirgrowth in a direction perpendicular to the prevailingcourse of the stream, frequently exhibit an asymmetricalcutting pattern that involves them in slow downstreammigration or "sweep" along the course. That pattern,were it to occur commonly in Mesopotamia, would havehad very destructive effects on many of the remains ofancient settlement that directly adjoin stream channels.However, comparing contemporary maps with naviga-tion charts drawn more than a century ago provideslittle if any evidence of "sweep." Sporadic cases of it canbe seen in the air photographs along former channels,one of which will be discussed presently, but for thisalluvial plain as a whole meanders generally are limitedto lateral growth followed by extinction through cutoff.

A substantial body of research indicates that meandersmust be understood as systemically derived from hydro-logical principles and cannot be dismissed as random orepisodic. Meander geometry can be characterized by aseries of more or less regular relationships between ob-served variables, such as stream width, stream flow ordischarge, and meander wavelength, amplitude, andradius of curvature. To be sure, not all of these relation-ships are equally invariant. Along the lower MississippiRiver, for example,

one considers natural levee patterns and spacing, notchannel width, to decipher channel history or correlatea channel segment with some upstream equivalent. Thebends along a channel with higher discharge will bemore widely spaced and the loops of meanders will belong in comparison to the meanders along streams withlesser discharge. I have found that average distancesbetween bends along a channel for an airline distance of100 miles or more provide a far better index of river dis-charge than conclusions based on criteria such as measur-ing radii of curvature. [Russell 1967, p. 74]

The nature of the relationship between meander spac-ing or wavelength and stream discharge can be furtherspecified. It might seem that the maximum rate ofmeander cutting would occur at bankfull stage, but ex-perimental and empirical studies suggest that a range ofdischarges during falling stages instead exercises the dom-inant influence over meander wavelength. This explains amuch lower standard error of correlation of wavelength(Lm) with either mean annual discharge (Q) or the meanof the month of maximum discharge (Qmm) than withbank full discharge (Qa). At any rate, these relationshipsare reported to take approximately the following forms:

8

oi.uchicago.edu

Water, Land, and Life

Lm = 106.1 Q 0.46 = 80.0 Qmm 0.46 = 8.2 Qb 0.62 (Carlston1965, p. 885).

Almost certainly, however, no statement of the rela-tionship that is confined to these two variables, dischargeand wavelength, can be wholly satisfactory. Empiricalstudies have shown that average particle size in the sedi-ment load also influences meander wavelength, and thatthe kind and density of local vegetation may also be afactor. Whether for these or other reasons, different in-vestigators have derived slightly different quantitativeexpressions for the key relationship between wavelengthand discharge under widely differing geographic andclimatic conditions (Gregory and Walling 1973, pp. 241-55).

As I have noted, meander development is directly re-lated to stream discharge. Meander wavelength increaseswith increasing flow. One can also express meander de-velopment in terms of increasing length of channel as amultiple of the linear distance traversed (index of sinuous-ity). Higher indexes of sinuousity imply a decliningchannel gradient, and it follows that volume of dischargeis inversely proportional to gradient (Schumm 1977, p.134). One must remember, however, that many otherfactors are at work, so that changes in discharge cannotbe uncritically imputed from changes in meander char-acteristics. In particular, meandering is a time-dependent,perhaps even cyclical, phenomenon:

It seems that once a meandering pattern is established,the hydraulic conditions in the bends and variations inbank material cause enlargement in meander amplitudeand decrease of radius of curvature until eventually acutoff occurs. Cutoff of one bend will by local steepeningof the channel gradient cause scour upstream and deposi-tion downstream of the cutoff. Both processes are likelyto trigger additional cutoffs by increased bank erosionupstream and by increasing flood heights downstream.In a reach meanders may enlarge to a critical threshholdof high sinuousity when, because of greatly reducedgradient, aggradation will precipitate cutoffs. Inevitably,meander growth will lead to a threshhold of channelinstability, at which point the channel will straighten out.However, this process may involve very different periodsof time, depending on water discharge, sediment load,and the nature of the sediments comprising the bankmaterial. [Gregory and Walling 1973, p. 142]

Next we may consider the question of sediment load,already presupposed by the processes of levee forma-tion-and, in fact, by the existence of the alluvium itself.An influential contribution of a generation ago on thissubject needs to be presented at some length, since it willpresently be seen to have an important bearing on con-temporary understandings of riverine history:

The Euphrates has an average silt content at Ramadi (at

a point near the surface of the stream at the centre of theriver) of 553 dry grammes per cubic metre and the aver-age discharge of the Euphrates at Hit throughout theyear is 710 cubic metres per second. Corresponding figuresfor the Tigris at Baghdad are 787 dry grammes per cubicmetre and 1240 cubic metres per second. Assuming aspecific gravity of 2 for the compacted silt this gives76.2 X 107 cubic feet of silt per year as the annual burdendistributed by the combined rivers; spread out over 100square miles it would have a depth of 3.28 inches. Asthe silt content in other parts of the river would be con-siderably greater than at the centre of its surface (averagefigures are being used for flow) this rate of accumulationcan be regarded as a minimum only. It has been esti-mated that the Karun carries down 1.1 million cubicyards of silt every year, which is equivalent to 0.13 inchesspread over 100 square miles. The Karun silt is carriedinto the Persian Gulf but the silt of the Tigris andEuphrates is deposited in the extensive marshes into whichtheir flood waters flow. The amount of silt carried tothe sea by the Shatt al-Arab is such a small proportionof the total that it can be neglected in these calculations.

The area of the lakes and marshy areas over which theTigris and Euphrates can distribute silt is about 1500square miles, so that if all the silt burden of these tworivers was deposited in them, its minimum average thick-ness would be 0.22 inches per year. The actual amountof silt distributed annually may be much greater and,taking windborne silt into account, it would seem thatthese marshes and lakes could not survive more than afew hundred years unless they were rejuvenated in someway. [Lees and Falcon 1952, p. 29]

I will postpone for later discussion the fundamentalthrust of this work, concerning the rate of recession ofthe head of the Arabo-Persian Gulf. But there are prob-lems with details of the arguments offered in the foregoingpassage that are relevant to our consideration of entirelydifferent themes. First, some additional doubt must beinterjected about the basic sediment load measurementson which Lees and Falcon relied, although they haveseemingly been cautious in citing them as "a minimumonly." Generalizing about such measurements, an author-ity on alluvial stream behavior has observed:

Estimates of transported sediment load have little mean-ing, except for the places where measurements are made.A river is far more turbid between a scour pool and thenext bar downstream than between that bar and the fol-lowing scour pool. Transported load always has animportant component of motion toward places whereturbulence is less intense. [Russell 1967, p. 70]

While an enormous mass of suspended material is un-deniably carried down annually by the twin rivers, thereare accordingly some grounds for doubt whether thespecific figures offered above can err only in the directionof being too small rather than too large (cf. Buringh 1960,

9

oi.uchicago.edu

Water, Land, and Life

p. 51). But a much more substantial difficulty arises fromthe authors' assumption that all this silt can be consideredas being deposited in lakes and marshes of relativelylimited area. Agronomists and irrigation engineers work-ing in the vicinity of Baghdad, at the opposite end of thealluvium from the low-lying lakes and swamps, have longbeen conscious of the recent and ongoing deposition ofsediment in that region. Apart from the very considerableaccretions represented by the broad river levees, the coun-tryside is crisscrossed by the great ridges of former canalspoil banks that are often 3-5 meters and more in height.And active deposition of course extends into every fieldthat irrigation water reaches. Buringh and Edelman, forexample, found from 20 centimeters to 1 meter of recentsilts over heavy basin clay in irrigated fields within atransect between the Tigris and Euphrates southwest ofSuwaira. Irrigation levees covered about half this trans-sect, they observed, and basins (which of course consistof thinner layers of finer-textured sediments, also ofriverine origin) covered the other half (1955, pp. 41, 45).An immensely wider region of ongoing silt distributionthan Lees and Falcon consider thus must be taken intoaccount-as wide, in fact, as all the land that is irrigatedand seasonally flooded.3

Aeolian erosion and redeposition is a further factor,to which I will return more systematically. Most of theunconsolidated, wind-borne material consists of fine,sand-sized particles, the greater part being crumbs ofsilty clay loam flocculated by salt, "pseudosand" (Schilstra1962). There is little doubt that by far the greater partof these components has been locally derived, from winderosion of the dry surfaces of basins and levees. Some isundoubtedly trapped as it moves by marshes and lakes,as Lees and Falcon indicate, but the observational evi-dence summarized in chapter 2 strongly suggests thatwind deposition takes place on virtually as general ascale as wind erosion. Hence the action of the wind isbetter considered as another of the diverse forces that actnot to concentrate riverine sediments or their derivativesin a few loci but to disperse them very widely.

There can be no doubt that soil deposits stemmingultimately from Tigris and Euphrates river sediments varytremendously in rate of accumulation across the alluvium.Depths of virgin soil underlying canal levee or occupa-tional deposits of as much as 7, 8, and even 10 metersbelow adjacent plain level have been recorded east ofBaghdad, of well in excess of 5 meters at ancient Isin,and of 6 meters below the great mound of Warka in thesouth (Adams and Nissen 1972, p. 6; Hrouda 1977, pp.19-20, 147). Those measurements are compromised, tobe sure, by uncertainty as to how much the "adjacentplain" level may have been elevated by slope runoff fromthe mound or levee in question, and by perhaps stillgreater uncertainty as to how much subsidence may haveoccurred owing to the weight of the mound or levee on

waterlogged soil (Adams 1965, p. 9; Adams and Nissen1972, p. 6). For much more modest and recent Indianmounds in the lower Mississippi valley it has been foundthat their "central part commonly sinks to a depth of upto two or three feet below its margins. In most cases themargins themselves have subsided" (Russell 1967, p. 18).But whatever the exaggeration in the recorded depthsreferred to earlier, surely even greater depths of deposithave accumulated over the same period along the majorriver levees.

In at least some formerly settled and cultivated areasthat now are desert, on the other hand, there appears tohave been a substantial net lowering through erosion ofthe land surface that obtained at the end of the thirdmillennium B.C. There is no other obvious explanationfor graves in ancient cemeteries being exposed abovecontemporary plain level by wind erosion. Similarly, thehundreds of small, low prehistoric sites that are reportedon in this study and elsewhere cannot all represent highmounds of which only the summits have somehow beenleft uniformly exposed above a deep blanket of lateralluvium. Perhaps even more significant in this respectare the extensive traces of third and fourth millenniumwatercourses that can still be easily followed both on theground and in the air photographs. At least in unculti-vated areas that have been heavily exposed to wind ero-sion during the last millennium or so, it thus appears thatmuch of the net increment from several millennia ofearlier alluviation has been removed and transformed intodunes.

Impressionistic as much of it is, the available evidencemakes it seem likely that the heaviest net increment, aswell as current rate, of deposition occurs not in swampsin the south but at the northern end of the alluvial plain.Taking into account the randomizing or at least dispersiveeffects over time of the widespread and changing dis-tribution of irrigation waters, of aeolian erosion and re-deposition, and of the periodic movements of the riverchannels, the area in which some deposition has occurredduring the past six thousand years must be nearly as largeas the alluvial plain itself. Hence it may be useful to em-ploy the figures on silt load given by Lees and Falcon notas an argument for tectonic rejuvenation of limited areasof swamp, but as a basis for calculating average deposi-tion over a much larger area of 60,000 square kilometersor so that constituted the core of the alluvium.4

This average is 3.59 x 10-4 meters per year, or lessthan 2.2 meters in aggregate for a six-thousand-yearperiod. Taking into account the more extensive deposi-tion on the upper end of the alluvium, probably extendingalong the Euphrates to somewhat below Babylon, wehave reason to expect considerably less than even thiscomparatively modest overburden in the heart of ancientSumer where this study is concentrated."

Lees and Falcon took the important step of casting

10

oi.uchicago.edu

Water, Land, and Life

archaeologists and historical geographers loose from anearlier, complacent mooring to the assumption of long-term tectonic stability. They maintained that we had toreckon with a concatenation of unstable forces, a geosyn-cline that did not necessarily remain conveniently inplace as other changes accumulated in and around it. Thesame perspective can be extended from the depths of thealluvium to its surface, to the plain that to the casualobserver appears to be perhaps uniformly rising throughalluviation but otherwise subject only to a random driftof dunes and other insignificant, localized changes.

Appearances to the contrary, the dynamism of theplain is concealed in this microtopography. Boundariesof levees and basins shift, and with them shift the localrate and character of sedimentation. River courses notonly build up levees but periodically break away fromthem; that is the only way the plain can remain a plain.Wind erosion works on exposed or abandoned levees,sculpturing them into tiny buttes and turning the un-consolidated mass that it loosens not simply into dunesthat drift harmlessly across the desert surface, but intoan extraordinarily powerful abrasive agent. Alluviationand wind deposition continue irregularly and inconstantlyover a wide area, intricately interwoven with erosiveforces to produce a shifting local outcome that defies easygeneralization. Yet all these processes of change fortu-nately take regular forms. Traces of ancient watercoursesand levees are recognizable as such, even when the greatmass of the sediments originally embodied in them hasbeen eroded away. The barren desert, seemingly variableonly in superficial detail, is in fact a palimpsest on whichthe repeated, profound modifications that have charac-terized its development can be disentangled and read.

LOCAL CLIMATE AND VEGETATION

Thus far I have dealt with the Euphrates as the primarydeterminant of a successful human occupation of thelower Mesopotamian plain that was essentially agricul-tural, and I have described the terrain-in the absenceof Euphrates flooding or irrigation supplies-as a desert.While not incorrect in a very generalized description,these characterizations must be qualified in several im-portant respects. To begin with, they fail to take cog-nizance Qf the pastoral, either non- or only semisedentaryaspects of Mesopotamian subsistence in all periods. Themaintenance of large herds, principally of sheep andgoats, cannot be understood without reference to pastur-age available for much of the year in immense semiaridand even desert tracts that were not in use by cultivators.Meat and dairy products from those herds may not havebeen an absolutely vital source of proteins and othernutrients in most circumstances, in view of the fairlygeneral availability of fish. But, without the wool fortextiles to be traded for natural resources that were

wholly lacking in the alluvium, it is difficult to believethat Mesopotamian civilization could have arisen asearly and flourished as prodigously as it did. And at leastequally important was the aspect of herds as a foodresource that could be held in reserve, not subject tothe same set of natural hazards as the crops, and capableof being moved from one location to another. From thisperspective, pastoralism was intimately linked in manyways with sedentary and even urban pursuits, and it re-peatedly served as the indispensable source of ecologicalflexibility and resilience in the aftermath of natural orsocially induced disasters (Adams 1975d).

Hence the climate of the lower plain must also beconsidered here as a factor influencing settlement andland use, if to a less significant extent than the land andthe irrigation water supplied by the Euphrates. As sum-marized by Guest (1966, pp. 17-18), its general featuresinclude high mean annual air temperature, large diurnaland annual ranges of temperature, low atmospheric hu-midity, and scanty, extremely variable rainfall that isconcentrated in the winter and spring. Rain is virtuallyabsent from late May through early October, and thelong, searing summer is the dominant season. The com-paratively short, cool winter extends from Decemberthrough February and is also well marked. Frosts occurperiodically, especially during January, and prevailinglylow temperatures at that time slow or may even com-pletely interrupt plant growth so that the normal wintercrops as well as natural vegetation are unable to makefull use of the rains. Except where there is flooding orirrigation, therefore, the conditions for plant life are"most rigorous." As Guest goes on to observe,

little growth can be made until the short spring season-just as the rains are beginning to cease. During the longsummer months there is no surface water available, whilethe intense heat and dryness of the air create conditionsof extreme desiccation. Thus the only plants able tosurvive in the lower plains of Iraq are the ephemeralannuals (which can rapidly complete their life cycle inthe spring and then lie dormant in the form of seeds forthe remainder of the year) or such deep-rooted and highlyxerophytic perennials as are sufficiently protected towithstand the rigours of the summer while drawing onunderground sources of water. This is reflected in thepaucity of plant species over the greater part of LowerIraq, since only a limited number of species come withinthe two above categories. [Guest 1966, pp. 20-21]

Table 1 amplifies and quantifies the somewhat im-pressionistic account of seasonal climatic variation thatwas given above. It has been based on meteorological ob-servations recorded at stations immediately west, east,and southeast of the area primarily covered by thisstudy-at Diwaniya, Hai, and Nasiriya, respectively-inorder to reflect conditions within that area as accurately

11

oi.uchicago.edu

Water, Land, and Life

TABLE 1 Climatic Data from Three Stations in South Central Iraq

Temperature (OC) Precipitation (mm)Month

Monthly Mean(all stations) Diwaniya (1929-58) Hai (1941-58) Nasiriya (1941-58)

Min. 6 A.M. 3 P.M. Max. z Min. Mean Max. Min. Mean Max. Min. Mean Max.

January -8.3 6.2 16.3 26.6 47.7 Tr.a 19.0 60.0 1.2 22.0 77.2 Tr. 17.0 39.0February -7.2 7.9 18.5 32.2 40.3 - 18.3 69.5 Tr. 22.0 53.9 Tr. 13.1 48.5March -2.2 11.2 22.8 36.1 35.7 Tr. 16.0 101.5 Tr. 25.9 78.7 Tr. 21.1 85.7April 1.6 16.4 29.3 42.8 28.3 - 15.3 64.4 Tr. 12.1 60.9 Tr. 15.5 66.0May 7.8 22.4 35.7 46.1 21.3 - 11.6 113.6 - 6.8 29.4 - 6.0 46.8June 14.4 25.7 39.7 49.5 18.0 - 0.2 5.3 - Tr. 0.3 - - Tr.July 18.3 26.7 42.3 50.0 17.3 - 0.1 3.7 - - - - - Tr.August 16.1 25.9 43.0 49.4 16.0 - - - - - - - - Tr.September 11.6 22.5 40.2 48.9 16.0 - - Tr. - - Tr. - - Tr.October 2.8 17.5 34.2 43.9 22.3 - 0.8 3.8 - 1.3 10.6 - 1.4 10.8November -2.8 12.6 25.2 37.2 36.7 - 15.1 64.2 - 20.4 70.6 - 22.7 69.8December -5.6 7.8 17.9 28.9 51.7 0.1 23.8 60.6 4.9 30.7 73.8 5.3 28.2 68.9

Water = year (Oct.-Sept.) precipitation totals 57.6 112.1 179.5 62.5 136.6 221.0 33.5 120.9 249.3

Sources: Ministry of Economics 1958; Hydrological Survey of Iraq 1959.a Tr. = Trace

as possible. The monthly averages given for temperatureand humidity are calculated from the averages for allrecorded years at all three stations. There is obviouslyhigh seasonal as well as diurnal variability, althoughpossibly it is only the occasional killing frosts betweenDecember and February that would have substantiallyinterfered with the agricultural cycle. Attention may alsobe called to the importance of increasing humidity be-ginning in October and continuing through the fall. In-itial plowing and seeding often is delayed until the soilsoftens as it takes on hygroscopic moisture from the air(Russel 1957).

Two features of the precipitation statistics recordedin table 1 particularly deserve attention. The first is thatthe amount of rainfall is, with rare exceptions, quite in-adequate to produce a winter crop and is better regardedas only an occasional supplement to irrigation. If 200millimeters of precipitation during the major growingcycle from October through April is regarded as theabsolute minimum for dry farming, that "water-year"figure was reached only three times in the sixty-six ag-gregate years of recording at the three stations. More-over, it is misleading to deal with precipitation in termsof aggregates. Its distribution at crucial intervals duringthe growing cycle is at least equally important. As notedabove, heavy rains concentrated in January may do littleto promote growth if prevailing temperatures are lowenough to approach the critical level of 100 centigrade(500 F). Heavy rains after March, on the other hand,come too late to influence the size of the grain harvest.When these circumstances are taken into account it must

12

be said that crop production without irrigation is virtu-ally never possible, except in rare instances at restrictedlocalities where runoff from a larger area can be con-centrated on a particular field.

The second characteristic of precipitation is its enor-mous variability. Heavy showers can be expected atany time from November through much of May, butthey may not occur before December or even Januaryand may be almost completely suspended for as much astwo or three months during the growing cycle. Thisagain reinforces the dependence of agriculture upon irriga-tion, since the Euphrates, for all the variability of flowwe have seen, is vastly more dependable as a watersource. At the same time, rain introduces a speculativeelement that affects both the size of the cultivated areaand the aggregate output. Particularly if there is heavyearly-season precipitation, farmers may be inclined toquickly enlarge the area they have planted, since highlevels of residual soil moisture may permit a crop withless irrigation than is customarily necessary.

There are obvious difficulties in seeking to generalizeabout the historic and prehistoric past from climaticconditions obtaining today. It is certainly correct, as hasrecently been observed, "that the evidence available tous is far too limited in its scope and quantity to supportany generalizations or far-reaching conclusions" (Oatesand Oates 1977, p. 115). Yet it must also be said thatnothing yet available in the palynological, geomorpho-logical, or archaeological record suggests that the climateof the region since the Pleistocene was for a time suffi-ciently wetter to permit sustained, significant dry agri-

oi.uchicago.edu

Water, Land, and Life

culture on the alluvial plains, except possibly for a near-piedmont band along the foot of the Zagros. Hence itseems incontestable that agriculture was introduced intolower Mesopotamia only on the basis of irrigation andthat the region has remained a classic example of irriga-tion agriculture ever since. That does not mean, however,that possible climatic variations would have had onlyinsignificant effects upon human life there. In the firstplace, shifts in the volume or periodicity of rainfallsurely would have been sufficient to have at least marginaleffects upon agricultural productivity. Particularly in cir-cumstances where the food requirements of a populationand the available irrigation supplies were approximatelyin balance, such shifts might decisively tip the balanceone way or the other. Additionally, attention has beencalled above to the potentially critical factor of Januarytemperatures. A climatic oscillation that led to a depres-sion of those temperatures would interrupt the growingcycle and therefore delay it for a longer period, un-doubtedly with adverse effects upon productivity.

But far more important would have been the effectsof even slight climatic shifts upon the pastoral componentof society and subsistence. We shall see in chapter 2 thatvariations in precipitation even during the eight-yearperiod of field reconnaissance have led to pronouncedchanges in desert vegetation, and there is every reasonto believe that far greater oscillations occurred repeatedlythroughout the ancient past. Most of them, to be sure,probably involved localized shifts in the availability ofpasturage rather than dramatic advances or declines inthe carrying capacity of the alluvium as a whole. Someimpression of the frequency and character of such shiftscan be gained from figure 2. Based on a succession ofeighteen "water-years" at the same three stations whoserecords provide the data for table 1, it distinguishes thehighest and lowest thirds in the range of annual precipita-tion totals at each station and plots their years of occur-rance. In about half these cases it appears that trends werefairly widespread; lows occurred simultaneously in 1942,

DTwaniya

Hai

Nasiriya

1944, and 1952, and highs in 1945, 1955, and 1957. Pre-sumably these were times when little was to be gained bymoving herds in search of better pasturage. On the otherhand, there were also years like 1951 and 1956, in whichlow precipitation was reported at two of the stations andhigh precipitation at the third. Clearly, these were timeswhen some movement would have been advantageous forherds and herdsmen not fixed in place by permanent fieldsand dwellings. Perhaps even more important than the rela-tively rapid oscillations and ensuing shifts were succes-sions of several years in which abnormally high or lowprecipitation occurred in one or more regions. AroundDiwaniya, for example, a lengthy period of low rainfall inthe late forties and early fifties would have led to a pro-gressive denudation of the fodder available for flocks onthe open desert, while the unusually high rainfall of thelate fifties would have had the opposite effect.

What is historically crucial about variations in precipi-tation like these, whether major and long-term or minorand brief, is that they bind the pastoral components ofMesopotamia to a set of ecological constraints somewhatdifferent from those affecting the cultivator. Hence mo-bility remains an important advantage for the former,while, in the absence of herds, the latter would have muchto gain by investing cumulatively in the improvement ofhis land. However, we must also bear in mind that culti-vating and herding are normally carried on simulta-neously, within the same agricultural communities andoften by the same families. As noted earlier, there arereciprocities in labor and subsistence that strongly en-courage this diversity. It could well be, therefore, thatclimatic variability has also had a profound but indirecteffect on historic patterns of cultivation. It appears tohave reinforced other natural factors such as salinizationin acting to retard capital investment in agricultural im-provements and to favor the retention of an extensivesystem based on alternate years in fallow.

Given the omnipresent fact of climatic variability, itfollows that extensive overgrazing must occur almost ev-

HIGH - e* *.... . __.

Fig. 2. Periodicity in high and low precipitation at three stations in South CentralIraq (for "water-years" beginning the preceding October). From Hydrological Survey

of Iraq 1959.

1313

oi.uchicago.edu

Water, Land, and Life

erywhere with great frequency, even if flocks and the sup-ply of natural vegetation remain aggregatively in roughbalance across the region as a whole. To this destructiveforce must be added the more selective activities of thefuel-gatherers, in this case aimed not at the ephemeral an-nuals but solely at the woody perennials. The effect, asGuest has noted, is a massive removal of perennial herbsand shrubs,

leaving in many places but few relics of the perennials: theleast palatable, the toughest and the most heavily-armedspecies. It is only in the more remote or protected placesthat we can get an idea of the form the vegetation wouldhave assumed if it had not been interfered with. . .. Insome parts of the plain there are large or small strips andpatches of vegetation which may appear to be natural butare in reality of secondary origin-still largely influencedby the agricultural history of the district, the proximity ofadjacent habitations and pastoral activities. [Guest 1966,p. 66]

Following Zohary, Guest accepts a definition of desertthat is arbitrary in the sense that it is based on extent ofplant coverage rather than on the occurrence of distinctivetypes of vegetation. Plant communities of the desert aresaid to be "open, often scattered, and usually more or lessrestricted to favourable habitats," while those of thesteppe are closed and cover the greater part of the landsurface, at least at certain seasons. In these terms, the un-irrigated terrain that was the special focus of this studyis perhaps best described as sub- or semidesert. Large,completely barren tracts do occur in some areas, as I willnote again in chapter 2. In most cases, however, they areeither the secondary products of overgrazing and exces-sive fuel-gathering or are caused by soil salinity or topo-graphic factors. "Broadly speaking the check to grazingthroughout the year is not lack of vegetation, except lo-cally in areas of edaphic or secondary desert, but lack ofwater points at which the animals may drink during thelong dry summer season" (Guest 1966, pp. 68-69). Finallywe turn to the characteristic plant communities them-selves, relying once again on Guest's authoritative study:

The typical natural vegetation of the sub-desert consists ofmore or less scattered perennial shrublets (e.g. Haloxylonsalicornicum, Artemisia herba-alba, Achilles fragrantis-sima, Rhanterium epapposum, etc.)-practically nowherecompletely closed and often very open, and including bar-ren tracts of edaphic desert and secondary desert. In springthe open spaces between the bushes are generally occu-pied by a relatively sparse crop of annuals. In depressionsand other favourable habitats the coverage of vegetationmay approach 70% or more, while after a wet winter theherbage between the bushes may become almost luxuriantduring its short-lived spring growing and flowering sea-son. Where the bushy perennials have been destroyed the

sparse ground vegetation is usually dominated by Stipacapensis, this being everywhere a sign of degradation.[Guest 1966, p. 71]

AN OVERVIEW OF MAJOR WATERCOURSESUCCESSION PATTERNS

In arguing that older archaeological theorizing aboutthe advance of the alluvium at the head of the Arabo-Persian Gulf was oversimplified, Lees and Falcon weresuitably cautious in advocating an alternative reconstruc-tion of its position. Yet, though it still remains obscure, theposition of the Gulf shoreline is fundamental if we are todescribe in basic outline the succession of rivers and land-forms that have characterized the Mesopotamian plain.Hence their interpretation remains an appropriate pointof departure:

The Tigris, Euphrates and Karun rivers are not buildingforward a normal delta; they are discharging their load ofsediment into a tectonic basin which is the successor to ageosyncline in which many thousands of feet of sedimenthave been accumulated in the past, over a period to bemeasured in hundreds of millions of years. The balancebetween subsidence and sedimentation in the recent pastseems to have been finely poised; subsidence was episodicand in the intervals the depressions tended to fill up withsediment. But in general subsidence has been dominant,with the exception of some minor local uplifts represent-ing a late movement of anticlinal structures....

We hesitate to suggest a position for the head of thePersian Gulf at the time of the Flood which gave rise tothe Babylonian legend, for it is equally impossible to haz-ard a guess at the date or the extent of this event. There isno acceptable historical evidence that the head of the Gulfwas ever very far up-country from its present position, andthe evidence which we have collected suggests on the con-trary a complex pattern of advance and retreat of the sea;precise dating is not yet possible. Subsidence of the Gulfbottom combined with a rise of the sea-level may evenhave buried the remains of many cities below river-bornesediment or below the waters of the Persian Gulf. [Leesand Falcon 1952, pp. 38-39]

More recently, considerable doubt has been cast notonly on the extent of tectonic instability that Lees andFalcon hypothesized but on the necessity of invoking tec-tonism as even a partial explanation for any of the phe-nomena they reported. Early critics of their thesis weresomewhat offhandedly dismissed as "far from adequatelyequipped to engage in geological battle" (Mitchell 1958, p.127), but much of their supporting evidence has now beenshown to be consistent with a variety of interpretationsother than the one they offered (e.g., Kirkby 1977, p. 283).Continuous submarine terrace formations have beentraced that seem to preclude any major tectonic movementsince at least late prehistoric times. Their attribution of

14

oi.uchicago.edu

Water, Land, and Life

raised marine terraces and drowned valleys to tectonicchanges has likewise been questioned, on the grounds thatit fails to take into account more recently accumulatingevidence of climatic change and consequent oscillations inHolocene as well as Pleistocene sea levels. 6 "Above all,alluviation and stream incision, not to mention drownedirrigation systems, do not in themselves indicate recenttectonism. Instead such physical changes point to varia-tions of environmental interrelationships far more com-plex than originally anticipated" (Larsen 1975, p. 56).Thus the case is today being argued with renewed vigorthat since the fourth millennium B.C. the shoreline at thehead of the Gulf has advanced a minimum of 150 to 180kilometers (Larsen 1975, p. 53).

This is not to imply that the issue is by any means set-tled. Turning from primarily geological considerations,the argument by Lees and Falcon that "there is no ac-ceptable historical evidence that the head of the Gulf wasever very far up country from its present position" alsoremains in sharp contention. Implicit in Sidney Smith'sreply to their article was the complaint that disproportion-ately higher standards of certainty were asked of historicaldocuments than of the unsystematic and geographicallyscattered geological data that were fitted together in sup-port of what remained a fairly speculative hypothesis. Theconsensus of those familiar with the cuneiform textualevidence, as he then stated it, was that "there was a con-tinual recession of the head of the Gulf from the earliesttimes onward" (Smith 1954, p. 396). Yet the nature of theinformation was such that the tectonic argument couldnot be completely controverted; no locations along theancient shoreline at properly specified times and placeswere ever attested.

How has the situation changed with regard to the ear-lier historical periods during the quarter-century or sosince the tectonic hypothesis first was formulated? Weknew then that some of the older Sumerian towns like Urand Lagash were in fairly close but tantalizingly unspeci-fied proximity to the sea; that remains the case. The num-ber of excavations in lower Iraq has of course grown sincetheir paper was published, but none has been conductedon the lonely ishans rising out of the swamps in the greatempty area east and southeast of Ur and Lagash. That isthe area that presumably would have been available forearly settlement if the shoreline has not advanced to itspresent position but has merely shifted back and forth inthe same general vicinity. None of the extensive, increas-ingly systematic surveys of lowland Iraq had been under-taken at the time Lees and Falcon wrote; but then it mustbe added that none has even yet ventured into the regioneast of the Shatt al-Gharraf that is crucial for this ques-tion. Such more limited reconnaissance as has been ac-complished, however, has failed to produce any evidenceof ancient settlement north of the Haur al-Hammar. Evenon the alluvial margins to the south of this great depres-

sion the only settlements yet recorded are mainly of Is-lamic date and apparently in no case earlier than the mid-second millennium B.C. 7

Later as well as earlier periods provide contributions tothe continuing discussion. The most recent reaffirmationof the position of Lees and Falcon is based largely onHellenistic sources. Proceeding from the firm identifica-tion of Failaka Island as the site named Icarus by Alexan-der by way of increasingly problematical ancient measuresof sailing distances, it has been argued that there is verylittle evidence of change'in the limits of the delta over thepast twenty-four hundred years (Hansman 1978, p. 60).However, the more convincing part of this case appearsto apply only to the extreme western part of the delta,where at any rate active deposition of sediment hadslowed or even ended earlier. And certainly the Islamicevidence is unequivocal in placing Abadan on an islandfacing the open sea at the mouth of a very wide Tigrisestuary extending inland as far as modern Basra, indicat-ing an advance of some 60 kilometers in at least the cen-tral portion of the delta shoreline over the course of amillennium or so (Le Strange 1895, pp. 302, 306). Ofcourse, the possibility remains open that this was theconsequence of a transitory phase of subsidence. Even forrather recent historical periods, therefore, relatively littleprogress appears to have been made in delineating the se-quence of changes in the position of the shoreline as a keyto the importance of tectonism as the underlying geologi-cal process.

There is one further development that may be of somesignificance. The Iraqi Directorate General of Antiquitieshas grown from a small organization with a handful oftrained inspectors to a major institution with many com-petent specialists. Chapter 2 will show that inspectors'field reports in the files of the directorate, necessarily com-bining older and newer records, are variable in quality.But the extent of coverage has increased progressively, andin that sense it has become progressively harder to believethat there could be substantial early remains in the south-ern swamps that have simply gone unrecognized. Apartfrom field inspections, moreover, there has been a vigor-ous, ongoing program of acquisitions of archaeologicalspecimens by purchase from local informants, and thatalso has yielded nothing suggesting early settlement in thearea in question. Without claiming that evidence is in anyway conclusive (particularly with regard to the possibilityof small, ephemeral sites), therefore, there is an increasinglikelihood that below the kingdoms of Lagash and Ur laya major hiatus of early settlement continuing well into thesecond millennium that is consistent with the presence ofa Gulf shoreline.

Two countervailing considerations may be raisedagainst what is admittedly a very tentative line of reason-ing. The first is that, if Lees and Falcon are correct in theirconjecture that the highest sedimentation rate occurs in

15

oi.uchicago.edu

Water, Land, and Life

the lakes and swamps of the south, early sites in this regionmay indeed all be very deeply buried. My contrary argu-ments have been given earlier, but the lack of a more af-firmative resolution of the question only underscores howlittle concrete information is available on the nature andrate of ongoing geomorphic processes. Second, even a totalhiatus of settlement does not necessarily imply an openarm of the Arabo-Persian Gulf. We may have to deal inthe past, as to a lesser extent we still do today, not with awell-defined shoreline but with a progression of swampsand more and more open, more brackish or saline lagoons.Such is a possible implication of Sumerian fish nomencla-ture, which distinguishes not only freshwater from salt-water species, but also brackish-water varieties, each withtheir correspondingly specialized groups of fishermen(Deimel 1931, pp. 98-99; Salonen 1970, pp. 31-50, 239-42). This possibility is even more difficult to set asideconvincingly, although the presence of small settlementsthroughout the swamps today suggests that nothing shortof open water would account for a seemingly total hiatusof settlement in the past. But here, for the present, at leastthe archaeological side of the question must continue torest-in a position not greatly different from that of ageneration ago. The old, simplistic assumptions aboutalluviation as the only significant process have been ex-punged by the work of Lees and Falcon and others whohave followed them. Now we are confronted instead witha maze of alternate possibilities, among which the availa-ble evidence does not yet permit a clear-cut choice.

Geomorphological progress of a more unambiguous,heartening kind has meanwhile been made at the upper,northern end of the alluvium. Detailed studies in the vicin-ity of ancient Sippar and Tell al-Dayr have been directedtoward relic systems of natural drainage in the area thatapparently follow the former channels of ancient water-courses. Two that run essentially parallel with the Euphra-tes are now thought to represent older courses of thatriver, stages in its irregular westward movement to a mod-ern position at the extreme western edge of the alluvium(Paepe 1971). This highly plausible reconstruction is con-sistent with the synoptic view of Euphrates evolution thatwill shortly be developed here, from an essentially differ-ent body of data. More radical is the further analysis of aperpendicular gully system that flows past Tell al-Dayr tojoin these ancient Euphrates channels near Sippar. Cau-tioning that his data are as yet insufficient for proof, Paepesuggests that this seems to represent a former Tigris courseand hence to document a primary Tigris-Euphrates junc-tion in the vicinity of Sippar. Citing evidence derivedfrom my own surveys in Akkad, he then argues also foran apparently somewhat later stage in which the Tigrisfollowed a more or less independent course southeast,roughly parallel with the present position of that riverbut about 30 kilometers to the south. Other than obvi-ously assuming that all of these stages or events are Holo-

cene rather than earlier, Paepe does not assign provisionaldates to them.

Although again resting on entirely different bodies ofdata, our findings once more coincide fairly closely. Whatcan be added here stems from more recent reconnaissancefarther downstream along the line whose significancePaepe correctly perceived, which for various reasons de-tailed in chapter 2 could not be initiated until a dozenyears after the initial reconnaissance or completed before1975. Along the southern fringes of the modern HaurDalmaj in an area that has been outside the frontiers ofcultivation since classical Islamic times, the powerfulmeanders of a major watercourse have been extensivelyexposed by wind erosion. As is detailed more fully belowin chapter 3, the adjacent settlements and sequence ofoverlying canal levees make it clear that the importantand fairly long-lived river course had begun to decline inuse by the end of the Early Uruk period and had beenpermanently abandoned by no later than the end of thefourth millennium B.c. This course appears to be a directcontinuation of the line whose upstream portion Paepeprovisionally identifies as the Tigris, and it can be fol-lowed 75 kilometers farther southeast until it disappearsunder the sediments of the modern Shatt al-Gharraf. Inso doing, it passes within 30 kilometers of Nippur and40 kilometers of Adab, both ancient towns whose laterhistoric associations with the Euphrates are close and ab-solutely unmistakable. If this is indeed the channel of thefourth-millennium Tigris, then the riverine sequence heretends to confirm Paepe's hypothesis that the cumulativeeffect of Euphrates flooding was progressively to force theTigris into more and more easterly beds. But there is alsoan alternate hypothesis-that this represents only theearliest and easternmost of a series of Euphrates bedsthat have moved progressively westward. Is further evi-dence available with which to make at least a tentativechoice between the two?

Here we must return to the regularities of meander ge-ometry that were adumbrated earlier. The quantitativeterms of the relationship may vary somewhat in differentsettings, but meander wavelength is closely proportionalto stream discharge. The meanders of this ancient water-course can be clearly plotted in several instances and hencecan be compared with modern Tigris and Euphrates me-anders. The point is not to estimate the discharge of the•ancient stream in absolute terms; there are, in any case,ambiguities in the concept of discharge as it applies tomeander geometry that would make this very difficult.But the proportionality of the meanders alone provides astrong if not entirely unequivocal indication of the sourceof the water.

Meander wavelengths on the Euphrates, taken from thevicinity of Falluja to minimize the effects of depletions forirrigation that might not have a fourth millennium equiva-lent, average about 7.5 kilometers. Tigris meander wave-

16

oi.uchicago.edu

Water, Land, and Life

lengths below Baghdad are more variable; a few exceed 20kilometers in length, but the majority clearly fall withinthe range of 8 to 12 kilometers. Meander wavelengths onthe ancient channel north of Nippur also vary, with a fewapproaching the upper figures for the Tigris, but the aver-age is approximately 7 kilometers (see fig. 3).

The interpretation that initially may seem to followfrom these figures is that the meanders and hence volumeof flow of the ancient watercourse are essentially identicalwith those of the Euphrates. However, that interpretationneglects the fact that several other channels of the Euphra-tes are known to have been carrying a substantial flowduring the Early Uruk period. Apart from the possibilitythat at least some water was reaching the western part ofthe alluvium, under the heavy silts laid down by the mod-ern Euphrates, the channel in question could not possiblyhave served any of the important towns-not to speak ofscores of villages-in the regions around Nippur, Uruk,and Eridu. Accordingly, the conclusion seems inescapablethat the ancient channel could not have taken the form itdid without a substantial admixture of Tigris water. Sinceit is perceptibly smaller than the single, modern channel

of the Tigris, on the other hand, we cannot identify it asthe equivalent of the whole of that river but only as oneof its branches.

The lower end of this course, it may be noted, has a tan-talizing but inconclusive bearing on our earlier discussionof the position of the shoreline of the Arabo-Persian Gulf.A marked change in the character of the channel stronglysuggests that it must have entered a large body of waternear the western edge of the modern zone of cultivationdependent upon the Shatt al-Gharraf (see below, pp. 31,62). But investigations of a different and much more de-tailed kind will be needed to ascertain whether seawater,as opposed to fresh or brackish swamps or lagoons, couldhave extended this far inland as late as the mid-fourthmillennium or even slightly later.

Drawing the threads of this discussion together, it ap-pears that a strikingly different general arrangement ofwatercourses existed at the time human settlements firstbecame widespread in the early fourth millennium. TheTigris and Euphrates did not remain distinct, as they dotoday, but were joined near the head of the alluvium. Atthat point, however, they did not form a single united

Fig. 3. Contemporary Tigris and Euphrates meanders compared with ancient mean-der traces.

17

oi.uchicago.edu

Water, Land, and Life

stream comparable to the Shatt al-Arab at the foot of themodern alluvium. Instead, they diverged once more intoan uncertain but probably considerable number of chan-nels that together may have constituted a shifting, bifur-cating, and rejoining combination of an anastomosingpattern and an alluvial fan as they crossed the lower Mes-opotamian plain toward a number of separate points ofoutflow into the Gulf., After the fourth millennium the Tigris passes largelyout of our ken for an extended period. Diverted farthereastward by the buildup of Euphrates sediments, it mayhave shifted abruptly into its modern, single-channel formin approximately its present position. A course even far-ther to the northeast is also possible, followed by a re-verse movement into its present position as sedimentsfrom the Diyala alluvial fan accumulated that would di-vert it southward once more (cf. Adams 1965, figs. 2-4).Still a third alternative is that it followed a number ofbraided or intersecting channels in the broad band sep-arating the first two possibilities that have been suggested.In any event, the relatively more manageable regime thatmust account for the density of Early Uruk settlement inthe area north of Nippur, presumably involving a channelon an elevated levee, apparently obtained no longer. Asidefrom a handful of towns identified with the Tigris andsporadic efforts to bring feeder canals from it into the heartof Sumer, the Tigris is not a significant factor again foralmost three millennia.

The theme of this brief overview henceforth is the spo-radic but continuing and cumulative westward movementof the Euphrates. If we assume that the Early Uruk chan-nel north of Nippur was a branch of the combined Tigrisand Euphrates rather than of the Tigris alone, then thatrepresents the easternmost point in the sequence insofaras we can at present trace it through associated archaeo-logical materials. And the abandonment of that branch isthen also the first step in the long westward sequence.

There are suggestions of not less than two and probablythree other channels that were coeval with this one, notto speak of shorter branches or connecting links betweenthem. Almost certainly there were one or more others stillfarther to the west, of which little may ever be knownbecause of the massive later sediments overlying them.Unlike the first channel, however, the others of which wehave definite knowledge remained at least intermittentlyin use for four millennia longer-until the whole of theregion through which they run was permanently aban-doned as part of the decline of the 'Abbasid Caliphate.

As the sequence of maps in the following chapters thatillustrate these courses makes clear, sinuousities and me-anders suggesting a prevailingly natural regime were in-creasingly confined to limited portions of these channels.Probably we should assume that natural, uncontrolledconditions prevailed during periods of political upheaval

and of demographic decline, and it is apparent that insome areas little or no effort was made at any time to im-pose another regime more conducive to intensive settle-ment and irrigation agriculture. But by no later than theend of the third millennium B.c., and already by the begin-ning of that millennium in some regions, the mode ofmaintenance along the greater part of these channels wasessentially canalized and artificial. Increasingly it washuman effort, rather than any predisposition of the Eu-phrates to maintain its channels in a stable, natural equi-librium, that accounts for their remaining open as impor-tant water arteries for so long.

The increasingly artificial, canalized character of thesewatercourses was punctuated by a relatively much moresudden and decisive shift during the earlier half of thesecond millennium B.C. Surface reconnaissance providesdata with little chronological precision, but there ap-pears to have been a fairly abrupt diminution of floweither at the end of the Isin-Larsa period or during the OldBabylonian period that followed. What persisted after thiscrisis, or perhaps were rebuilt after an interval of generalsocial and economic collapse, were canals following thecrests of the old levees but with much more limited carry-ing capacity. And since sediment carrying capacity isproportional to cross-sectional area, this implies that thework of maintenance involved in keeping them openmust have increased proportionately.

There is a decisive westward shift in the center of grav-ity of the canal and settlement patterns for the ensuingCassite period. A reasonable although still tentative ex-planation for the earlier crisis, therefore, is that the bulkof the Euphrates flow had shifted westward during theinterim. This is surely related somehow to the contempo-raneous rise of Babylon, both as a demographic center ofconcentration and as the political capital. A planned di-version of a greater part of the Euphrates flow into itsmore westerly branches, directly downstream from Baby-lon and hence more subject to its control, might even havebeen a central aspect of Babylonian politicomilitary strat-egy. But the conscious intervention of some human agencyneed not have played a decisive part, as the parallel case ofthe shift from the Hilla to the Hindiya channel in thenineteenth century A.D. demonstrates. Arrested only bythe erection of a barrage and control works at Hindiya, amassive diversion of flow into the more westerly channelhad earlier forced widespread abandonment of cultivatedlands in the Hilla and Diwaniya regions (Longrigg 1925,p. 311; Gibson 1972, pp. 26-29). And though this dis-astrous diversion may have taken the form of an earliercanal that gradually began to run out of control, the cu-mulative westward movement of the Euphrates continu-ing over many millennia suggests that no undue im-portance should be attached to the specific actions orevents that triggered individual shifts within that process.

18

oi.uchicago.edu

Water, Land, and Life

There need not have been any substantial human factorin the westward movement during the Old Babylonianperiod. If there were one, this might as easily reflect thegrowing incapacity of the First Dynasty of Babylon to takepreventive action during its later years as a conscious as-sertion of strength during its ascendancy.

The course or courses that became dominant during thelater Old Babylonian period must have been well to thewest of those that predominated earlier. Known ancienttowns like Borsippa, Dilbat, and Marad, in addition toBabylon, presumably were situated on or close to the newor newly enlarged branches, and the positions of thesecenters suggest that no less than two branches were inuse. But at present they cannot be more accurately locatedor described, nor do we know what further shifts mayhave occurred before the Euphrates came to occupy itsmodern beds. It is clear that the Shatt al-Hilla, the moreeasterly of the two main branches today, follows a coursecrosscutting and overlying the remains of Parthian andSasanian canal levees (see below, pp. 209-10), so that itspresent conformation is quite recent. But knowledge ofarchaeological sites in this heavily alluviated terrain is toolimited to encourage speculation as to what those leveesmay overlie in their turn. 8

SWhat can be seen, in sum, are two complementary shiftsin the distribution of water over the plain. The first in-volved atrophy of the fanlike network of natural chan-nels that apparently had its apex in an early junction ofthe Tigris and Euphrates near Sippar. Some of those chan-nels disappeared entirely, while others diminished in vol-ume and took on an increasingly canallike character thatinvolved increasing dependence on artificial maintenance.Most of the Euphrates discharge moved west, to one ormore branches along the western periphery of the allu-vium, where it is still found today.

The second shift was a direct corollary of the first. Theold natural channels had had their origin to the north-northwest of the region with which we are dealing. Nowthe main body of the Euphrates had turned almost di-rectly south, following the alluvium's western boundary.New canals became necessary, therefore, that had noneof the characteristics of natural streams and could be dugand maintained only with unremitting human effort. Toreduce their length, they were supplied by left-bank off-takes along the Euphrates, more nearly to the west or west-northwest than north-northwest. Over time, therefore,systems of levees grew up that crossed or intersected oneanother rather than taking essentially parallel directions.And the effect of those new canals, as their prevailing di-rection shifted from south-southeasterly to easterly, wasto carry Euphrates sediments out into the center of thealluvium where the waters of the Euphrates and the Tigrishad once intermingled. Gradually accumulating there,they could only reinforce the hydrological processes that

had led to the separation of the Tigris and Euphrates inthe first place.

EFFECTS OF HUMAN AGENCIES

Irrigation agriculture has appeared at several points inthis account of topographic succession as little more thana surrogate for natural alluviation. It surely is, in part, ameans by which riverine sediments are distributed evenmore widely and uniformly than by natural floodingalone. Surely also, the processes of levee formation thatare set into motion by canal construction are essentiallythe same as those by which an aggrading stream slowlyelevates its bed with the sediments it carries. Even theformation of spoil banks through the periodic desilting ofcanal beds may be likened to the growth of a natural leveeeffected by a stream as it regularly overtops its banks.And, though canals are generally excavated initially infairly straight lines, they frequently show the same pro-pensity as rivers to develop meandering courses over time.On air photographs or large-scale maps of southern Mes-opotamia, successive reaches of individual watercoursesoften appear to follow a bewildering mixture of "natural"and "artificial" regimes, defying any effort to classify thewhole of a particular watercourse as one or the other.Quite logically, then, the primary terminological distinc-tion made today is one of size, shatt for the handful oflargest rivers and canals, and nahr for the enormous arrayof smaller ones.

This might seem to suggest that the basic topography ofthe alluvial plain owes little to human influence. Perhapsit also seems to imply that the adaptive requirements forsuccessful agriculture are those imposed by relatively con-stant natural processes and constraints, or that humanoccupation has merely reinforced or intensified trendsthat sooner or later would have been manifest anyhow.But none of these possible conclusions is in fact justified.Human agencies do not merely supplement but in parttransform some of the dominant forces and forms of eventhe physical landscape. It is to a description of the distinc-tive effects of these agencies, both in subtly altering thetopography and in setting new demands upon humansettlement and social institutions, that we must nowturn.

In relation to the unstable, continually shifting naturalprocesses that tend to distribute water and sedimentsacross an alluvial landscape, an agricultural civilization isa powerful countervailing force. Growing crops in fields,gardens, and orchards, with palms and other perennials,irrigation canal systems, and storage and transport facili-ties, all are illustrative of means of subsistence and cumu-lative capital investments that are placed in jeopardyby channel movement. Towns and cities must be simi-larly regarded, with social and symbolic incentives play-

19

oi.uchicago.edu

Water, Land, and Life

ing an increasing part in the urge for stability. Enduringstructures of social relations are mapped out in, if notgenerated by, tenure systems and arrangements for access,exploitation, and control of land as the primary produc-tive resource. The effort to maintain those more or lessfigurative "maps" against hostile natural forces accord-ingly is a part of the effort to assure the continuity of thesocial system itself. The response to destructive changesof channel, therefore, normally is persistent, unremittingeffort to reverse them once the flood is spent. To antici-pate and prevent such changes, still other efforts are calledforth to construct bunds and similar control works.

Much but not all of this concern is related to major dis-ruptions in course and flow. Even localized meander cut-ting jeopardizes canal headgates and the orchards thatnormally are sited as near the banks as possible to assurethe water supply. Perhaps more important, watercourseswere the principal arteries for the movement of goods,including not only long-distance commerce but the trans-port of harvested crops into the cities from outlying es-tates and villages. Meandering channels are characteristi-cally highly variable in width and have deep scour poolsalternating with shoals and riffles (Gregory and Walling1973, p. 247). These conditions seriously impede riverinemovements by boat and barge and make the tasks of tow-men slower, more onerous, and much more costly. Forsuch reasons, then, efforts were directed not merelytoward diking and maintaining but also toward straight-ening the major watercourses. Figure 32, showing theconformation of a late third millennium channel, illus-trates the apparent outcome in one region south of ancientAdab.

Several effects of these activities can be traced, apartfrom merely the straightening and stabilizing functionsthat were intended. The natural system of iiterlaced, an-astomising networks of channels that characterized thecentral lower plains involved local fluctuations in flowthat made water management difficult. Over time, irriga-tors sought to replace this with a more canalized, branch-ing, dendritic system. Major levees were kept almostindefinitely in use, both because the traditional towns andcities were situated along them and because canals alongtheir crests maximized command of the land along bothbackslopes for irrigation purposes. As they took on anincreasingly artificial character, levees increased in heightrelative to their backslopes and adjacent depressions, andalong the line of their summits there were increasingly tobe found the great artificial ridges or spoil banks that hadto be removed more or less annually for the canals to con-tinue to function. Long-continued reliance on the immedi-ate hinterlands of individual towns for the agriculture tosustain them, particularly if population growth edged theperimeter of cultivation down the levee backslopes andover the depression margins, unavoidably increased thedangers of salinization. So also did the extension, modifi-

cation, and renewal of the tails of the dendritic canal sys-tems on which irrigation increasingly came to depend,since the levees and spoil banks associated with thesesmaller branches interrupted natural patterns of drainageand encouraged a rise in saline groundwater levels(Buringh 1960, pp. 153-54).

This is not to imply that salinization is to be under-stood exclusively as an unintended by-product of humanagricultural activity. As I noted earlier, it is endemic onsemiarid, subtropical alluviums where high evaporationand slow drainage gradually concentrate even the verylow salt levels that are present in rivers like the Tigris andEuphrates. Traditional Mesopotamian agricultural prac-tices, moreover, are often exquisitely adapted to confiningsalinity within margins that permit continuing agriculturalproduction (Buringh 1960, pp. 249-52). But the mecha-nism of a fluctuating, saline groundwater table that threat-ened the roots of the crops went unrecognized until re-cently, so that the lack of systematic attention to drainageas a part of traditional agricultural methods surely in-tensified an already existing problem. The subsistencefarmer, in addition, is typically caught between short-term uncertainties as to the adequacy of irrigation sup-plies to assure this year's crop and ultimate salinization ofhis land as the long-term consequence of overwatering(Adams 1975d, pp. 3-4). This is a classic bind in whichoverapplication of water, though hastening and intensify-ing the effects of salinity, all too often has been the onlystrategy that could be considered.

Salinization is in some respects an archetype of theconsequences of human intervention mentioned previ-ously. Short-term stability and security were sought-instream flow, in settlement location, in movement of com-modities, in basic agricultural output. Attaining at leastsome of these objectives may have been an indispensableprecondition for urban growth and for devoting increas-ing surpluses to the aesthetic achievements and social dif-ferentiation and complexity that we associate with civili-zation. But there were attendant costs that were not atonce apparent: mounting ecological fragility, reductionsin the productivity of agricultural labor, and a perilouslyincreasing dependence upon labor and capital inputs formaintaining the wider system of watercourses that couldno longer be mobilized locally. The successful contain-ment of small floods, for example, increases the dangerfrom larger ones that cannot be contained. With growingdependence on external resources, prospects for a quickand effective response to a variety of crises are diminishedif the larger sociopolitical unit is also threatened. Generalelevation of groundwater levels in a region, even if for atime farmers manage to adapt to it, increases the likeli-hood of an ultimate, general agricultural collapse. In avery real ecological sense, therefore, the seeds of abandon-ment lie in the development of more large-scale, labor-and capital-intensive, ostensibly better "stabilized" forms

20

oi.uchicago.edu

Water, Land, and Life

of land use out of more flexible, extensive, seemingly"primitive" ones.

Any description in broad terms like these runs the riskof overgeneralizing. We are considering a five-thousand-year span of agricultural civilization in a substantial re-gion, and the record is subject to important temporalfluctuations as well as to long-term, cumulative trends indevelopment. It was also locally diverse, beneath a patinaof uniformity that has largely been imposed by a pro-longed interval of aridity since the latest extensive aban-donment. This geographical diversity both qualifies andenriches what has been said.

The most obvious form of local variability involves themajor watercourses we have been dealing with. Perusalof the maps showing patterns of settlement along themduring successive periods makes clear that in most periodsthere are some striking discontinuities. As perhaps is bestillustrated by the virtual hiatus in settlement in much ofthe area between ancient Nippur, Isin, and Shuruppak,some of these discontinuities remained in place more orless permanently even though textual sources support theinference from the maps that waterborne connectionswere maintained between the surrounding towns. Associ-ated with these hiatuses, moreover, are well-developedmeander-cutting patterns on the air photographs, suggest-ing that the forms of channel straightening and mainten-ance mentioned above were for some reason seldom at-tempted here.

This curious and unanticipated pattern of selectiveneglect is at first glance difficult to explain. Distinctivetopographic or pedological features that would have pre-cluded settlement are not apparent, and at least the in-stance just cited includes considerable areas of moderncultivation that argue against the existence of such fea-tures. Nor is there any apparent reason for more rapidalluviation here that would have buried older sites beneatha heavy overburden of recent sediments. Perhaps the hi-atuses can best be accounted for not as neglected areas inthe usual sense but as necessary parts of a larger scheme ofland management and irrigation maintenance. Continu-ous diking and channeling of streams may have beeninadvisable, on this view, since no flexible response or"safety valve" then would have been available in the eventthat destructive floods, or even merely excess water, beganto move through the system. Accordingly, certain regionswere set aside as planned, seasonally filled depressions torelieve the pressure of high water in the spring. Such apractice is attested for the time of Hammurabi, it has beenargued (Klengel 1976, pp. 130-31), and there is no reasonto believe it was limited to the Old Babylonian period. In-tentional release of floodwaters of course would serveother purposes as well, such as to impede siege operationsby an invading army. Worth noting, although probablynot directly intended, is the effect such flooding wouldhave on stimulating the growth of natural plant and ani-

mal resources. Tending to confirm the existence of exten-sive (or even partly permanent) swamps around ancientIsin, for example, is the relatively high frequency withwhich the bones of water birds are found to occur in occu-pational debris (and hence presumably in the diet) at thatsite (Hrouda 1977, p. 147). By all odds most important,however, must have been the employment of selectiveflooding to encourage the growth of natural fodder forthe herds in uninhabited areas.

Having introduced the theme of the complementarityof cultivation and grazing patterns on a local scale, wemust note its much more comprehensive application to theregion as a whole. The primary settlements of at least thethird millennium form a relatively dense and narrow rib-bon along a series of parallel water arteries, rather thanbeing dispersed fairly uniformly across the landscape.That is understandable, as we have seen, in terms of thegreater ease of irrigation and channel maintenance it per-mitted. It may also reflect attempts to impose a moredurable, rationalized, and hierarchical social order, sincesuzerainty could better be maintained over dependencies,and taxes and other forms of surplus could better be ex-tracted from them, if they were arranged along a limitednumber of navigable channels. But in any case the patternleaves open vast areas in which cultivation as well as set-tlement presumably must have been both much less in-tensive and less permanent. This appears to have beenso along the right bank of the major series of channels,to the west of Nippur, Isin, Shuruppak, and Uruk, untilthe very end of the third millennium. It was still morestrikingly so to the north well into the first millennium,where a very wide area extending from the abandonedchannel of the fourth millennium all the way to the rightbank of the modern Tigris and probably beyond seemsto have been devoid of permanent towns or even villages.

The picture of Mesopotamian land use and settlementthat emerges for at least the older periods is thus quitedifferent from what is usually visualized. Most of thesedentary population, and virtually all of the urban devel-opment, was confined to a relatively narrow, fairly inten-sively cultivated green ribbon or tube down the center ofthe alluvium. Long-continued use of several parallel wa-tercourses running through this tube had straightened andelevated their accompanying levees, improving conditionsfor irrigation along their backslopes and providing a kindof drainage away from the major concentrations of fieldsthat would have alleviated salinization. To the sides, inboth directions, lay very large regions in which muchmore "pristine" conditions obtained: uncontrolled runoff,seasonally filled depressions alternating with unwateredareas, swamp, semiarid steppe. Here conditions favoredextractive activities like cutting firewood, hunting, dig-ging clay for pottery, and perhaps collecting reeds. Here,preeminently, lay zones suited for specialized grazing.Later we will need to examine in more detail the degree to

21

oi.uchicago.edu

Water, Land, and Life

which the available evidence supports this reconstruction,but it at least clarifies and synthesizes a number of unre-lated characteristics of the sequence of maps of irrigationand settlement patterns.

The culmination of our sequence, at least in terms ofthe maximization of human influence on the character ofthe floodplain, saw the final transcending of the olderboundary between the green tube and the steppe. In earlymedieval times large-scale canalization was extended tocover almost the whole of the alluvium, supporting ruralpopulation densities probably greater than those of today.Urban population, too, reached unprecedented levels.Scale, however, is not to be equated with long-term ad-vantage or even viability. For the urban as well as lessersettlements, a largely convex rank-size distribution (seebelow, p. 183) argues for limited political and economicintegration. Similarly, even in its greatest extension, theaccompanying system of irrigation displays the weaknessof the foundation on which it was built.

Its construction, for example, seems to have primarilyinvolved the replication of modular units of moderate size,rather than the development of new forms of integrationreflecting durable, genuinely centralized control. Simi-larly, the bulk of settlement, and presumably populationas well, continued to hug the old central levees ratherthan moving out into the newly opened areas. Probablythis not only implies attachment to the older towns butindicates that many of the areas served by the newly ex-panded irrigation system were characterized by inferior,poorly drained soils, not suitable for the intensively culti-vated, regularly manured summer cash crops that nor-mally were planted close to the towns. In other words, theexpansion in gross output came at the expense of decliningproductivity of land and labor. Almost certainly it alsoinvolved a decline both in long-term capacity to surviveserious environmental perturbations and even in short-term economic well-being. A large-scale, increasingly arti-ficial canal system under the aegis of the state also couldonly be introduced together with a corresponding reduc-tion in lands formerly devoted to grazing. This implies aloss of some of the resilience always represented by largeflocks, held in reserve as an alternative subsistence re-source. Finally, the emphasis on an enlarged, more inte-grated canal system increased the scale and complexity ofits routine maintenance-let alone the problems of at-tending to its repair after any severe damage or disruption.Tasks that had been within the capacities of local com-munities were so no longer, so that even the smaller, morerural components of the system became more and moredependent on the effective functioning of an inherently un-stable and politically vulnerable imperial bureaucracy.

These structural weaknesses, though of growing im-portance and perhaps ultimately decisive in the generalcollapse that accompanied the later 'Abbasid period,should not be overstated in their initial impact. For at

least a few centuries in late Sasanian times, and possiblyagain in the Early Islamic period, the magnitude of irri-gation activities altogether dwarfed all other forms ofland use and may have come close to realizing the full po-tential of the region. Hence the more immediate conse-quence was an unprecedented prosperity, reflected in thescale of urban construction as much as in the enlargementof the canal system itself. The capacity to design and builda system linking the Tigris and Euphrates in order to meetagricultural requirements that could not be met by eitherriver alone was, it must be stressed, a historic achievementneither matched nor superseded before modern times.

Returning to the unintended consequences of prolongedsettlement and canalization, we must mention desertifica-tion. The "natural" pattern of the Euphrates would beto spill over into backswamps and depressions, to altercourse, perhaps not to water all areas uniformly (certainlynot in any short period of time), but to create a mosaic inwhich many local areas of periodic flooding and sedimen-tation lay alongside other small areas that were tempo-rarily unwatered. After abandonments of the kind men-tioned above, however, the Euphrates and its adjoiningperimeter of cultivation moved far to the west. Only rare,great floods now would carry water inland. A vast, moreor less permanently dry region now was exposed to winderosion. The crests of the spoil banks and levees of formercanals, most exposed of all, had been flocculated by saltsthat had accumulated there through capillary action, be-coming "puffed solonchak" soils of very loose, soft struc-ture that were easily carried away by the wind (Buringh1960, pp. 89, 161).\hese wind-borne particles in turn be-gan to abrade away others, initiating a process of massivewind erosion and the consequent development of exten-sive dune fields as one of the dominant landforms of theregion. And dune fields, in their turn, tend to make theprocess irreversible by seriously deterring the reopening ofthe region to agriculture.

FEATURES OF A PREMODERNEUPHRATES LANDSCAPE

As we have seen, the topography of the floodplain as awhole has been shaped by intersecting natural and humanforces. Partly complementing and partly offsetting oneanother, they account for an ordered sequence of west-ward riverine movement, for intercalated networks oflevees and depressions, and for shifting zones of settle-ment and cultivation. Seen from a distance, or over pro-gressively more inclusive intervals of time, the dominantimpression is one of broad, systemic change.

These more detached, inclusive perspectives are pre-cisely the ones imposed by most of our data. The findingsof archaeological survey cannot (at least with the tech-niques employed here) be differentiated into units shorterthan several human life spans. Frequently, therefore, the

22

oi.uchicago.edu

Water, Land, and Life

settlement patterns we detect may be conflations of shortercycles of occupation and abandonment, as well as beingnaturally more incomplete than if we could encounterancient villages and towns as living entities. Even thetextual sources, precisely dated though they occasionallymay be, generally view the agricultural regime from themore distant perspective of the urban scribe or absenteeestate administrator whose problems are those of account-ing for flocks, crops, and agricultural laborers in aggre-gated units. Idealized versions of the calendar round ofagricultural activities do exist (Jacobsen 1958, pp. 55-58;Salonen 1968, pp. 202-12; El-Samarraie 1972, pp. 64-71),but they are abstract, static statements that take no ac-count of real patterns of individual perception, choice, andbehavior in a forest of uncertainties.

Yet as we descend through levels of inclusivenesstoward the individual community at a particular point intime, at least some of the aspects of structural change thatpredominate in a larger view quickly diminish. Ceaselesschange of a different kind in the destinies of individualsand groups naturally continues, but within enduring pa-rameters of behavior and ranges of expected outcome.Behind a screen of intense local variation, continuity froma very remote past probably governs not only the seasonalsubsistence cycle but many other social and economic re-lationships that are closely tied to the circumscribed hori-zon of a community's own agricultural activities. Whatwould be desirable, therefore, is to supplement the fore-going account of long-term ecological changes at theregional level with an impression of more or less momen-tarily coexisting interrelationships in a much narrower or-bit. Only in very recent years, with the growing penetra-tion of modern health, education, communications, andtransport facilities and with the provision of mechanicaland chemical aids to agriculture itself, has all this changedirreversibly. Hence the microcosmic view that is needed-or a series of such views, if the data permit-must predatethe modern era.

Consistent with the primary focus of this study on set-tlement and irrigation patterns, two contrasting docu-ments may exemplify these supplemental impressions. Atthe beginning of this century the German expedition toFara, ancient Shuruppak, prepared the first, a detailed ifsomewhat imprecise sketch map of their environs (Andrae1903). Recast onto the base map prepared from the airphotographs of the region, their findings provide at leasta glimpse of the traditional rural landscape within theradius of less than a day's foot journey. 9 Still somewhatgeneralized in that field boundaries, ownership, frequencyof cultivation and similar details are not provided, figure 4nonetheless brings us considerably nearer to the reality ofa rural settlement pattern than is possible from the find-ings of archaeological survey or cuneiform texts alone.

Perhaps the dominant impression to be drawn fromthis map is that at least this tribally organized group of

agriculturalists had a comparatively slight effect uponthe control of water and land, its primary factors of pro-duction. The parent watercourses seem to have beenallowed to maintain essentially natural, uncontrolled re-gimes, with irrigation depending upon flooding behindsmall earthen dams and distributary canals so small andinfrequent that they were not even illustrated. Insequentchannels varied greatly in width and depth as they woundfrom depression to depression, sometimes being lost tosight altogether as they passed through swamps. Regularlayouts of fields were distinguished by their absence. Theselection of lands for cultivation depended upon casuallyopportunistic considerations of ready availability ofwater, so that labor inputs were kept absolutely minimal.Clearly, this was an extensive rather than intensive systemof land use, at first glance seeming to imply that neitherwater nor land was regarded as being in any way limitedin supply relative to needed crop production or availableagricultural labor (Andrae 1903, pp. 24-26, trans. in Ad-ams and Nissen 1972, pp. 81-82).

Yet we know that the map was made at a time whenthere had been a catastrophic loss of water in this immedi-ate region through the shift of the Euphrates from theHilla to the Hindiya channel, so that widespread aban-donments had already occurred during the immediatelypreceding years. Hence a somewhat different interpreta-tion must be substituted, to avoid the distortions of anassumed equilibrium of factors governing agriculturalpractices. Deep and prevailing uncertainties-as to watersupply, security, and the tax and conscription demandsof a predatory central government-militated against any-thing more than minimal investments in either agriculturalfacilities or settlements. Flexibility of response instead wasthe key to any successful adaptive strategy, tied to themaintenance of large herds as an alternative subsistenceresource and to provision for rapid movement of tribes orother local groups with all their belongings (Adams 1975d,1978). Seen in this light, the relative superficiality of themodes of land use that this map documents is not an ata-vistic feature, but rather one dictated by conditions thatmust have recurred frequently throughout the historicalsequence.

We cannot assume, however, that the picture affordedus in a particular instance may be extended to the coun-tryside as a whole in all periods. Andrae contrasts themassive canal banks near Babylon with their virtual ab-sence here, and there is every reason to believe that agri-culture in the hinterlands of ancient cities was generallymuch more capital- and labor-intensive-as indeed it stillis today. Part of the difference, to be sure, lies only in thefact that in such settings cultivation may have continueduninterrupted for centuries; spoil banks built up over longperiods convey a misleading impression of the laborneeded at any one time for canal construction and main-tenance. But urban continuity, as well as the nearness of

23

oi.uchicago.edu

'Ibra

S • ,, ' . ' . , . ..... & 1414

' * ' . * \ ~ 1459

- M•rd .. "•Ab Hatab 1

/..-.06 . . .. ..... "", , . . ./ 41394

" ." ,„ . . ... ,,,, ,4 ', ,M, Bism ay.

0.2'. ' ..

. .. ,, .. . ... . .old, 1356 3ell9lle001031

0 1 . H VILLAGE 045... . ... . .. / ... . 01 084

ill's AA 14059

L E .. . . . ,,,,,, T "O Dl 07

(AFTE W. ANRE 1903 C P . ' .M D'IFIE

:: . ~/ : :- * ... . ,,,, * \ .......igo#

A'3,, * :" : *:.4 2. . ...... . ...... " . ..... ,." : *\ *- .* " b. :,,,, . .-. ' . . . . ... . ... . .. .

, \: i i\ ' :, : .: ..... *" ._ .... .. .......

• • "ARSH (P R' SEAS NAL * " - -" : .. " . " - " .*-. . . . . 00 ,, .... ',,.

S..E MOUN 05... b u ...0, . " "" 0 4. .... ~

MAAFTER SH AR A014L • ' "R "0 C G MO IFED

wKha,.zi..ya - 013A

1394 -el. l ,-, • ."

a,...20- ......• 74 ,,13,, - .8...

PER IODICAL LY CULTI VATED MODIFIED)

(AFTER W. ANDRAE, 1903; CARTOGRAPHY MODIFIED)

Fig. 4. Settlement and cultivation in the Fara region, ca. 1900.

24

oi.uchicago.edu

lmdmý al K 'zi Bag

Jerachhlya\ ... Salman Pak

Chukr, Tak KisraaKh•a d Khan .. s Nshirvan C•',G,,

Khan Iskander Ja'afir Tai ,Gal'a Shi

Zuba'id Qala

sayyb - irzakchi

hm mmar Shaykh oro 'akatl-Tdjiye

HasswTya Ka hte,

",al'a

,;Dover

Gaa Khnhawi .Dover tents

al-HiIIoh

'Abdullahibn Ali

Qoal -a,lnan-ye

SMohammed, ar aj

.4'.

a , rod a -l ,h. Mehdi

., -h o ,my•' i -5 : .

G 'l •hiy ;' " a

,'Nuwwdm

Gala tC) .Divaniya a uf mm ra

SRabi'a Q iSavhiF QGalraSCid : .QaOd 0Gala Lam,

'Atik I"A Qal'a Lam-a1 a

0 JedidHamzaMus

al-Hayy

lmaimSad Moajib 's

•7"-'-4 · .. "

QGa \ ooK /SSutan

Im Bender al-Wasit

GNusoyri Ekrem

SayyidNoam .' tentsibn H.akim

Immo"Mehdi :

Samawa U S H 'Humayda tribet" GQal'a ShaykhtHumayda Q

ao-Shatra former fields ofala Ayn at -Sayd 'Atik ° Muntafiq shaykhl b Maqamledid Hudhr

o 10 OTTOMAN HOURS d

aI Arja 5d

Fig. 5. An Ottoman view of the plain below Bahgdad, A.D. 1848.

25

oi.uchicago.edu

Water, Land, and Life

an assured urban market, surely provided incentives togradually undertake many improvements like ditching,manuring, land-leveling and field-bordering that wouldenhance productivity. Valuable as is the picture recon-structed for the turn of the century around Fara, there-fore, the extent of its applicability needs careful qualifica-tion. In brief, it is probably most representative of situa-tions remote from cities, and of periods of political fluxrather than stability.

The prevailing absence of security at the time of An-drae's work is emphasized by the ubiquity of small mud-walled fortifications. Most are found along the more sub-stantial watercourses, not infrequently accompanyingsmall weirs and canal offtakes that they were intended toprotect (cf. Adams and Nissen 1972, pp. 75-76). At leastsome, however, apparently were near the enclosing fron-tier of the steppe and may have been intended to supplyimmediate protection for herdsmen and their flocksagainst sudden, short-lived seminomadic incursions. Itwas observed during the survey that the ruins of thesesmall defensive works were almost invariably accompa-nied by habitation mounds implying small adjoining set-tlements. Andrae fails to record most of these, perhapsbecause they had already been dismantled when watershortages first became serious, whereas the little forts andtowers remained for him to see, since they could not becarried away.10 No doubt these fortifications were in-tended primarily to meet conditions of endemic pettyraiding, with refuge from more serious attacks beingsought in flight. Again this suggests that we are dealinghere with an acephalous rural pattern, perhaps occasion-ally capable of providing a loose defense in depth againsturban encroachments, but in the main flourishing only intheir absence or remoteness.

Around the major ancient towns, we may speculate, theonly fortifications allowed would have been intended pri-marily for the protection of officials and travelers. Assuch, they would have been centrally maintained andmore impressive as well as much less numerous. But it isto be regretted that premodern maps of urban hinter-lands with a comparable recording of detail are not avail-

able. To the urban-based official or traveler it was thepermanent towns, caravanserais, and interconnectingroutes that were important, not the volatile dispositionof a restless tribalized peasantry over the countryside.Hence the contrasts between more stable, densely settledurban peripheries and genuinely rural areas- or perhapsbetween well-ordered kingdoms and interludes of prevail-ing civil strife-that were suggested above cannot be es-tablished. Incompletely representative though it may be,the map of the Fara environs remains our most substan-tial clue to the convergence of natural and human agen-cies within a local setting of the kind that constituted thetraditional Mesopotamian landscape.

Looking beyond that closely circumscribed setting,however, it may be useful to compare the reality of thepremodern countryside with an urban perception of itembodying all the systematic biases just mentioned. Fig-ure 5, redrawn from an Ottoman War College maproughly half a century earlier than Andrae's work, reflectsa concern only with the garrisoned strings of towns alongthe principal watercourses." Even the rivers themselvesare only impressionistically mapped, and the canals andsubsidiary streams upon which the life of the populationdepended were clearly of no interest to the cartographer.The entire populous district of which Andrae's map showsonly a small part falls in the vast tract between the tworivers and hence was also outside his purview.

The Ottoman cartographer's conception of the land,I suggest, was typical of that of most of its historical ad-ministrators and literati. The reality always has been thatcity and countryside form a fluid continuum, an interac-tion or systemic interdependence that this study seeks totrace out in greater detail. But cognition mediates reality,sometimes simply excising patterns of life when they can-not be fully controlled or even comprehended. This testi-fies to an enduring feature of Mesopotamian urban con-sciousness, a pattern of selective perception that tincturesmost of its prodigous outpouring of written records. Theonly available corrective, it can plausibly be argued, is onethat the archaeologist seeks to supply through reconnais-sance and excavations in the countryside.

26

oi.uchicago.edu

2The Recovery

of Ancient Settlementand Irrigation Patterns

The foregoing chapter provides a very generalized histori-cal overview of the morphogenesis of a landscape, inwhich the Euphrates River played a basic and initiating, ifnot always immediately decisive, part. The cumulativerecord is one of broadly increasing but uneven human in-teractions with natural forces and constraints, ultimatelyleading to a profound modification of the lower alluvialplain away from its "pristine" condition. Interrelation-ships among human groups were of course also increasingin scale and complexity and shifting in character, but a de-velopmental account of human institutions is within ourpurview here only insofar as it was importantly mediatedby factors of land and water.

An archaeological survey supplies the connective fabricand body of comparable data from succeeding periods onwhich this account ultimately depends. Surface recon-naissance alone, without the enrichment of historical andethnographic insights or the modifications and improve-ments in precision that would be made possible only bya closely associated program of excavations, furnishes atbest an exceedingly limited framework of interpretation.But its complementary virtue is that it helps to overcomesome of the biases and accidents of discovery that suffusemost bodies of ancient textual and archaeological datapertaining to specific periods. It permits us to perceivegross differences and similarities that span millennia, andto seek explanations for problems of long-term ecologicalinteraction that our ancient informants either never con-sidered or dismissed as commonplace and hence not worthrecording.

This is not the place for a critique and rationale ofarchaeological settlement survey as it has been elaboratedin a variety of other regional settings (cf. Parsons 1972;Johnson 1977). Ancient settlements, like most other relicsof past human behavior, reflect social decisions made un-der a variety of inducements and constraints. Some aspects

of their distribution and internal structure may be of alargely cultural character: they may reflect an ethnic orlinguistic boundary, persist in a particular place becauseof its religious associations, or maintain this or that formof urban hierarchy in response to this or that set of ad-ministrative, commercial, or other considerations. Manyfeatures of settlement are responsive at least in part toconditions of scarcity: of water, of building materials, oftransport routes and facilities. Still others, probably thegreat majority, articulate these responses to scarcity withhistorically derived perceptions and potentials: the extentof functional differentiation and integration that a par-ticular society encourages and will sustain, its level oftechnological proficiency, its readiness to cluster or ag-glomerate, and its subsistence choices or imperatives.There is nothing inclusive about this listing. The point isonly that all human settlement is patterned in many com-plex ways, the unraveling of which can provide insightsinto social change and stability by no means limited toman's spatial disposal over the landscape. The refinementand proliferation of settlement pattern studies thereforereflects the widespread recognition that it is a legitimategoal of archaeologists and anthropologists-and, a forti-ori, of geographers-to make those patterns intelligible.

A number of the presuppositions and avenues of ap-proach that are essential for archaeological survey insouthern Mesopotamia have been set down in a previousstudy of a generally similar kind (Adams 1965, AppendixA), and they will not be repeated in detail here. The mostbasic principle is that settlement accompanies agriculture,and that both are dependent upon assured supplies of irri-gation water. The vestiges of ancient settlement survivefor the archaeologist to discover in the form of mounds ortells, built up layer by layer out of decomposed mud brickand other architectural and living debris. Where tells fallinto a linear pattern, as they generally do, a kind of least

27

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

effort principle allows us to hypothesize that the line isapproximately that taken by the watercourse serving thosesettlements during their floruit. With the addition of airphotographs, the individual watercourses, and even super-imposed sequences of watercourses dating back to fairlyremote periods, can often be seen. Then, adding a finalmethodological element, the superimposed occupations ofthe tells can also be detected from the ceramics and otherdebris profusely littering their surfaces. Mounds and wa-tercourses thus can be brought together into an inte-grated, changing system or sequence of systems, the majorepochs tied to ceramic and other chronologies that neces-sarily depend in large part on excavations.

Such is the rationale of the general approach. Evenmore important, however, is the mode of its detailed ap-plication in concrete, varied circumstances. It is to thespecial requirements, procedures, opportunities, and lim-itations of the survey itself that we must now turn.

MAPS, AIR PHOTOGRAPHS, AND THEIRINTERPRETATION

The region covered by this account heretofore hasreceived little cartographic attention. Potentially of great-est utility is an Arabic 1:50,000 map series, but the exist-ing sheets are restricted in availability and vary uncon-trollably in quality. They naturally concentrate uponregions of current economic importance within the fron-tier of cultivation, moreover, and virtually ignore thedesert topography beyond it. More immediately useful,therefore, is the British one-quarter-inch (1:253,440) mapseries, perhaps most widely available in a slightly modi-fied reproduction by the United States Army Map Service.This furnishes a detailed (though now badly outdated)guide to cultivated areas, particularly valuable for thenames that are carefully transliterated and copiouslysupplied. But again the extent and reliability of coveragedecline abruptly beyond the frontier of cultivation. Onlya handful of the most prominent landmarks (principallyancient mounds) are located with any precision. Else-where the desert topography either seems to have beensketched in from informants' imprecise and not whollyfactual descriptions-to be rendered in such conven-tionalized form that its relationship to existing featuresis often very difficult to discern or to be dismissed withthe legend "unsurveyed desert," which is at least morehonest. As for the aerial navigation charts available fromvarious United States military services at scales rangingfrom 1:250,000 to 1:1,000,000, they are so generalizedas to be useless for ground reconnaissance in uncultivatedareas.

Archaeological survey, under these conditions, includesa heavily time-consuming but altogether indispensablecomponent of primary mapping. The means at hand, cru-cial in its own right as a source of voluminously detailed

information on subtle differentiations in soil texture,moisture, and surface vegetation that are the relics ofancient watercourses across an alluvial plain, was a setof KLM aerial photographs at an approximate scale of1:35,000, made temporarily available for my use by theDirectorate General of Antiquities after being supplied bythe Directorate General of Surveys. No mosaic composedof these photographs was available, however, nor couldadequate ground controls be introduced by my own lim-ited mapping efforts to compensate for the inevitable,generally small distortions arising from variations incamera elevation and perpendicularity.

My procedure in mapping was to tape together amosaic consisting of a dozen or more individual KLMphotographs, anticipating that the matching of differentparallel runs would help average out distortions in in-dividual photographs and runs. After a quick tracing ofa few salient features, the mosaic then was disassembledand the individual photographs studied in greater detail.Prolonged examination was always necessary, undermany different conditions of lighting. Gradually a pat-tern of linear discolorations emerged, generally consist-ing of the faint traces of ancient levees, to supplementmore contemporary and obvious features like outlines ofseasonally swampy depressions and fields of dunes. High-lighted with a marking pen, these faint lines could thenbe added to the mosaic tracing. At some point a provi-sional map of sufficient area was ready to serve as a basisfor fieldwork.

The field survey was largely concerned with ancientsettlements rather than with the intricate overlay oftracings of canals and rivers of various periods that con-stituted the dominant evidence on the air photographs.Most, though by no means all, archaeological sites wereassociated with at least a slight rise in elevation over thesurrounding plain surface, and any perceptible eminenceroutinely became the object of a visit. Perhaps half of thesites that were located had been previously identified ten-tatively on the air photographs as suggestive discolora-tions, either lighter or darker than surrounding areas de-pending on complex conditions of salinity, moisture, andtime of day when the photograph was taken. Differences insurface texture also proved suggestive, low mounds oftenshowing up as more uniform or "smoother" areas, perhapsbecause of some reflective property of the sherds litteringtheir surfaces. But, while most sites could be fairly un-ambiguously identified once their location had beenpinpointed accurately, it must be stressed that the photo-graphs were in no sense an infallible guide to the archae-ological sites. Hundreds of suspicious discolorations werevisited in vain, often with considerable difficulty in reach-ing and exactly locating them on the ground, while in afew other cases fairly large and prominent sites werefound to which the photographs provided no clue evenwhen closely examined with a strong hand lens.

28

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

Thus the matching of the archaeological sites to theprovisional mapping of ancient river and canal traceswas a pervasive, demanding problem. In a sense, twomaps had to be elaborated simultaneously on a singlebase, of which only one, the watercourses, was essentiallydependent on the aerial photographs from the outset.The archaeological mapping had to have as its startingpoint one or more major mounds, canal junctions, orother features that could be unambiguously located ona photograph. From there it was carried forward by tri-angulation with a prismatic compass (multiple readingspermitting something on the order of V/2 o accuracy)and measured odometer distances, transferred to theprovisional base map with a three-arm vernier pro-tractor.

Periodic opportunities to bring these two semi-inde-pendent mapping systems into correspondence were ob-viously vital. Sometimes they were provided by additionalcanal junctions or similar features visible on the photo-graphs, from which further sightings could be taken onmounds already provisionally located. Even more usefulin this respect were the ruins of small mud-walled watch-towers and enclosures, the remains of nineteenth centuryand earlier settlement in areas now wholly abandoned,that occur on the summits of many mounds in the region.Study of the approximate area on a photograph wouldcontinue until, with the aid of a pocket microscope ifnecessary, a structure as small as 4-5 meters in diameterhad been clearly identified at the 1:35,000 scale. With anew position thus fixed, the locations of sites found onearlier traverses could then be reconfirmed and if neces-sary corrected.

The need for major, continuing corrections in a processof this kind must be recognized as an essential conditionfor accuracy. Simple errors in measurement account forpart of the difficulty, but far more common and seriouswere misidentifications of the mounds and other features-low, lacking distinctive form, frequently mirage-riddenand obscured by dunes and blowing sand-upon whichtriangulation sights were taken. To reduce these errors,I frequently built temporary cairns out of bricks and othersurface debris on accurately located mounds that mightlater be useful for bearings. But the guiding principle inboth detecting and overcoming them was to rely onredundancy of measurement, recording as many bear-ings as possible on mounds in all directions rather thanmerely two as a requisite minimum.

The linear traces of ancient watercourses also wereadded to as the mapping of archaeological sites pro-ceeded. More detailed examination of the air photo-graphs, in connection with locating particular sites, addedsome. Others were first observed on the ground and onlylater found on the photographs as well. Still othersemerged from more problem-oriented study, as unfold-ing awareness of. the pattern formed by sites occupied

during a particular period led to hypotheses about wherethe associated canal routes should be sought.

A penultimate stage in the mapping involved calibrat-ing latitude and longitude intervals from know land-marks in the air photographs, principally along the Tigrisand Shatt al-Hilla courses. At this juncture certain dis-tortions became evident, arising from the use of individualaerial photographs as a surrogate for a map base. Mapsections that had been drawn from temporary mosaicsof photographs, progressively extended northward fromthe Euphrates near Warka, had to be joined with thosedeveloped from air photographs of the Tigris coursewest of Kut. Cumulative discrepancies over this distanceof about 140 kilometers, before partial correction byaveraging errors and redrafting, varied from 400 metersto about 1.4 kilometers; it is reasonable to suppose thatmost or all have now been reduced to less than 1 kilo-meter.

Finally, the draft of the base map prepared in the fieldwas photographically reduced and completely redrafted.At this time an effort was made to apply consistent stan-dards and symbols of representation, and the originalsite-numbering system that followed the order visitedwas replaced by a system based on contiguity that per-mitted easier reference. It is regrettable that the air photo-graphs were not available to me for continued use as theanalysis of the survey's findings proceeded after com-pletion of the base map. The discovery of long linesand consistent patterns of contemporary settlements, dis-played herein in a sequence of maps of sites occupied insuccessive historic periods, argues forcefully that certainapparent lacunae might be filled by more closely examin-ing the photographs for those areas. But any iterativeprocess of closer approximation and improvement ofdetail ultimately reaches a point of diminishing returns.It seems rather doubtful that the additions that might bemade by further restudy of the air photographs-usingonly these methods, at least-would be significant ones.

There are several features of the base map, conventionssystematically followed in its preparation, that one shouldbear in mind. Differing fundamentally from air photo-graphs, any map is a system of symbols, a selection ofcertain aspects of reality for representation at the expenseof other aspects. This map, in the first place, seeks tominimize inclusions that are of modern origin. No indi-cation is given of roads and communications, even themost important trunk routes. A number of the more im-portant contemporary towns of the region are conven-tionally shown for convenience of reference, but noattempt is made to give the names and locations of thelarge number of smaller settlements. The major riversare in their modern position, or at any rate in the bedsthey occupied at the time the air photographs were takenin 1961-62. Modern canals are not shown, however, tothe extent that they can be differentiated from earlier

29

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

ones through the evidence on the aerial photographsalone.

This last-mentioned convention obviously introducesa degree of ambiguity, since I will argue later that earliercanal levees generally have been selected as the optimalcourses to be followed in the design of subsequent irriga-tion networks. But, fortunately for our purposes, com-prehensive modern systems have only begun to be intro-duced in the last few decades. And the traditional system,for a long time haphazardly maintained and more re-cently subject to small-scale, eclectic extensions, fails toobscure at least its larger-scale antecedents of the Sasan-ian and Early Islamic periods. The latter differ not onlyin size but in straightness and in regularly branchingelements of an overall design. These features in mostcases can be identified beneath the irregular and discon-tinuous segments of later centuries and are the only onesto appear on the map within the zone of cultivation.Beyond the limits of cultivation, on the other hand, es-sentially all traces of former canals and meandering, pre-sumably more "natural" watercourses that are visibleon the air photographs have been recorded on the mapas of relatively ancient origin.

Clearly, the air photographs permit a recording offormer watercourses in the desert areas much superiorto that in the cultivated zone. But there are limitationsand uncertainties to be kept in mind even there. Only afew of the very largest ancient canals can be readily fol-lowed on the photographs as continuous lines. Morecommonly, parts are quite apparent, but most can befollowed only with considerable difficulty, and in a fewplaces the lines cannot be detected at all. To increaseintelligibility, all of these conditions have been renderedon the map by continuous lines of uniform width, insofaras the existing gaps are small enough to leave a reason-able assurance of the continuity of the original canalcourse.

The use of lines of uniform width (except in the caseof a few major arteries like the Shatt al-Nil) obviouslysacrifices information on ancient canal size that wouldbe useful for some purposes. One can imagine studiesconcerned with population density, for example, in whichestimates derived from aggregate areas of settlement arechecked and complemented with estimates of aggregateagricultural output derived from calculations of canalcapacity. But the variation in the clarity of the photo-graphic traces indicates the difficulties that will beset suchan effort, at least until it is undertaken on such a scalethat the photographic evidence is massively supplementedby ground recording. Levee width alone, moreover, isat best a somewhat dubious measure of canal size. Long-continued use of a small canal may lead to the accumu-lation of spoil banks as massive as those from a brieflyused larger one, and changes in slope and cross-sectionalarea can also have important effects on the siltation rate.

An approach to ancient demographic, social, and eco-nomic conditions through a network analysis of flowsin the canal system at a given period, therefore, is verylikely to require an extensive program of levee cross-sectioning.

Still a further difficulty is that elevated levees are ex-posed to wind erosion, particularly as the balance ofsurface material in an area shifts toward fine, uncon-solidated sand. Sharply contoured, substantially elevatedspoil banks of former canals are absent in this entireregion except where they are associated with settlementsthat demonstrably were abandoned within the last cen-tury or so. Hence the mass of remaining levee materialmay be in no way proportional to the original mass.However, this does not mean that levees can simply beblown away without leaving any trace. Probably thereis a sorting process in which heavier particles are leftbehind, for even very low levees that are almost im-perceptible on the ground frequently are associated witha clear line of discoloration on the air photographs. An-other possibility, suggested by a Cassite canal adjoiningsite 1590, is that compacted material in the bed of a canalmay tend to be cemented together by clay and fine siltparticles. In that instance, at any rate, a slightly elevatedwhite ribbon of canal bed deposit remained behind al-though the spoil banks that once must have adjoined ithad completely disappeared.

The largest dune field in the region, extending fromancient Adab to beyond Umma in a southeasterly direc-tion, directly adjoins and largely overlies the traces ofthe largest and most long-lived of the region's ancientwatercourses. It cannot be doubted that the bulk ofthis aeolian material is locally derived, and that in factthese dunes embody in an altered state a considerablepart of the original mass of the river levee. Figure 32,traced directly from the air photographs, illustrates thepatterning of the dunes in relation to the surviving tracesof the watercourses, which in many places have beenetched into low relief by the same wind-borne sand thatelsewhere covers them.

How should we interpret these traces? Generally theyconsist of close-spaced, varvelike patterns parallel tothe direction of flow in the bed. Probably they are to beexplained as relics of alternating intervals of high andlow flow, for such alternations would necessarily beaccompanied by differences in bed-load particle size. Buta more difficult question concerns the dating of the exist-ing surface traces. If a very large proportion of the parentmaterial of the levee has been scoured away, then thevarves should represent a conflation of the beds occupiedby the watercourse during a large part of its existence.In the absence of evidence of successive meander-cutting,however, this is not the appearance given by the existingtraces. Instead, therefore, we are probably witnessingthe bed only as it was situated during the terminal period

30

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

of full-scale flow, covering all or part of the Isin-Larsaperiod. That the bed as shown is not later than the earlyOld Babylonian period is shown by sites 1175 and 1460,both of which date from that period and are positioneddirectly over the watercourse after it must have ceasedto function.

A somewhat different interpretation appears to berequired in the case of the watercourses joining in thevicinity of site 039 and continuing south-southeast tosupply Uruk. The visible traces of a bed here closely ad-join a series of sites (039, 097, 100, 131, 242, etc.) thatcontinued into the Old Babylonian period but that werenot occupied later. On the other hand, site 1555, whichwas apparently already abandoned by the end of theIsin-Larsa period, is not associated with the visible tracesof this or any other watercourse. Here, therefore, thebed that can be followed on the ground and in the airphotographs is apparently Old Babylonian, that is, slightlylater than the larger bed that can be traced southeast ofancient Adab. But the two cases are similar in that onlythe terminal period of use of the watercourse appears tobe represented in its surface traces.

A third instance of a major natural watercourse, re-quiring an interpretation substantially different fromeither of the foregoing, is the complex series of meandersnortheast of Nippur. The upstream traces of this water-course have been first identified northeast of ancientKutha, adjoining the important late prehistoric town ofTell 'Uqair (Buringh.1960, fig. 72). Shortly disappearing,they reappear again near site 670 more than 70 kilometersdownstream, emerging at that point from beneath theheavy overburden of a series of major, much later canallevees that all tend east-northeast. The course is lost toview again under the recent sediments and disturbancesof cultivation associated with the Shatt al-Gharraf, morethan 75 kilometers southeast. Over much of this distanceit can be traced not as a single, slightly sinuous but onthe whole relatively linear bed, as in the two foregoingcases, but rather as a series of alternate courses of con-siderable breadth, variability, and obvious duration.There is evidence, for example, for the formation andpinching off of several meander loops, while elsewhere(near site 1171) a process of lateral meander movementor "sweep" can be followed over a distance of more than6 kilometers. This is a strikingly less controlled, more"natural" river regime than in the other two cases, andit clearly reflects the successive beds associated with thatregime over a much longer period.

The lower end of the observable part of this ancientwatercourse, northeast and east of ancient Adab, appearsto differ in important respects from the remainder of it.Traces of alternate course-cutting and meander migra-tion disappear. There are also several bifurcations inthe channel that appear curiously stable and unaffectedby the processes of movement that are so evident farther

upstream. Interpretation is unfortunately made more dif-ficult by the fact that, immediately below these bifurca-tions, the massive overburden of recent Shatt al-Gharrafsediments obscures whatever further traces there mightbe. But what can be seen is perhaps best explained bythe hypothesis that stream velocity was greatly reducedat this point, and that we are witnessing the remains ofa deltalike series of mouths through which the channelfound its way into a much larger body of water. Anattractive candidate for the latter is a long arm of theArabo-Persian Gulf, extending more than 300 kilometersnorthwest of its present shoreline. But of course it isequally possible that the channel at this point merelydrained into a large inland swamp or series of lagoons.

The dating of this course, as I indicated in the pre-ceding chapter, can be confidently assigned in the mainto the Uruk period. It is most extensively traced out byaccompanying settlements in the Early and Middle sub-phases of that period, seems to have undergone partialabandonment in Late Uruk, and survived into JemdetNasr times only as a few vestiges. Most or all of thatspan is presumably reflected in the sequential develop-ment and then permanent demise of the meanderingwatercourse that can still be traced out on the surface.In this case, it thus appears, either accompanying alluvia-tion was less extensive to begin with or wind erosionof the former levee has been extensive enough to exposeunderlying beds that still remain buried in the formertwo cases. Such wind erosion, if it did occur, probablywas at least partly antecedent to the many later canalsystems whose levees now overlie the bed, their cross-cutting patterns suggesting many alternative directions offlow. This implies that a major cycle of wind erosionmay have occurred before the Third Dynasty of Ur. Sucha reconstruction can only be somewhat speculative atpresent, but it argues plausibly for the first appearance ofdune fields in the region, embodying these wind-erodedlevee sediments, by no later than the mid-third millen-nium. Whether the substance of dunes of that antiquityhas remained permanently unconsolidated since then,while shifting from place as atmospheric circulation ortopographic conditions changed, can only be a matterof conjecture. We have noted, however, that wind deposi-tion of sheetlike deposits tending to merge imperceptiblywith the plain surface is going on simultaneously today.That suggests a shifting balance of wind erosion anddeposition, a process of recycling from cultivable plainto dunes and back again, which may continue in-definitely.

While these three cases are of considerable importanceto the reconstruction of settlement and irrigation patternsin the periods to which they pertain, my purpose in dis-cussing them here has not been to deal with their historicalimpact. They are intended to illustrate the range of topo-graphic information on early watercourses that the air

31

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

photographs and surface reconnaissance make available,opening up rich interpretive possibilities for unexpectedlyearly periods. It must also be clearly recognized, however,that these possibilities are beset with limitations andambiguities that are unlikely to be overcome withoutmuch more detailed study.

As the base map profusely illustrates, ancient canallevees of various periods repeatedly bifurcate, intersect,and cross one another. In many cases the air photographssuggest which canal line is superimposed on, and hence ispresumably later than, the other. But that evidence ofteninvolves a large element of uncertainty. Since canal leveeswere so often reused, moreover, what might be inter-preted as the stratigraphically superimposed, later canalmay be no more than a late reuse of a line that on thewhole precedes the one it apparently crosses.

There is a still more intractable difficulty in trying toillustrate the set of junction points that defines thebranches of an ancient canal system as it functioned atany given time. Such systems were continually subjectedto major as well as minor alterations in an adaptiveprocess that took account of reductions and uncertaintiesin flow as well as many other variables. From a veryearly time, certainly no later than the early third millen-nium B.C., it was a design characteristic throughout thisregion that districts could be supplied with irrigationwater from at least two different trunk canal systemsthat presumably would be unlikely to suffer simulta-neously from impediments in flow. Hence the actualpattern of flow must have been at least in part a matterof local option. To speak of the nodes and branches thatconstituted a canal system "at any given time," there-fore, only begs an unanswerable question. Was the "time"at which we attempt to offer a map representation of asystem the moment it left a Sasanian engineer's draw-ing board? A normal year of operation or a year ofwater shortfall? A summer or a winter growing season?Or even, a particular day in the operation of the system?The answer would quite possibly be very different in allthese instances. Except in a few cases that are completelyunambiguous (again, the Shatt al-Nil provides most ofthese exceptions), all crossed canal lines accordinglyhave been broken at their junctions to indicate uncer-tainty as to the sequence or contemporaneity of the canalsinvolved.

One other modern human artifact appears on the basemap-the frontier of cultivation. The one that is con-tinuously delineated is taken from the air photographsand hence corresponds to the early 1960s. More recentextensions in the cultivation zone have in some cases beensubstantial. This is particularly so to the north and eastof the surveyed region, where the relatively more ade-quate waters of the Tigris and the Shatt al-Gharraf haveencouraged not only government-initiated schemes butalso a rapid lengthening of the tails of existing canals

by local cultivators. Arrows are employed on the mapto suggest the direction and extent of these recent move-ments. No attempt was made during the fieldwork tokeep a continuous record of the new position of the fron-tier, however, so that the map furnishes only a very ap-proximate indication of what has been a significant shiftin the proportions of cultivated and uncultivated landover a relatively limited period.

Shown in stippling on the base map are areas of dunes.In large part these lie outside the cultivation perimeter,although in places farmers have sought to stabilize someof the looser, less actively moving dune groups and toextend field canals and cultivation into their midst. Therehas been no systematic study of dune formations in thisarea, but the dune fields whose outlines have been tracedfrom the air photographs generally consist of symmetrical,lunate forms of barchans, sometimes closely grouped intooscillating, wavelike ridges with alternating barchanoidand linguoid elements (Cooke and Warren 1973, p. 288).In this region individual dunes are generally small, thoughsome can cover a hectare or more and rise to a heightof 7 or 8 meters. It is not the individual, isolated dunethat furnishes the principal obstacle to archaeologicalsurvey, but rather the much larger, dense grouping thatmay extend almost impenetrably over several squarekilometers. Such groupings may completely cover archae-ological sites, and in any case they make it much moredifficult to detect sites. Even with four-wheel-drive ve-hicles, travel is impossible except on rare occasions aftera heavy rain, so that a significant reduction in the effec-tiveness of survey must be assumed within the dune areasthat are shown.

As I have described more fully in a separate sectiondealing with the configurations of a changing landscape,the dune represents only a portion-and probably arelatively small portion-of the aeolian bedforms of theregion. Low hummocks or nebkhas that have formedaround desert shrubs also occur over vast areas, con-tinuing a very rough microtopography within a contourinterval of 50 centimeters or less even after the surfacevegetation has disappeared and the roots have been re-duced to a brittle skeleton lacing the sand together. Leastobvious to the casual observer are very low, sheetlikedeposits, sometimes slightly undulating. They tend tomerge with dune fields or to blend imperceptibly intoirregularly wind-eroded plain surfaces, so that their thick-ness and area are usually very difficult to estimate. Occa-sional gullying after intense local rain squalls sometimesexposes aeolian deposits 10 or more centimeters thickthat rest unconformably on the underlying alluvium,even where no hint of such deposits was apparent fromthe surface. For the moment this serves only to illustratethat aeolian deposits should not be thought of as alwayscoterminous with the dune fields shown as stippled areason the base map. Subsequently, we will need to consider

32

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

further the implications of these deposits for the complete-ness of the record of the survey.

A final mapping exercise was undertaken on a some-what experimental basis. Figure 6, covering essentiallythe whole of ancient Sumer and Akkad, tests the utilityfor archaeological purposes of the Earth Resources Tech-nology Satellite (now LANDSAT) images. These imagesare now readily and inexpensively available, includingmultiband and other image options that offer importantadvantages for specialized interpretive purposes. Thetwo used for this map were merely standard prints, en-larged to the same 1:250,000 scale as United States airnavigation maps. Traces of ancient watercourses shownin the figure are essentially those that can be detectedin the LANDSAT images, while the modern river coursesand certain other details have been taken from othermaps. But the essential point is that the interpretation ofthe images was done independently from the main map-ping based on ordinary aerial photography, so that thetwo provide largely independent analytical statementsabout the layout of main features of ancient canal systems.

Certain limitations of the satellite images emerge atonce. Their resolving power is extremely low, which ofcourse is why (unlike military satellite photographs) thereare no security restrictions on their distribution. It mighteven be said that they are not strictly suited to the delinea-tion of ancient canal lines and other fairly small-scale,precise features, since the minimum unit of ground defini-tion is about 80 meters in diameter. There is a continuousgradation between linear discolorations that are reason-ably unambiguous and a much larger number of othersthat oscillate between appearing genuine and appearingto be possible figments of the imagination. A somewhatsimilar difficulty on occasion characterizes the use ofordinary aerial photographs, to be sure, but only in aminority of cases, with little overall effect on the pre-ponderant mass of substantive detail recorded on thebase map. Here, on the other hand, the great majorityof cases must be classified as more or less doubtful.

Poor resolution perhaps also accounts for the absenceof ancillary indicators (e.g., disused segments alternatingwith contemporary reuse) as to the antiquity of a lineartrace that presumably represents a watercourse levee.Only by comparing the LANDSAT-based map with onedrawn from the other photographs can the modern fea-tures be identified and eliminated. However, some otheromissions are more difficult to explain. Coverage in desertareas is particularly poor. Major ancient levee systemsthat dominate the air photographs of uncultivated areas(e.g., the Shatt al-Nil leading to Uruk, the second Shattal-Nil northeast of Nippur, and the third, classical IslamicShatt al-Nil flowing from south of Kish to Na'amaniyaon the Tigris) are essentially untraceable on the LAND-SAT photographs.

One might wonder, with so many substantial draw-

backs, whether on balance the LANDSAT images canpossibly have any utility for the reconstruction of ancientcanal systems. I think figure 6 argues conclusively thatthey do. To begin with, inadequate representation ofdesert regions is not a problem; those are the regions forwhich the ordinary air photographs have already suppliedan almost indigestibly voluminous record. What theLANDSAT images supply is a very broad overview ofalmost the whole of the alluvial plain, permitting themajor features of canal and watercourse systems in atleast the vast cultivated region to be identified veryquickly. And the essential congruence of that overviewwith the findings of many months of field survey andwork on other maps and air photographs can be con-firmed easily by comparing figure 6 with other maps in-cluded in this study.

The LANDSAT system apparently records some verybroad aggradational features (i.e., levees) at least as wellas ordinary air photographs. Perhaps in the latter thesebroad but diffuse traces tend to be submerged by massesof subsidiary detail. At any rate, they appear as light linesagainst a darker field on the LANDSAT images. Possiblylevee sediments are in fact lighter in color; they may alsoappear lighter only because they are slightly elevated andtherefore drier. But, whatever the explanation, the re-construction of these levees in cultivated areas allowsthe main outlines of the ancient canal systems to be ex-tended far beyond the limits of the surveyed area. Surfacereconnaissance confirms that at least some of the leveesso reconstructed stem from as early as the second mil-lennium B.C. The bulk are Parthian, Sasanian, or evenlater, and their interdiction by the Shatt al-Hilla and thereunified course of the Euphrates below Samawa helpsto confirm the relative recency of the latter.

Other features emerge from examination of theLANDSAT photographs that are of considerable his-torical importance, even though they lie outside the regionthat systematic survey data allow us to deal with in anydetail. Alternative courses of the ancient lower Euphratespast Eridu and Ur can be tentatively identified. A com-plex picture also emerges of river and canal successionsand bifurcations in the region around the Islamic city ofWasit, differing considerably from the usual reconstruc-tion in which the medieval Tigris is said simply to haveadjoined that town. In general, then, this map serves notonly to check and supplement the findings of recon-naissance within the surveyed area and its immediateenvirons, but also to provide at least provisional coverageof a much larger region including the entire settled coreof the ancient alluvium.

BOUNDARIES OF SURVEY COVERAGE

While aeolian deposits furnish an undeniable obstacleto desert survey, the obstacles within the cultivated zone

33

oi.uchicago.edu

' IG -

K

0 10 25KMrIR . I " '

Fig. 6. Major ancient levees identifiable in LANDSAT imagery.

34

-I-

k-/

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

tend to be very much greater. In the first place, ancientcanals and watercourses are much more difficult to tracein areas that already lay within the cultivation perimeterat the time of the first comprehensive aerial photography.Further, the small, low sites that are a major element inthis survey's analytical framework, particularly for under-standing the late pre- and protohistoric periods, are ex-tremely difficult to detect when they occur in cultivatedfields. Not only are they easily hidden by the crop plants(and sometimes to an almost equivalent extent by weedsthat grow rapidly during the fallow cycle), but they aregenerally plowed through by farmers rather than left asslightly elevated, uncultivated islands. The result is thatthe low and subtle but distinctive contour that is createdby the cap of sherds on these sites, often visible for akilometer or more in the open desert, is altogether lostto view. Only by gridding these areas on foot, in traversesnot more than one or two hundred meters apart, couldthe same intensity of survey be achieved as can be car-ried out in the desert with a primarily motorized mode oftravel at grid distances five or even ten times greater.Finally, of course, the presence of canals and cultivationgreatly impedes vehicular access to the margins of anarea where foot reconnaissance is thought to be necessary.

For these reasons the survey tended to stay outside thefrontier of cultivation. As I will presently detail, excep-tions were made to take advantage of more favorableconditions or to avoid important lacunae in coverage ofmajor ancient watercourses and urban hinterlands. Onthe whole, however, it was fortunate that the primaryfocus on courses of the ancient Euphrates happened tocoincide largely with uncultivated areas. As the desertboundary recedes before advancing waves of cultivators,and as the remaining desert land surface is increasinglycrisscrossed with roads and other improvements thatgravely jeopardize its archaeological sites, those areaswere the highest priority of study that could be met withinthe limited time and means available to a single in-vestigator.

The effective boundaries of the surveyed area are sug-gested by the distribution of recorded sites, as well asby the cultivation limits plotted on the base map. Adashed line indicating the outer boundary of reconnais-sance also is given on the maps showing the distributionof known settlements in particular protohistoric andhistoric periods. But a fuller discussion of conditionsalong particular sections of this boundary is needed torelate the survey's findings to the broader pattern of settle-ment distributions within southern Mesopotamia as awhole.

To the northwest, the airea of this reconnaissance ad-joins a region that was briefly and superficially examinedby the author and Dr. Vaughn E. Crawford as part of theAkkad survey in 1956-57. Much of it lay outside theprimary focus of concern at that time, and in fact was

visited primarily to record a site fortuitously exposed bythe excavation of the main drainage canal of the Mussayibirrigation scheme (A 221; cf. Harris ahd Adams 1957).At that time the region was wholly uncultivated, but ithas since been included within the expanding Mussayibproject. I had expected to use the classical Islamic Nilcanal, the ancient bed of which can be easily followedeast northeast to the environs of the modern town ofNa'amaniya on the Tigris, as the northwest boundaryof the present survey. A look at the base map will showthat this was done in part, but that in places dense cul-tivation has now pushed southward across this otherwiseconvenient boundary.

Continuing south along the western perimeter, theAkkad survey extended to Tell Abu Salabikh (A 275)and Nippur. I visited only a limited number of sites inthis area on short field trips during the early 1960s, how-ever, and almost all of them have been rerecorded andassigned new numbers within the present, more compre-hensive system. There are a few exceptions shown onthe base map for which new data are not available (A 221,259, 261, 264, 266, 273, 274, and 275). These are givenunder their original numbers both on the base map andat the end of the site catalog; the data originally pub-lished (Adams 1972) on them has been supplementedwhere possible from the original field notes and fromphotographs of the collections. Since all of this areafarther to the northwest has received some study as partof the Akkad survey, the whole of it has been desig-nated here as one of "limited survey."

South of Abu Salabikh is an area of cultivation pro-vided for by the Shatt al-Dhagharah and its effluents. Itforms a long, broad peninsula extending southeast overthe towns of 'Afak and Al-Bdayr almost to the site ofFara, ancient Shuruppak. West and northwest of 'Afakthis is a zone of dense cultivation, limited drainage, andintermittently marshy conditions. No archaeological siteswere reported by residents of the area, nor were anylisted in the-records of the Directorate General of An-tiquities. Consequently no attempt was made to survey it.

East and southeast of 'Afak is a different matter. Watersupplies for canal irrigation grow progressively sparseras one moves toward the tails of the Shatt al-Dhagharahsystem, and reportedly also have grown sparser duringrecent years as upstream users have claimed a greatershare of the available flow of the Euphrates. Soil salinity,which regular, adequate irrigation might otherwise holdin check, has become a serious problem in many areas.This is accordingly a region in which, as the absence ofarrows on the base map shows, the frontier of cultivationhas remained relatively static since the early sixties. Infact, the frontier can more accurately be said both to havereceded in places and generally to have "softened." Itno longer marks a relatively abrupt transition betweena zone of continuous cultivation and desert. Instead,

35

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

there is a wide zone in which large, irregular tracts havegone out of cultivation altogether, while elsewhere someagriculture is continuing. The remains of numerous aban-doned villages are to be seen (cf., e.g., Nissen 1968), theirinhabitants reportedly having moved to better-favoreddistricts along the Shatt al-Gharraf and elsewhere. Itfollows that many of the difficulties of conducting anarchaeological reconnaissance in cultivated terrain men-tioned earlier are greatly ameliorated here.

Two other conditions combined with this to make a rel-atively systematic survey of the region highly desirable.One was that a number of sites within it were listed in Di-rectorate General of Antiquities records. Information ontheir dates and location proved of variable quality, butin at least some cases it strongly suggested early occupa-tion. Second, reconnaissance of the Warka area in 1967(Adams and Nissen 1972) and the initial phase of re-connaissance around Nippur in 1968 had made it clearthat several important early canals or rivers flowedthrough the area. Hence the region was given carefulscrutiny during the 1975 campaign. Coverage is un-doubtedly somewhat less complete and systematic thanin areas where cultivation is altogether lacking, but it ismore accurately described by placing the region withinthe primary zone of intensive survey rather than in oneof the peripheral zones of limited survey.

South of this peninsula of cultivation is another areathat today is given over to nomads and semisedentaryfolk. It is a region for which the great mound of IshanBahriyat, ancient Isin, may perhaps be thought of as themajor focus. To the west lies another frontier of cultiva-tion, this one expanding vigorously eastward. Here theboundary was not "soft" but continuous and relativelyimpenetrable. It also constitutes the boundary of thesurveyed region, therefore, except in a few instanceswhere mounds were found to be accessible from thesecondary road network serving the cultivated area. Alsoshown within the cultivated area are a number of sitesthat were not visited but that have been assigned numbersbecause they can be approximately located (and in mostcases, dated) from information in the Directorate Gen-eral of Antiquities files. The site catalog summarizes andgives the source of the available information, and a tri-angular symbol is used on the map to indicate that thesize and location have not been accurately assigned onthe basis of direct inspection with the aid of the air photo-graphs, as was the practice for the main body of sitesmaking up the survey.

As I indicated earlier, there are obvious difficulties indetecting small, low sites within the cultivated zone. Inaddition, however, both the records of the directorateand my own unsystematic observations (while repeatedlydriving through the cultivated zone from bases in townsalong the lower Euphrates) suggest that substantial ele-vated mounds are relatively much less numerous than in

the desert. Probably this is because the cultivated zonehere adjoins the Shatt al-Hilla and other western, demon-strably fairly recent Euphrates branches. Relatively small,scattered sites of any period might be expected alongminor streams or canals, but the substantial, closelygrouped, long-continuing mounds of second millenniumB.C. and earlier date are to be sought instead along moreancient Euphrates courses that parallel one anotherthrough what happens to be desert today.

It is also clear that very active alluvial deposition hasbeen continuing along the modern Euphrates branchesfor a full millennium or so, with little or no counterbalanc-ing wind erosion in that immediate region. Hence anymounds there that once were of prominent height arelikely to have been partially or even wholly buried beneatha thick blanket of sediment. But, if that is so, one mustconclude that overbalancing tendencies toward alluvialdeposition also were characteristic along the olderEuphrates courses to the east when the latter were inprimary use. Given those conditions, most of the numer-ous low Uruk sites that constitute a major theme of in-vestigation in this study may well have been completelyburied for a considerable time. The opportunity to studythem today thus is an outcome of the westward shift ofthe Euphrates and the shrinkage of the cultivation frontiersince classical Islamic times, both of which have replacedalluviation with wind erosion as the preponderant geo-morphic force in the desert.

The entire southeastern part of the region covered bythe site map has been previously described as the outcomeof the Warka survey (Adams and Nissen 1972). A sum-mary of the same data is included here to provide as com-prehensive a picture as possible of ancient settlementpatterns on the lower Euphrates floodplain. To avoid con-fusion of reference, the numbering system applying tothe Warka survey has been retained, covering the 466sites in that part of the region. Starting again with site501, 1,139 newly numbered sites are cataloged here. Manyof these include components of widely different ages, andcomprise numerous mounds not immediately adjacent toone another, so that the actual number of sites is some-what larger.

The eastern boundary of the survey is more simply dis-cussed: with only insignificant exceptions, it follows thefrontier of cultivation that is dependent upon feedercanals from the Shatt al-Gharraf. To the south, where theWarka survey was conducted only five years or so afterthe air photographs were taken, no information is avail-able on recent advances in cultivation. At least in thenorth, however, the relatively more plentiful waters ofthe Tigris have led to a wide advance in virtually all areas.Moreover, the newly cultivated lands seemingly arefarmed with greater reliance on agricultural machinery,and in wider, more continuous tracts with dense standsof crops not interrupted by old field canal levees and

36

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

other topographic irregularities. Effective survey undersuch conditions was altogether impossible without agrossly disproportionate expenditure of time.

There is a further area that has been labeled "limitedsurvey" on the period maps. Irregularly triangular andcovering about 800 square kilometers, it occupies thenorthern end of the eastern boundary. This area is largelywithin the cultivation perimeter, although the intensityof agricultural use appears to have declined considerablyin the wake of localized political shifts accompanyingthe Iraqi revolution of 1958. An archaeological surveyof it appeared desirable in order to clarify the source ofthe water that must have flowed past the important earlytown of Tell al-Wilaya. Records in the Directorate Gen-eral of Antiquities indicated that third millennium dateshad been assigned to a number of sites in the area (627,878, 882, 891, and 898) that, if the dates are correct,possibly traced the line of a canal from the Tigris.

As "limited survey" implies, a brief reconnaissance wasundertaken in this area along the secondary road network.It was possible to reach a number of the previously re-ported sites in this manner, including three of the fivereportedly early ones that had seemed to define the thirdmillennium canal line. None proved to have early pottery,nor was pottery earlier than the first millennium B.c.seen on any of the other sites visited in the area. Hencethe special rationale for conducting an intensive studyof the area disappeared, and its exclusion from the normalsurvey perimeter because it lay within the cultivated zoneled me to cease work there. The problem of the sourceof Tell al-Wilaya's water unfortunately remains as enig-matic as ever.

The final survey boundary needing comment is internal.Immediately west of the area just described is a largedepression known as the Haur Dalmaj. This was alreadyrecorded as a perennial swamp in the late nineteenthcentury (Kiepert 1883), at a time when it must have beensupplied primarily by periodic Tigris floods and runofffrom adjacent areas of desert. A half-century later theBritish quarter-inch map series indicated that it had grownsubstantially in area, supplied by the tails of pump-fedcanals that were being rapidly introduced along the Tigristo the north. It continues today, now receiving massivefurther water supplies from the main drainage canal ofthe Mussayib irrigation project that has its outlet in thevicinity of site 646. The total area permanently affectedis about 150 square kilometers, including broad ex-panses of more or less perennial water interspersed withthe remains of ancient levees and other topographicirregularities. Although the water obviously rises and fallsaccording to the season, an attempt to survey this areain late August 1973, theoretically under the lowest waterconditions, was abandoned when I discovered that tem-porarily exposed areas of the bed of the depression werecomposed of treacherous and impassable sabkha soil. In

the spring of 1975 therefore, survey was continued only tothe margins of the standing water. The depression itselfmust be considered wholly outside the zone of study.

RECONNAISSANCE SEQUENCE, PROCEDURES,AND EFFECTIVENESS

The fieldwork associated with this study took place infour unequal phases, necessarily separated by long in-tervals. Initially it had seemed possible to plan for aseries of campaigns of moderate duration in sequentyears. The first, lasting from late October through Decem-ber 1968, led to the recording of 389 sites northwest,northeast, and southeast of the Oriental Institute's fieldheadquarters at Nippur. This led to a heightened aware-ness both of the overwhelming abundance of Sasanian-Early Islamic surface remains and of the considerableuncertainties surrounding the ceramic chronology of theseperiods. Hence the remainder of the first season wasdevoted to a stratigraphic sounding in a small site oc-cupied during those periods, to provide an improvedtypological sequence with which to undertake the furtherreconnaissance that was expected to resume the follow-ing autumn (Adams 1970).

A few months later, however, a formal application tocontinue the reconnaissance was denied. I devoted severalensuing years to research on other themes and to a Uni-versity of Chicago administrative assignment. During avisit to Baghdad in early 1973 I once again obtained anapproval in principle. A very brief second campaign wasthereupon conducted in late August and September 1973.My intention to penetrate certain swampy areas northof Nippur during the season of low water proved un-feasible, but some of the procedures needed for a moresystematic reconnaissance were elaborated. An additionalsixty-seven sites were logged in areas northwest of Nippurand north of Tell Abu Salabikh. Administrative responsi-bilities precluded continuation of the work for longerthan a few weeks, but on this basis plans were laid fora major resumption in the succeeding academic year.

There were moderate further delays, and for a timeit appeared that the Iraqi authorities might once againrescind the needed approval. Ultimately, however, thethird and longest phase of fieldwork got under way inearly February 1975 and continued through May. Tem-porary field camps were established in a succession ofplaces indicating the wide geographic coverage to whichthis campaign was devoted: initially the German ex-pedition headquarters at Warka, then sequentially in thetowns of Rumaytha, Diwaniya, 'Afak, and Na'amaniya.In all, 707 sites were recorded, although in this as in thepreceding cases the number of provisional listings wasslightly modified by later analysis.

Finally, about a fortnight in December 1975 was ex-clusively devoted to certain methodological problems for

37

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

which further field data were needed. Several additionalsites were recorded that had been overlooked, but thebulk of the effort was devoted to sampling proceduresand to intensive collections from many of the small pre-and protohistoric sites. This activity was limited to whatcould be accomplished within a day's journey from thebase at Nippur, and thus it coincided for the most partwith the area initially surveyed in 1968.

The breakup of the work in this unforeseen fashion hadseveral important consequences. In particular, succeedingphases embodied improvements in method and theory thathad been worked out during the interim. In 1968, for ex-ample, I had not yet made the acquaintance of any of thediverse and promising formal approaches to locationaltheory. Until after the end of the reconnaissance that year,moreover, I did not have access to the clarification of theSasanian-Islamic ceramic chronology that stemmed fromthe excavations at Tell Abu Sarifa in the months immedi-ately following. By the 1975 season similar improvementswere also available for the protohistoric ceramic chronol-ogy as detailed in Appendix A to chapter 3.

To some degree, these and other cumulative improve-ments were retrospectively applied to the initial findingsduring the final phase of methodologically oriented field-work. However, this was necessarily limited to a relativelysmall number of sites, and particularly to those in the pre-/protohistoric and Sasanian-Early Islamic periods. Thereremain many sites to which somewhat different datesmight have been assigned had they been visited in 1975 in-stead of 1968 or even 1973. Those for which this seemsmost likely, either on the basis of entries in the originalnotes or on the basis of restudies at adjacent sites wheresimilar original datings were later modified, have beenmore or less tentatively corrected in the site catalog and inthe relevant tabulations. But some differences based onsequential changes in the dating criteria obviously remain.

There is another consequence of the repeated, involun-tary setbacks to the research schedule that is less obviousbut no less important. Uncertainties surrounding this pro-gram of archaeological survey varied in intensity but werealways present. Apart from serving as a brake on long-term forward planning of the research, these uncertaintiestended to place a premium on rapidly completing at leastan initial reconnaissance of a broad area rather than onapplying a more elaborate and time-consuming surveymethod to a restricted area. That judgment on priorities,dictated by external circumstances, remained uniformover the /entire seven years that the field research waseither in abeyance or in progress. It had a decisive bearingon the selection of the site-detection and dating proce-dures followed, with which I now must deal in furtherdetail.

As I suggested earlier, conditions over a great part of thesurveyed area did not permit a uniform, gridlike coveragein the search for ancient sites. Belts of dunes and season-

ally filled depressions were the most evident obstacles,but other impediments to rectilinear travel were almostas serious. Sand hummocks that have formed arounddesert shrubs, for example, have converted much of theuncultivated plain into a surface that is at best rough andoften all but impassable. Wind erosion, concentratedalong the raised surfaces of ancient levees, often gougesout formations of buttes and steep-sided channels that areeven more difficult for a vehicle to traverse than theworst of the hummocks. In places, as in the desert aroundancient Isin, which was abandoned only late in the pastcentury, the steep spoil banks and trenches of formercanals still act as barriers for considerable distances. Asa subjective estimate, less than a third, and probably lessthan a fifth, of the area within the zone of intensive studywas an open desert land surface permitting unimpededmovement and easy site detection.

In the limited areas where the going was good, some-thing approximating a series of parallel traverses was gen-erally carried out, at distances of about 1 kilometer. Differ-ent tactics had to be adopted where this was not possible,varying according to both the potential importance of thearea to be surveyed and the nature of the obstacles en-countered. Two slightly condensed excerpts from my fieldnotes, both dealing with unusually difficult areas for re-connaissance that were also of considerable significancefor the data they might furnish on third and fourth mil-lennium watercourses and settlement patterns, illustrateboth the general orientation and how it tended to beapplied in varying circumstances:

Some description of desert wasteland between Jidr (004)and Bismaya (ancient Adab) is necessary in order to under-stand the limitations on the physical possibilities of surveyin that region. The land surface is extensively carved bywind erosion into badland formations with as much as30-40 centimeters of relief, making jeep travel virtuallyimpossible in a consistent direction and tortuous under allcircumstances. The ancient canals generally survive asslightly raised strips of compacted earth, although rough-ened into "riffle" patterns by wind erosion and gulliedby rain erosion as well. Thus one frequently cannot evenfollow the canal-strips in a vehicle. Add to this the sandhummocks, and the sheet deposits that may become in-distinguishable from alluvial deposits after a few years ofrain, plant growth, and compaction. And then there arethe dunes, not infrequently rising 5 meters or more andsometimes forming such tightly packed clusters that noinspection whatever of the underlying land surface is pos-sible. [For a representation of dunes in this area, see fig.32.]

Within the realistic constraints of the time available, Ihave bulldozed through along the major watercourses,one way or another-particularly the major watercourserunning southeast into Jidr. Elsewhere, however, it is justnot possible or productive to cover the terrain systemati-cally. Instead, I have picked up the likely-looking loca-

38

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

tions for mounds from the air photos and have foughtthrough to those locations as nearly as I can find them,keeping an eye open for other, small or less likely sites onthe way. Fortunately, erosive processes are very wide-spread in this region, while on the whole the mounds arerelatively immune to them (because of their cap of sherds).Hence the tells have turned out to be generally distin-guishable from this ravaged landscape on the air photos.Site 1315 is an exception, but in most cases one can im-mediately identify the location of a particular tell on thephotos if one can place it within a square kilometer or so.What we are most likely to lose, of course, are preciselythe small, early tells like site 1315, but there is just nopractical way of digging them out within the constraintsof a massively obscured land surface and a relatively brief,one-vehicle operation. Even with quite prominent tells,one frequently has to cruise around in an area, climbingsuccessive dunes for a look at the hollows between them,before a particular canal intersection can be recognized.Small, low, early sites, even if they are not totally ob-scured, could be picked up reasonably satisfactorily onlyby systematic low-level helicopter traverses. It's only avignette, but finding site 1316 from site 1315, being reason-ably certain that the discoloration on the air photo didindeed represent a tell, took about twenty minutes ofcircling. Yet site 1316 is 250 NNW x 180 x 3 meters high!

I assume that I have been able to locate the bulk of thelarger sites in the area with the air photos-those exceed-ing 4 hectares in area and with appreciable elevation. Ihave no idea how many small sites we may be missing inthe region, although, if one tries to judge from the plainnorth of Fara (ancient Shuruppak) that lies west of thissand country, it appears that this was not a region of in-tensive prehistoric settlement anyhow.

The second, shorter excerpt emphasizes some of thesame themes-in particular, focusing on ancient water-courses as clues to settlement location. It concerns site1306, an important protohistoric town:

Site 1306 was discovered only after the watercourse tothe south of it was confidently transcribed from the airphotos as being indeed an ancient watercourse. Now thelatter seems in a position to have carried the effluent fromthe former. Note that there is no evidence of meander-cutting along it, perhaps suggesting that it flowed underdeltaic rather than alluvial conditions. But while a subse-quent serious attempt was made to detect ancient sites tothe south of site 1306, it was wholly unsuccessful-inde-terminate rather than negative. Dense dunes and, moreimportant, heavy wind-laid sheet deposits, have elimi-nated land surfaces on which sherds from low, early tellsmight be encountered. Tamarisks may have played acrucial role in this and similar cases-growing to takeadvantage of available groundwater in the coarse-tex-tured sediments of former watercourse beds and then, bytheir presence, attracting dunes that grow up around theirroots. All this was particularly clear south of site 1306. Ifollowed the probable course south for about 5 kilome-

ters, crisscrossing back and forth by compass over wherethe course was likely to be, and can only say that thereis nothing to mark what is seen in the air photo except anunusual concentration of tamarisk.

It should not be thought, however, that omissions incoverage were exclusively confined to areas like these,where topographic conditions made regular gridding im-possible. There were subtler but significant difficultieseven in areas where dune formations were wholly absent,some of the most insidious of them subject to variables ofwhich the investigator only inadvertently becomes aware.The revisits in 1975 to a number of low Uruk sites thatwere originally surveyed in 1968, in order to make inten-sive collections that would aid in more accurate dating,offered one such occasion. Great difficulty was found inrelocating some of these, in spite of reliable bearings thatpermitted the search to begin within 100-200 meters ofthem. An excerpt from field notes, concerning the restudyof sites 804, 805, and 837, again provides an illustration:

none of them was found in less than twenty to thirty min-utes of looking. With a less irregular surface we could havegridded the area more quickly by driving over it, but inthese circumstances we had to do so on foot amongbushes that were prevailingly 20-30 centimeters high.These bushes were not densely distributed, but they ab-solutely prevented sensing any gradient in sherd densityoutside an area extending more than about 2 meters inradius from the viewer. And since a sparse surface scatterof Uruk pottery occurred throughout the area, we had tocrisscross it on foot until the major concentrations finallywere found. I should also note that widespread hummockformation and sheet deposition was beginning aroundthese low bushes, further obscuring the surface. Clearly,it would be quite impossible to duplicate my initial 1968coverage of this region under present conditions.

When I sought an explanation of this phenomenon,local informants indicated that there had been unusuallyheavy precipitation in the immediate area during the pre-ceding four winters. This seems to explain the prolifera-tion of desert shrubs and the ensuing entrapment of air-borne sand. Even if there is now a termination of thatcycle, the effects will continue for some time. The dor-mant or dying roots and stalks of these shrubs will con-tinue to act as traps for aeolian particles-perhaps even atan increasing rate, if the volume of airborne dust is posi-tively related to a decline in winter precipitation. In short,the uniformity and effectiveness of survey is inextricablytied to local variations in microclimate as well as to morereadily perceived differences in gross topography.

It is apparent that survey under conditions like thesecannot be exhaustive. Small, low sites are particularly dis-favored in the detection process, but even fairly large onesmight well be overlooked under topographic conditions

39

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

that are by no means rare. For the protohistoric sites,which are preponderantly small and which in any case arethe most likely to be buried beneath an overburden oflater debris, the possibility of substantial loss is especiallysevere. Thus the need is clear for measures that both mini-mize the loss and provide some measure of its extent.

Several steps were taken in all phases of the study toincrease the recovery rate for protohistoric sites. First, amuch more complete record was kept of occurrences ofprotohistoric sherds (see Appendix A to chap. 3) than ofthose attributable to any later period. Even single exam-ples on large, late mounds, possibly to be dismissed asonly "strays" since many tens of thousands of surfacesherds could be seen, were uniformly noted. Their signif-icance will remain to some degree a matter of conjecture,at least until a number of such sites have been strati-graphically sampled to determine whether the major occu-pations overlie small, early ones. For purposes of thisstudy, I have more or less arbitrarily assumed that threeor more early sherds on a site with a heavy later over-burden are indicative of an early occupation while onlyone or two may not be. Second, it became apparent veryearly that the protohistoric sites tended to cluster in cer-tain areas and to avoid others. Hence I followed the prac-tice of driving slower to increase the intensity of coveragein any area where even a few protohistoric sherds hadturned up on later mounds. Third, it was found thatsherd distributions around early sites can greatly aid theirdetection. The mechanism of dispersion is obscure, but itmay well involve the cumulative effects of later plowing.Henry Wright's account of the Ur survey in the Appendixto this book mentions the presence in that region of strayclay sickles (attributable to the Ubaid or Uruk periods) inbelts 3 to 5 kilometers wide. He suggests that these mayindicate the perimeters within which cultivation was con-centrated. On the plain between the Tigris and the Eu-

phrates to the north of the Ur region, however, no strayearly sherds of any kind were noted at such substantialdistances from early sites as this suggests. Isolated exam-ples were indeed seen from time to time on an otherwiseempty plain. A process of systematic gridding was initi-ated in these cases, either on foot or by vehicle at whateverintervals were indicated by the local terrain, and one ormore protohistoric sites regularly were found within akilometer or so. If there is any substance to an unsystem-atic impression, moreover, clay sickles were found withno greater relative frequency in purportedly cultivatedareas away from sites than on the sites themselves. To thatit must be added, of course, that on some Early Uruk sitessickle fragments outweighed all other identifiable ceramiccategories considered together (see below, p. 124).

A less generalized, more formal step also was taken toassess the efficacy of survey procedures, on the basis of arestudy of a sample of previously surveyed areas. Usinga grid of 10-kilometer (100 km2) squares in uncultivatedareas both northwest and southeast of Nippur, extendingroughly from Tell Abu Salabikh to Bismaya (ancientAdab), a stratified systematic unaligned sample (Haggett1966, pp. 196-98) of one-kilometer squares was drawnwith the aid of a random-number table. This samplingdesign assures a wide dispersion of locations while main-taining randomization within each larger square in orderto avoid the effects of possible periodicities in the phe-nomena being studied. A map of the distribution of theresultant one-square-kilometer plots is given in figure 7.Having designated loci for restudy in the fashion shown,without reference to sites already known within thesesquares, in December 1975 I attempted to delimit theboundaries of each of them and conduct an intensive re-survey within those boundaries. The results of this under-taking are described in table 2.

To summarize the table briefly, thirteen one-kilometer

Fig. 7. Sample one-kilometer squares for restudy of survey coverage.

40

oi.uchicago.edu

TABLE 2 Results of Resurvey of Sample Square Kilometers

Square A Topography: Old, stabilized, but very rough hum-mocks around shrubs.

Sites initially recorded: None.Modifications or additions: None.

Square B Located within the limits of the Haur Dalmaj (cf.pp. 000-00). Unreachable for survey in 1968and 1973, or for restudy in 1975.

Square C Topography: NW is old, stabilized aeolian depositaround bushy tamarisk, roughened by recentwind erosion; SE is clear and slightly elevatedcanal levee.

Sites initially recorded: 859 immediately adjoinssample area.

Modifications: Description of 859 modified to in-clude small NE outlier possibly just within thesample area. Dating modified to include proba-ble post-Samarran as well as the Sasanian-Sam-arran occupations originally assigned.

Additions: None.

Square D Topography: Much recent wind deposition.Sites initially recorded: None.Modifications or additions: None.

Square E Topography: Many dunes, much recent wind dep-osition, including hummocks around sparsevegetation. But obstructions still widely enoughscattered to permit thorough vehicular ratherthan foot reconnaissance.

Sites initially recorded: None.Modifications: None.Additions: A small, low mound was identified 400

m NW of 1195. Largely contemporary with thelatter and given the same site number, it mayalso have a small, underlying Uruk occupation.

Square F Topography: Stable plain surface, generally freeof hummocks or other wind deposit. Moderatevegetation.

Sites initially recorded: 1290, 1291, 1292.Modifications or additions: None.

Square G Topography: Numerous dunes, but widelyenough separated to permit vehicular recon-naissance. Generally vegetation-free.

Sites initially recorded: None.Modifications or additions: None.

Square H Topography: Level surface except for small,sparse hummocks. Moderate vegetation.

Sites initially recorded: 750.Modifications: Description of 750 modified to in-

clude smaller ENE outlier. Dating modifiedfrom Sasanian-Samarran to Middle Islamic.

Additions: None.

Square I Topography: Old, stabilized wind deposit aroundmoderately dense vegetation. Low sites mightbe completely obscured in the NW third of thesquare.

Sites initially recorded: 838.Modifications or additions: None.

Square ] Topography: Dense dunes in SE portion of

square; considerable recent wind depositionelsewhere.

Sites initially recorded: 1034; only one of twoadjacent mounds given this number falls withinthe sample area.

Modifications: Assessment of the major period ofoccupation remained unchanged, but traceswere found of an earlier occupation that hadnot been noted previously.

Additions: A third mound was located 400 mWSW, contemporary with and much smallerthan the other two.

Square K Topography: Low, discontinuous wind deposit,alternating with rough, wind-eroded surface.

Sites initially recorded: None.Modifications or additions: None.

Square L Topography: Much low, sheetlike aeolian deposit,small dunes, little vegetation.

Sites initially recorded: 1216 adjoins sample area.Modifications: Sasanian-Early Islamic reidentified

as the major occupation of 1216, but a small,localized Neo-Babylonian-Achaemenian settle-ment also found on one end of the mound thathad not been noted originally.

Additions: A low-lying, early mound found onfoot reconnaissance 400 m SE and includedwithin the description of 1216.

Square M Topography: Abandoned cultivation, manyditches; vehicular travel impossible.

Sites initially recorded: None.Modifications or additions: None.

Square N Assigned square fell in extremely dense dunes be-yond Adab, and hence outside perimeter ofstudy. Not visited.

Square 0 Topography: Old canal levees with sparse vege-tation. There are low dunes in N corner ofsquare, but vehicular reconnaissance is possibleamong them.

Sites initially recorded: 904, 905, 908.Additions or modifications: None. As usual with

Sasanian-Islamic canals, there is a little scat-tered cultural debris at intervals along the levees.

Square P Topography: Originally scattered dunes and vir-tually vegetation-free plain bordering cultiva-tion. Recently there has been a massive govern-ment-sponsored settlement project coveringmany square kilometers in this area, accompa-nied by land-leveling, extensive canalization,and deep tractor-drawn plowing. These condi-tions precluded a revisit, but in any case small,low sites that were originally overlooked wouldhave been destroyed.

Sites initially recorded: 962, 963.

Square Q Topography: Dense, active dunes bordering cul-tivation. No sample square was assigned sincevicinity of Nippur is uniquely well known fromexpedition activities and hence unrepresenta-tive.

41

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

squares were restudied, of which eight were left un-changed. A total of nine sites had been identified duringearlier surveys of this area of 13 square kilometers. Threesites were located that had not been recorded originally,one in each of three squares. And the original descriptionor dating of a site was modified in two of these lattersquares as well as in two others.

What do these figures imply, not only about the grossorder of completeness of the survey as an exhaustive rec-ord of surface remains in the area, but also about thenature and significance of the sites that have been omitted?If the proportion of originally recorded to newly discov-ered sites holds true generally, it appears that the total of1,605 sites (including 466 published in the Warka survey)may be deficient by as much as one-third. Or one couldcalculate alternatively on the basis of areas and not sites,adding 3/13 additional site for each of the approximately6,250 square kilometers that the intensively surveyed areacomprised. This would suggest an even larger incrementof unrecorded sites, only slightly smaller than the numberrecorded. However, it is apparent that there is a high de-gree of clustering of sites along a small number of majorancient watercourses from the northwest that coincidewith the area of the restudy. Hence the proportion ofomitted sites is likely to be much closer to the formeralternative than to the latter.

In fact, even the lower proportion is likely to be some-what misleading inasmuch as several of the additional,previously unrecorded sites (1195, 1216) lie on the bound-aries of one-kilometer squares. Were the squares morerigorously delimited than mere compass bearings andair photographs allow, the number of additional sitesfound in the sample might shrink from three to two oreven one. But the more inclusive approach of counting allof them is preferable. My point in the exercise is not toemerge with a misleadingly precise statistical constant toapply indiscriminately, but rather to illustrate concretelythe omissions inherent in a survey conducted under theseconditions-and to suggest only very roughly their proba-ble order of magnitude.

Perhaps more significant than the absolute number ofunrecorded sites is their character and placement. All wererelatively small and low, in fairly close proximity tolarger mounds or mound complexes. They argue stronglyfor the obvious point that sites of moderate to substantialsize and elevation are much less likely to be missed. Forthe later historical periods in which the upper echelons inthe site hierarchy play the preponderant part, it appearsthat the omissions do not significantly distort the aggre-gate picture of settlement. For the pre- and protohistoricperiods, on the other hand, the omissions may well leadto a substantial underestimate of site numbers and a cor-responding overestimate of average site size.

These observations can be reinforced and extended byconsidering the sites for which the restudy suggested mod-

ified descriptions or datings. At two of the four sites, minoroutlying mounds were observed that had not been notedpreviously. Both were contemporary with the larger,neighboring mounds that were originally recorded, andthe substitution of different datings reflected only an im-proved knowledge of thejIslamic ceramic sequence stem-ming from intervening excavations at Tell Abu Sarifa. Atthe other two, localized early components were noticedthat had originally been overlooked. Considered together,these modifications suggest that one can slip too quicklyand easily into the practice of identifying a site by itsmajor mound and dating component. Multicomponentsites are common, and more systematic efforts to identifythem may in fact be a more economical and effective wayto increase the representativeness of survey findings thangridding the entire survey terrain at smaller and smallerintervals. In any case, it again appears that overlookedcomponents, like overlooked sites, tend to be those of verymodest size.

The massive ruins of the Warka area furnish a prac-tical example of the difficulties presented by extensive,multicomponent sites. Both around the ancient city and tothe north of it, along the broad old levee of the Shatt al-Nil, sprawling low mounds and surface debris are almostcontinuous. There is an evident temptation, to which it isnow apparent that the Warka survey fell victim, to dealwith an area of this kind by making periodic transectsacross it and especially by circling its outer peripheries,where sand accumulations are less dense and the going isgenerally easier. Such an approach obviously means thatone is generalizing from what has been sampled to whatis not seen directly. That problem can never be completelyeliminated no matter what intensity of survey procedureis followed, and a trade-off of greater intensity for widerextent of survey coverage was in any case an explicit partof the approach that was taken. But one must rememberthat journeys along the outer edges of ruin fields do notconstitute an unbiased sample of them, but instead givean account that overemphasizes the period of maximalextent of settlement.

It has recently been shown that we failed to observeSasanian remains in the area immediately southeast of theWarka city wall, and both Sasanian and Early Islamic re-mains along the ancient canal north of the city (Finsterand Schmidt 1976, pp. 164-66). The extent and durationof the settlements concerned remain an open question; atleast the former one is apparently of considerable size. Aneffort is made to correct the omission by incorporatingthe new data in the appropriate period maps accompany-ing this volume, and it should be clear by now to anythoughtful reader that future discoveries of additionalomissions at least as substantial as these are to be ex-pected. But what is important from a methodologicalviewpoint is that the more extensive scale of Seleucid-Parthian (not to speak of earlier) settlement in and around

42

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

Warka probably helped to thwart our identification ofthese relatively more modest later remains because of anoverreliance on search of the more topographic featuresand along the outer circumferences of ancient settlement.

A final area of concern for the effectiveness of site de-tection procedures involves the mounds in cultivatedareas that have been incorporated into the findings of thesurvey directly from the files of the Directorate Generalof Antiquities. In all, there are about fifty-five of these,concentrating in one group extending irregularly northfrom Rumaytha and a second group southeast of Na'ama-niya. A number of additional sites were similarly recordedin the files, in these areas and elsewhere, but are pub-lished here as integral parts of the survey since they wereindependently visited, described, located, and dated. Com-paring the findings for these latter sites with the reportsin the directorate files provides a basis for establishingthe degree of accuracy of the original reports. To what de-gree do the reports permit an extension of survey cover-age into regions beyond its own self-imposed limits?

The answer must be carefully qualified. Individual re-ports have been assembled by a succession of inspectorswho have varied greatly in experience, over a span of morethan four decades. Files have been kept on the basis ofsite name rather than location, and location has almostalways been transcribed from existing maps of varyingstandards rather than independently confirmed withcompass bearings. Hence individual sites have been foundto be as much as 10 kilometers out of position. The reli-ance on site names for file designation also produces anobvious concentration on large, prominent mounds forwhich names can be easily elicited and a correspondingneglect of small, low sites. The reliability of dating assign-ments is also somewhat variable, although in general-with earlier periods an important exception-the reportedperiods of major or terminal occupation are congruentwith my own findings.

Such selectivity and variability is a severe but by nomeans crippling limitation on the utility of the files forcertain purposes. Named sites are at least placed in theirapproximate region of occurrence, so that they can gen-erally be found with the aid of a local guide. Viewed asevidence of a regional pattern rather that as a grouping ofsites individually conforming to fairly uniform standardsof accuracy, the files permit at least certain tentative gen-eralizations about the time of onset of major phases of set-tlement or abandonment in certain areas. It is in this cir-cumscribed fashion that records on sites not directlyvisited during the survey are employed here.

SITE SURVEY, COLLECTIONS, AND DATING

Previous references to "sites" have perhaps tended toleave the impression that they form a coherent, self-evi-dent category. One might wish that this were so, but it is

not. The recognition and measurement of individualclusterings of ancient debris involves a host of interactingobservational variables that frequently cannot be keptseparate from matters of subjective judgment.

Density of debris is one important indicator. Randomlyoccurring sherds or other fragmentary remains may beexpected almost anywhere in a region as long and as in-tensively settled as Mesopotamia. At what point in anascending scale of density must a particular occurrence nolonger be dismissed as insignificant or random but re-garded as indicative of an ancient settlement? No attemptwas made in this study to establish a quantitatively rigor-ous standard for this determination, although such a stan-dard is obviously feasible. Its drawback is a spurious im-putation of precision and consistency, since like effects canbe produced by a wide variety of unlike circumstances. Aspare distribution of sherds, for example, may reflect unim-portant "strays," a low occupational mound completelyburied beneath a blanket of later alluvial sediments, orperhaps a specialized structure like a police post in whichthe contemporary use of pottery was minimal. In any case,sparse, widely dispersed distributions were not directly re-corded as indicative of sites unless they were accompaniedby unambiguous evidence of architecture or mound build-up. Where such distributions were encountered, however,every effort was made-in the vast majority of cases suc-cessfully-to identify a nucleated zone of more concen-trated debris from which the outlying traces could beassumed to have spread through various forms ofdisturbance.

Site dimensions raise other problems of uniformity anddefinition. At one extreme we may think of a "pot drop,"a tiny clustering indicative of an isolated, transitoryevent. Or, even if less transitory, should an area of debrisso small that it may stem from a single rural farmstead begiven the status of a site? Unquestionably it should forsome purposes, but it is doubtful that so minimal a defini-tion will be helpful in a very wide-ranging and prelim-inary reconnaissance of the kind described here. At theother extreme, after all, are a considerable number oflong-lived and important cities that have not previouslybeen identified. Though it is beyond dispute by anyonethat they are sites, even these remains of urban size fre-quently pose their own problems of definition. Their ruinfields, for example, are seldom completely continuous.How do we avoid creating a misleading impression of sizeand nuclearity by giving a sprawling, vaguely defined areawith variable densities of debris a single site number andset of dimensions?

The standards adopted for dealing with this problemwere again not made rigorous; in fact they were con-sciously kept variable. For pre- and protohistoric sites,individually the smallest and the most subject to lossthrough alluviation and other forms of later disturbance,essentially no clustering of debris was too small to be re-

43

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

corded (the problem of "pot drops" fortunately did notarise, since nothing on so small a scale was encountered).For the Sasanian and Early Islamic periods, characterizedby vast areas or relatively well preserved but amorphoussettlement, small clusterings of less than about 25 metersin diameter were less often recorded independently unlessthey occurred at a considerable distance from larger set-tlements or provided a date for associated canal remains.Remains of the third and second millennia fell somewherebetween these differing intensities of study-on the whole,rather closer to the prehistoric end of the continuum.

There is seldom a clear demarcation between an ancientmound and the surrounding plain. Quite apart from un-certainties as to the extent of refuse accumulations or out-lying habitations around its foot, erosion transports slopematerial outward for considerable distances if a mound'selevation is substantial. Plowing and canal constructioncarry the process further. Very commonly the line of alater canal seems to have been sighted on an abandonedmound, with the canal then dug directly through themound (in spite of the greater labor involved) rather thanrouted around it. Tending to counteract these centrifugalforces is an uncertain and probably shifting combinationof alluviation and wind deposition, both acting to burythe outlying slopes beneath a blanket of sterile material.Without the aid of excavations, therefore, the measure-ment of a mound's dimensions is useful for relocating itand as an index to its approximate size but should not betaken as an accurate gauge of the area of originalsettlement.

With the emphasis on maximizing the extent of cover-age, I devoted time to preparing sketch maps of mounds ormound complexes only in exceptional cases. More nor-mally, in the prevailing case of an elliptically shaped singlemound, I paced the longer axis and merely estimated theshorter one as a proportion of this. I also estimated heightsfrom a position on the plain far enough from the moundfoot so that its elevation could be scaled against the hori-zon. For higher mounds, I took lines of sight toward anintersection of the horizon with the outer slope. I thenwalked to this point, repeating the process if necessaryuntil the summit lay below the horizon. But while I oftenrelied on horizontal and vertical estimates of this kind, Ichecked them at frequent intervals (in the latter case alsousing a hand level) to increase general reliability andminimize any tendency to develop unconscious biases.Obviously, I could have used techniques giving muchgreater precision. My impression is that the fundamentaluncertainties would remain the same, however, and that asubstantial investment of time toward this end wouldhave produced a largely spurious accuracy.

Surface examination directed toward dating requires amore differentiated approach. Low sites with relativelybrief spans of occupation fortunately predominate in thekey pre- and protohistoric periods. But multiperiod

mounds are common, particularly in the third and earlysecond millennia, and frequently involve some lateralmovement and transport of materials for building ratherthan simple superimposition. Digging of foundations,wells, and graves plays a part in churning up earlier ma-terials during the later life of a settlement, and slope washcontinues indefinitely to blanket its lower slopes withdebris from higher levels. Intensive sampling proceduresdirected toward eliciting the size of occupation duringsequent phases are unconvincing under most such condi-tions, even when research designs deal exclusively withsmall sites and relatively limited areas. On the other hand,one can generally identify at least the presence of partic-ular periods of occupation through ceramic "index fossils"(Adams 1965, p. 121), and one often can make some esti-mate of the extent of occupation at a particular time bynoting whether these chronological markers are localizedor general in their distribution. This procedure was fol-lowed for substantially elevated, multiperiod sites. It in-volved systematic coverage on foot of all the major areasor components of a site, comparing the evidence for dif-ferent periods of occupation on each to synthesize a pic-ture of the site's settlement history.

We must again note that the density of surface debrisvaries widely, on low single-period sites as well as highmultiperiod ones. Several factors appear to be involved,seldom having a uniform influence or acting in the samecombination. Sheet erosion and slope wash are accentu-ated on higher mounds, concentrating sherds from themound's uppermost levels upon an underlying erosionsurface as those levels are gradually removed. Wind ero-sion has the same effect, although the local availability ofunconsolidated abrasive material (e.g., dunes) to act as awind-borne erosive agent probably is more important thanmound elevation in establishing the rate of downcutting.In not a few cases the survey encountered a mound fairlysoon after it had apparently been crossed by a highlyerosive, rapidly moving dune, and we then found the sur-face littered with complete or relatively complete vesselsleft behind after the entire topmost layer had been re-moved. On the other hand, the gradually increasing con-centration of sherds on a mound's surface slows its rate oferosion by acting as a protective cap. Similarly, the heavyreliance on baked brick in Sasanian and Islamic times,surely for private house foundations as well as for wallsin larger buildings, has greatly slowed the rate of erosionfor sites with these as terminal levels. Hence high concen-trations of surface sherds on such sites are rare, their spo-radic occurrences perhaps indicating no more than a localabsence of building activity.

These observations on erosion and dating have a bear-ing on the methods of estimating mound height that Idiscussed earlier. To begin with, they suggest that mea-surements of any kind are likely to be valid only for a lim-ited and uncertain period. This supports my judgment

44

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

that efforts at time-consuming precision are on the wholeinadvisable. In addition, the volume of a mound is seento depend heavily on the nature of the erosive processesto which it has been exposed and on the resistance offeredby surface materials in different segments. Variable alluvi-ation and wind deposition (or erosion) on the plain sur-face adjoining the mound further obscure calculations ofmound volume. Recently it has been proposed that moundvolume offers an attractive approach to estimating popu-lation size, by way of uniform assumptions about housevolume and duration of use, lack of reuse of building ma-terials, and the absence of significant contributions tovolume other than from the decay of domestic architec-ture (Ammerman, Cavalli-Sforza, and Wagener 1976, pp.41-53). Each of these assumptions can be seriously ques-tioned in the Mesopotamian case, and the further uncer-tainties besetting even a rudimentary calculation of moundvolume once again render the attempt at additional accur-acy largely futile.

No reference has yet been made to recent, purposivehuman disturbance. Casual surface collecting of sherdsand certain other kinds of artifacts, a cumulatively sig-nificant source of distortion in many other areas, seemsalmost unknown in this region. On the other hand, thereis an ongoing problem with illicit excavators, perhapsmade bold by the desert solitude. Most of their depreda-tions are fairly small-scale and shallow, possibly no morethan the idle picking by a passing camel-herder at a partlyexposed pot or bones suggesting a burial with offerings.But at least one large, organized gang was at work duringthe field activities of the survey in 1975. Moving rapidlyby truck and concentrating on Early Dynastic and Akka-dian sites, it was gutting virtually the entire surface ofimportant towns like site 1188 with thousands of pitsdug over intervals of a few days or at most weeks. Ulti-mately, of course, erosive processes will again do theirwork and restore a variegated assemblage of surface frag-ments with which an archaeological surveyor can work.But the subsurface loss is irretrievable. Fortunately, theoverall number of sites that have been attacked in thisway is still relatively small.

This discussion of sherd densities should not be takento imply that there is generally a uniform gradient towardmaximal sherd concentrations on the oldest sites. It isprobably true on the average that the pre- and protohis-toric sites have the densest surface sherd assemblages, butthere are many exceptions. Low, early tells in cultivatedor periodically flooded regions may be submerged beneatha mantle of alluvium, thus suffering no surface erosion atall for an extended period until reexposed by wind erosionof the entire surrounding plain surface. In fact, somesparse scatters of early sherds may not be primary sitesat all, but only bits of debris from entirely buried sites thathave been brought to the surface by continued plowing orother disturbances.

Most early sites, however, have a uniform surface ap-pearance characterized by relatively high sherd density.This invites a more intensive approach to surface collec-tion and dating. Sites of the Uruk period, in particular,provide access to a crucial and poorly understood chapterin the development of urban civilization in Mesopotamia.Hence observed "index fossils" on these sites were moresystematically recorded during all phases of the survey.In the 1975 season, moreover, general presence-absencerecords were supplemented by intensive, quantitativelyrecorded collections.

The unit of analysis for most of these intensive collec-tions was a circle of 5 meters radius, drawn by circum-scribing one stake around another tied to it by a lengthof cord. All diagnostic sherds and stone utensils withinthese circles were collected, separated according to type,and counted. We made one or more such collections oneach site, depending on it size, its single- or multicom-ponent structure, and the time available. In a few caseswhere density did not permit an adequate collectionwithin this limited area, all the diagnostic material wefound was brought together and counted. At other timeswe recorded only a generalized presence-absence tabula-tion-for example, when the presence of a later over-burden severely limited the amount of early material, orwhen we had to leave quickly because of dusk or heavyrain. All available counts and tabulations, covering pri-marily sites first located in 1975 but also a number thatwere found earlier and revisited for the purpose, are as-sembled in Appendix A to chapter 3. In the intensive col-lections as in all others, essentially no materials werepermanently retained after I had noted their presence. Therequest to adopt this procedure was transmitted by theDirectorate General of Antiquities representative initiallyassigned to accompany the survey, both to reduce theburden of storage on Iraq Museum facilities and to retaina maximum number of specimens in direct associationwith the sites on which they originated. Exceptions weremade only for a few highly unsual finds incontrovertiblyrequiring salvage.

Quantitatively oriented specialists will recognize at oncethat this is an extremely modest beginning in the use ofintensive collecting methods. Work is already well ad-vanced on Uruk sites in Iranian Khuzestan that uses morenumerous, rigorously defined variables in an effort todistinguish asymmetrical exchange relationships, centersof regional variation, and patterns of functional specializa-tion among communities (Johnson 1973). Here the pri-mary objectives were more narrowly chronological, todifferentiate among patterns of settlement pertaining tothe sequent subphases of the Uruk period, although nat-urally this also implies some extension in our under-standing of other socioeconomic patterns (see chap. 3).But, as I indicated earlier, the general approach fol-lowed in southern Mesopotamian surveys was consciously

45

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

and consistently a contingent one. With prevailing uncer-tainties as to continuation of the project, the first prioritylay with extending the geographical extent of coveragewhenever trade-offs had to be made. When and if condi-tions permit restudies devoting massively increased timeto each site, the sites now known in Iraq will lavishlyrepay the efforts of investigators pursuing the same kindsof questions that are now beginning to be asked on alarge scale elsewhere. At that time, too, an extension ofthe approach fom prehistoric periods to the later, better-documented ranges of time that are characterized bygreat urban centers probably will deserve systematicapplication.

Initially, in fact, I had planned as a part of this projectto introduce a somewhat similar system of recording forSasanian and Early Islamic sites. This at least might havepermitted greater precision in the understanding of spatialand temporal aspects of a social and historical transforma-tion that was almost as important as that occurring duringthe Uruk period. Because of the subsequent permanentabandonment of so many of these sites, a further similarityto the prehistoric case, I thought it likely that their surfaceremains would be unusually rich and illuminating for thepurpose. But two circumstances combined to defeat thisintention. The first was the relative sparseness of surfaceceramics in most areas of late sites, for reasons alreadyoutlined, so that numbers of sherds adequate for mean-ingful quantitative analysis usually could not be assem-bled without increasing the radius of the collecting areato much more than 5 meters-and hence also greatly in-creasing the time needed to complete work in each col-lecting area. Second, and even more seriously, the scaleof the undertaking turned out to be grossly disproportion-ate to the resources of a single investigator. There are some36 square kilometers of Sasanian settlement alone withinthe intensively surveyed area, more than four times thearea of all known Uruk settlements.

The initial compiling of a list of Sasanian and Islamicceramic markers nevertheless did provide a basis for somerefinement in dating. A covariation in the frequency ofcertain glazed styles, appearing to have chronological sig-nificance, was tentatively established as a result of the TellAbu Sarifa excavations in 1969. That was used selectivelyduring the 1975 season of survey, on the basis of grosscounts and rough impressions reached for certain sites asa whole. Perhaps this can identify the terminal ceramicassemblage more accurately, at least when surface ma-terial is relatively abundant. On the assumption that thisis so, we have gained a few new leads to the disjunctivechanges in settlement that marked the end of the Sasanianperiod, as well as to the accelerating processes of abandon-ment that appear to have been under way by the earlyninth century. But these increasingly become substantiveissues rather than matters of survey method and henceare better left for full discussion in chapter 5.

REINTERPRETATION OF LOWER DIYALACANAL PATTERNS

Although this study is primarily concerned with theregion between the Tigris and Euphrates rivers, the sig-nificance of survey findings for that region obviously canbe better understood if it is considered in its wider geo-graphic context rather than in isolation. The larger unit ofinteraction that would be most appropriate, at least forlater historical periods characterized by increasinglydense and extensive settlement, is likely to be no lessextensive than the whole of the Mesopotamian alluvium.A considerable part of the alluvium unfortunately has notyet received any systematic archaeological surface recon-naissance, especially along the Hindiya branch of theEuphrates and in the swamps adjoining the lower Tigris.To add to our difficulties, both of those districts probablyplayed a greatly enlarged historical role in Early Islamictimes. In this respect they contrast sharply with currentlyavailable surface evidence, which is drawn mainly fromdistricts where the Arab conquest was followed not by anexpansion of the frontiers of cultivation but by their re-traction. Lacunae in our detailed knowledge about trendsof settlement thus are very likely to introduce distortionsinto any general reconstructions of historic populationtrends, at least insofar as those reconstructions are basedprimarily on the Mesopotamian archaeological evidence.

The plains adjoining the lower Diyala River, east ofBaghdad, serve as a source of data with which we canpartially circumvent this danger. On the whole marginaland of secondary political importance until the last fewcenturies B.C., their subsequent place is indicated by thefounding there of the Parthian and Sasanian capital ofCtesiphon and, on a still more grandiose urban scale, ofBaghdad itself. If the concept of an urban heartland inMesopotamia can usefully be extended from that civiliza-tion's formative era into the past two millennia or so, thenfor that later range of time the lower Diyala plains wouldhave to be regarded as no longer marginal but instead asconstituting a central part of it.

An extensive archaeological reconnaissance of this re-gion has been carried out and published (Adams 1965),and to summarize its procedures or recapitulate the sub-stance of its findings in this volume would be redundant.However, the major and final phase of that fieldwork wascarried out in 1957-58. At that time the full utility ofaerial photographs for understanding changing ancientcanal and watercourse patterns had not yet been per-ceived; only during later years were the procedures fortheir study that have been described above graduallyelaborated. The older, less-developed method has littleeffect on reconstructions of irrigation patterns along thelower Diyala before the latter part of the first millen-nium B.C. at the earliest. Alluviation was relatively muchmore rapid than on the central floodplain of the Eu-

46

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

phrates, and there was no comparable process of Middleand Late Islamic wind erosion to expose long-buriedsediments providing traces of ancient hydrographic pat-terns. But for the later periods the more recently elab-orated methods of utilizing aerial photographs hold thepotential of providing much new information.

Figure 8 summarizes this new information, in a formreasonably comparable to the base map for the currentstudy. The photographic coverage on which it is based isinferior in quality to the coverage employed for the majorpart of this study. There has obviously also been no oppor-tunity for iterative improvement of impressions initiallyderived from close study of the photographs, guiding field-work on that basis and then in turn refining the photo-graphic interpretations with field observations. And onemust further bear in mind that the greater part of theDiyala region was once again under cultivation at the timeof the original study (Adams 1965, fig. 1), curtailing bothphotographic and archaeological surface recovery in waysalready suggested. But the map given in figure 8 nonethe-less permits a substantially fuller interpretation of thelater irrigation sequence. While the numbered sites andtheir locations and dating are identical to those previouslypublished, the new data permit the main stages in the de-velopment of the accompanying irrigation system to bemuch more clearly and unambiguously distinguished. Asis discussed more fully in chapter 5, this discloses a num-ber of interesting contrasts between trends of settlementand land use in the oldest urban heartland and those in atleast one of the regions that succeeded it in this role.

MAJOR LIMITATIONS OF THIS SURVEY ASA SOURCE FOR SETTLEMENT HISTORY

It is perhaps worthwhile, in concluding this lengthydiscussion of methodological issues, to draw togetherwhat appear to be the principal limitations of at least thistype of survey approach. Several that become apparentdirectly from a description of its procedures have alreadybeen suggested. Others will emerge from details given inthe appendixes to each of the three succeeding chapters,where the descriptive data on sites and dating criteriapertinent to various periods have been concentrated topermit a more connected flow of narrative within thechapters themselves. And still others, especially thoseconcerning estimates of population size and density andevaluations of the extent to which there was continuityof occupancy at particular sites, appear only when thedata are utilized for analytical purposes. The cumulativeweight of these considerations, here only briefly summa-rized, should be evident at once. Together they will con-stitute, I hope, a sharp, continuing reminder of the pro-visional quality of much of the synthetic reconstructionthat follows. The reader should note, therefore, that thislisting of them is intended to obviate the long and tire-

some succession of cautionary or qualifying phrases thatwould otherwise thickly intersperse the main account con-tained in the next three chapters.

To begin with the most obvious point, this surfacereconnaissance was not accompanied by a program ofsoundings in which inferences from surface collectionsand observations could be checked and refined againstexcavated data from a sample of the same sites. Hence itcan only provide conditional and unsystematic confirma-tion for the more or less plausible hypotheses generatedfrom it. But, while that is unavoidable, it should not be adeterrent to further consideration of either the hypothesesor the supporting data. What is involved is a view of sci-entific investigation not as a pristine discovery of certaintybut as a cumulative, unfolding process in which thetesting of hypotheses and the search for new data areclosely linked. Very few new findings of any kind areimmutable and decisive. We engage in the collective, andtherefore (in principle at least) cumulatively self-correctingact of pushing back the frontiers of contingency that con-strain knowledge-only to discover that the contingenciesand constraints assume new shapes as they recede beforeus. A survey approach, in other words, differs not in kindbut only in degree from the promises and limitations of allother approaches available to the archaeologist or ancienthistorian.

A second more considerable problem is raised by un-systematic aspects of the data base to which detailed ref-erence has already been made. Continuing uncertainties asto the future of the operation dictated an emphasis on abroad-ranging preliminary investigation rather than onexhaustively detailed gridding of small areas, and differ-ences in coverage also arose from the incidence of culti-vation and from the many topographic irregularities. Stillanother problem of the same kind stems from the fact thatthe work had to be conducted over a seven-year period,during which there were some cumulative improvementsbut perhaps other subtle shifts in method or perceptionthat took place too slowly to be noticed and that mighthave been avoided had it been possible to complete theinvestigation without interruptions. Greater emphasisfrom the outset on systematic sampling and exhaustiverecord-taking would have reduced these biases, but itwould not have eliminated them. Moreover, a greaterinvestment of time in sampling and recording could havebeen made only at a sacrifice in the extent of coverage.That trade-off appeared, and still appears, injudicious inthe circumstances, but it does undeniably qualify the com-pleteness of the findings.

Third is the silence with regard to function that charac-terizes most of the data. Surface indications support anoccasional suggestion about specialized activity at a par-ticular site-for example, walled enclosures that may haveserved as fortified khans or storage facilities, sites pre-ponderantly composed of the debris of glass- or brick-

47

oi.uchicago.edu

oi.uchicago.edu

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

making, clay cones suggesting wall mosaics on earlypublic buildings, remains of pottery kilns, or unusual con-centrations of chipped stone tools. But such observationspertain to only a tiny minority of the sites cataloged inchapter 7. Were the remainder undifferentiated as to func-tion, or differentiated in no important respect but that thelarger ones contained progressively increasing proportionsof nonagricultural specialists? This seems hardly likely,which can only mean that the available data leave usignorant of many differently specialized communities.The tendency exists and must be resisted, as a conse-quence, to see a homogeneous field of autocthonous, agri-culturally oriented cells, rather than the web of supra-subsistence interrelationships by which the whole regionalstructure was articulated.

Fourth, the specialist will no doubt note cases in whichthe ceramic and other dating criteria relied upon through-out this study do not correspond fully with the temporalunits into which the analysis is divided. There are refer-ences to a number of these apparent discrepancies in theappendixes to chapters 3, 4, and 5, and others will surelyappear from time to time as a result of newly published-not necessarily the same as recent-excavations. It isfrankly rather unlikely that a whole set of typological fea-tures, sufficiently large and popular to identify a phaseof settlement history in surface collections, should haveexactly the same times of origin and disappearance. Evenwhere that may have occurred, as with the advent of mass-production methods in protohistoric times and again dur-ing much later periods of widespread diffusion or experi-mentation affecting new decorative techniques like glaz-ing, the correspondence of typologically defined periodswith periods demarcated on largely political grounds isstill less likely. Surface dating criteria must be recognizedfor what they are: hypotheses, of a plainly approximateand provisional character, whose justification rests partlyon the currently available findings from stratigraphic ex-cavations and partly on the meaning and coherence thatparticular criteria seem to introduce into the interpreta-tion of changing settlement patterns. Ultimately, we canretain only those criteria that are definitely found to cor-respond to a single temporal unit or that still are usefulfor some purposes even though they span two or moresuch units. The test in the long run, in other words, willcome only with excavations.

But not just any excavations, I hasten to add. Particu-larly in historic ranges of time, Near Eastern archaeologystill sometimes involves work of depressingly primitivestandards. Since there is widespread reliance on localpickmen with inadequate numbers of supervisors, fullquantitative control is very often not kept over sherdfrequencies. "Full" recording (subject to the skill and at-tentiveness of the pickman) of whole vessels and completeprofiles is emphatically not an adequate substitute, sinceit generally leaves us unaware of differential patterns of

breakage and discard in antiquity. Worse still, in the ab-sence of careful, continuous control it is often not possibleto distinguish ceramic assemblages that are reasonablylikely to represent a single period of use. Debris along anancient floor meets this qualification. Sherds that mayhave been included in collapsed walls above such a floordo not, nor does material associated with floors that havebeen badly cut by intrusive pits. The spade, not the sur-face collection, is ultimately the arbiter of the sequence,but only when certain minimal standards have been met.Until that time, and explicitly disregarding reports thatconvey a subjective and unverifiable impression of thestandards as well as the findings of excavations, the datinghypotheses that occasionally have had to be introduced inthis study may well continue to be as valid as any other.

A fifth area of difficulty involves our inability to dealwith settlement size except in the rather gross terms ofphysical dimensions of sites. Population is the index whosevariability we seek, and we can indeed reach crude ap-proximations if assumptions about constant density areapplied to site dimensions. But available demographicfindings from the modern Middle East make it clear thatdensities are not at all uniform (cf. below, chap. 4, n. 1),without establishing what archaeologically recognizablefactors can explain the variability. Density in privatedwelling areas may increase with increasing populationsize, for example, but this effect can well be completelyreversed if monumental buildings and public spaces aredisproportionately concentrated in the larger urban cen-ters. Further complicating the problem, at least underMesopotamiam conditions, are local differences in alluvi-ation and erosion that affect site surfaces. Calculationsbased on sherd densities might appear to be a useful re-finement of population estimates dependent on site sizealone, but in fact the density of debris appears to be largelydependent on natural processes subsequent to a mound'sabandonment. Hence we are left with hierarchies of mea-sured site sizes that surely have some relationship, but notnecessarily a very useful one, to hierarchies based on an-cient populations. Any discussion of breaks or tiers inhierarchies of central places, an important analytical ele-ment in many formal approaches to the interpretation ofsurvey data, must take place through this screen ofimprecision.

Sixth, ambiguities of a different kind are inherent inderiving ancient population estimates from archaeologicalsurvey data. Insofar as the recognizably distinctive sets ofdating criteria that have been employed do not define asequence of periods of uniform length, it is an open ques-tion whether the population estimates made from re-corded areas of occupation during each of them are di-rectly comparable. At the extreme, for example, would thetotal population of fifty sites aggregating five hundredhectares of settlement that were occupied during (un-known, presumably different) portions of a five-century

50

oi.uchicago.edu

The Recovery of Ancient Settlement and Irrigation Patterns

span really be equivalent to the total population of sitesof approximately the same surface area that can be as-signed to a one-century span? Or, to introduce anotherfacet of the same question, does a population estimate fora given site that is derived from the maximum area it at-tained during an unknown, possibly brief portion of afive-century period correspond as closely to the averagepopulation of that settlement over the whole half-millen-nium as it would if the known period of occupation werelimited to a century? The answer, in both cases, is proba-bly not. For the group of fifty sites, it is not unreasonableto suppose that there would be less simultaneity of occu-pancy during the longer, half-millennium period. Somesites are even likely to have succeeded others that hadbeen abandoned earlier in the same interval, so that theirinhabitants would be "counted twice" in estimates of re-gional population.1 For the single site, whose populationalso presumably varied with time, lengthening the periodwould similarly tend to increase the difference between itsmaximum extent of occupation and the average. Theseconsiderations recently have been carefully reviewed byHarvey Weiss (1977, pp. 357-59), who concludes thatthey require standardizing the length of the periods insome way if accurate assessments of population trendsare to be made.

While he is probably right in this general prescription,the weakness of the available data makes the course he ad-vocates difficult to follow. Periods assignable to groups ofdating criteria (that in any case waxed and waned indi-vidually in frequency) are notoriously difficult to datewithout large margins of arbitrariness and uncertainty.Unrealistically "uniformitarian" assumptions are implicitin any attempt to standardize period lengths, and theseare at times directly contradicted by the historical record.In particular, they often lead to serious underestimates ofregional population densities that were at least brieflyattained during some very long intervals like the Seleucid-Parthian period (see below, pp. 196-97). A less rigorousprocedure than Weiss advocates accordingly seems morereasonable, at least for the present, in which it has been

possible in most cases to designate periods of roughlysimilar length but in which period length is not otherwisetaken into account.2 Some further series of correctionsultimately may be necessary, if and when it becomes ap-parent that the successive sets of dating criteria used hereare identified with periods of widely varying length. Butfor the moment it seems possible only to acknowledge ex-plicitly the difficulties that may lie in this direction, andthen to set them aside for later study.

Finally, as discussed in greater detail in the appendicesto chapters 3, 4, and 5, chronological grossness is an im-portant barrier to improved understanding that simplycannot be overcome at present. In most cases the availabledata do not permit us to place a particular surface collec-tion in a span narrower than something on the order often human generations. To speak of processes of historicalchange within such a constraint is to deal with changes soslow and amorphous that they may have entirely escapedthe conscious notice of individuals taking part in them.Changes visible in settlement patterns therefore constitutea set of phenomena that in most periods are almost en-tirely disjunctive from the phenomena with which histori-cal and other written records deal. But to that the archae-ologist is justified in replying that they are no less "real"for the lack of any felt and communicated sense of im-mediacy. To be sure, he must avoid unconsciously impos-ing a "gradualist" bias upon his findings-for example,the assumption that site occupations were generally sta-ble and population trends consistent over long periodssimply because he cannot easily detect volatile, quicklyreversible patterns. But there is also no inherent hier-archy in which immediate events take precedence oversystemic processes that penetrate more deeply. The taskof historical synthesis, as Braudel has persuasively shown,is one of working vertically through a succession of super-imposed strata: the enduring, the slowly changing, andonly last the evanescent flux that dominates the minds ofprotagonists of a passing generation or moment (Braudel1972, 1: 20-21; 1973, p. ix).

51

oi.uchicago.edu

3Urban Origins

(Ubaid-Early Dynastic I Periods)

The emphasis now shifts from environmental and meth-odological considerations to an analysis of the primaryrecord of human settlement on the Mesopotamian flood-plain. It is important at the outset, however, to note oncemore several prominent themes of earlier chapters that wemust continue to take into account as the focus of in-terpretation changes.

To begin with certain features of physical geography,the present serves as a guide to the ancient past only inan exceptionally limited way. The courses of the rivershave moved repeatedly and are surely continuing to do so,though modern technology probably can rectify or evenforestall major, disastrous shifts and limit most futurechange to slower processes of meander-cutting. Since theentire plain is composed of silts, differentiated only bycoarseness or fineness of texture, there is absolutely noreason to assume that the major courses or branches oftoday have any close relationship to those of earlier pe-riods. The only basis for defining the latter, then, is whatcan be empirically demonstrated through historically ori-ented investigations of soils, landforms, or, as in thisstudy, the remains of ancient settlement.

Similarly, even the physical limits of the plain are amatter on which the present provides no reliable guideto the past. Earlier assumptions that one could recon-struct the past by assuming a regular rate of infilling atthe head of the Gulf, and a correspondingly steady rate ofadvance in the land available for human settlement, wereshown a generation ago to be simplistic in their failure totake into account the inherent tectonic instability of theMesopotamian geosyncline. More recently, the supposedeffects of such instability upon post-Pleistocene landformshave been cast into doubt, but in a context that stressesthe complex interactions of many other factors and soprovides no encouragement for a return to the earlier as-sumption of smooth and easily calculable regularity.

But if the layout of physical features in earlier periodsbears no necessary relationship to that of the present, theecological, hydrologic, and geomorphic processes that canbe observed at work only in the present are crucial forundestanding the past. The dynamics of stream flow, leveeformation, and soil salinization must have been implaca-bly the same in every premodern epoch, permitting onlymodest human moderation of their effects and in themain imposing a relatively fixed set of requirements andperiodicities on agricultural and urban life. The proces-sual perspective that emerges particularly stresses the needfor cultural adaptations to risk and uncertainty connected

Swith both land and water supply. Much of this must al-ways have been perceived as environmentally imposed andoutside human influence-disastrous floods and watershortages, channel siltation and subsequent deflectionaway from settled areas, loss of productive fields throughthe rise of saline groundwater into the root zone. But in adeeper sense, as we have seen, the dangers confrontinghuman societies in the Mesopotamian setting are not tobe understood solely as external impositions of the en-vironment. Instead they are in important respects pro-ducts of long-term interaction of those societies with theenvironment, in which dangers or deleterious changes,traditional subsistence practices, institutional forms, andenvironmentally constrained decision-making composean interacting web of cause and effect.

A second finding that must strongly tincture the an-alysis that is to follow runs in a direction that may in-itially seem somewhat contrary to the first. There is abroad plasticity or substitutability among viable adapta-tions to the conditions imposed by the Mesopotamianclimate, soils, vegetation, and river systems, and there isevery reason to expect not only that alternative adapta-tions coexisted in different parts of the zone but that theyfluctuated widely in their respective importance.

52

oi.uchicago.edu

Urban Origins

Large-scale irrigation agriculture, "dense" not only inthe population it could support but in its managerialand capital requirements, accordingly should not to beregarded as the historic culmination toward which somesort of inner logic or driving force led the whole systemto evolve naturally. Inherently complex and lacking eco-logical resilience, it appeared relatively late in the his-torical record. Even when it did appear, it did noteradicate alternative forms of adaptation but only pre-dominated for a spectacular but relatively brief interval.

y Pastoralism, a way of life devoted preponderantly toherding, was of course the other extreme of the con-tinuum. Within the Mesopotamian plain itself, it wasgenerally relied upon only by relatively small, marginalgroups except during certain "dark ages" of politicaldissolution. Much more important were a variety ofadaptations stressing a shifting balance of husbandry andcultivation, for they could usually support far greaternumbers than unadulterated pastoralism while maintain-ing, at the expense of some loss of productivity, greaterlong-term resilience than specialized cultivation. Theseminomadic, "tribal" units pursuing these shifting pat-terns may frequently have had centrifugal, antiurbantendencies, as when they can be observed in most con-vincing detail in the nineteenth century. But howevereffectively they were incorporated within larger, urban-centered polities, they remind us of the range of adapta-tions that was always possible.

Another way of taking cognizance of the long-con-tinuing range of subsistence and social variability is toemphasize the region as the unit of this analysis. Archivesand monuments that are concentrated in the major townsfoster a predisposition to begin with a particular site orcity as the paradigmatic unit, viewing its hinterlands onlyas they sustain its life processes as an independent or-ganism. Here the quite different starting point is a se-quence of patterns of settlement in the countryside,geographically bounded by the disappearance of thosepatterns beyond the limits to which irrigation watercould be conducted in order to sustain an at least brieflysedentary community. This difference in perspective doesnot negate the importance of cities as organizational andpower centers and as the principal loci of most forms ofcultural initiative. But placing them in a regional frameof analysis lays greater stress on cities, towns, villages,and encampments as they compose an interacting sys-tem, fluctuating in size and prosperity and relying upondifferent parts of the spectrum of subsistence resourcesnot wholly by independent choice but at least partly asan outcome of their changing relations with one another.

The existence of diversity with regard to subsistenceadaptations is paralleled by similar diversity in topog-raphy and geomorphic processes. In the traditional ap-proach, alluviation has been seen as a uniformly dominantif not necessarily constant process. Upon closer inspection

this idea requires replacement, as we have seen, by apicture including a complex, sometimes slow and some-times very rapid, array of erosional as well as depositionalprocesses. The locally variable outcome of these processes,which was often in evidence but impossible to record indetail during the course of a rapid, wide-ranging archae-ological surface reconnaissance, introduces unmeasuredbut presumably large uncertainties into the findings ofthat reconnaissance. This must have some effect on esti-mates of site areas, since in different circumstances greateror lesser proportions of the sloping lower flanks ofmounds would be buried and hidden from view. Morevulnerable still is the proportion of originally existingsmall, low sites that can be found and recorded withpresent survey techniques. As I noted earlier in connec-tion with a restudy of small prehistoric sites in the Nippurarea, even local wind deposition owing to a few wet yearswith correspondingly rapid plant growth can have a veryadverse effect on the rate of site recovery.

Uncontrollable variation in the quality and quantityof data recovered is a problem not limited to archae-ological surveys. Short of exceptionally costly and time-consuming programs of sampling, it extends with almostequal force to archaeological excavations. A variant ofthe same problem can be said to lie at the root of mostcontroversies over the use of ancient textual sources forthe reconstruction of broad patterns of belief and be-havior. There is also much truth in the response that thebasic features of a pattern can often be recognized evenwhen a very large number of elements are missing in thetotality.

But in the present context this unevenness of data muststill be recognized as a deterrent to the use of analyticalapproaches that elsewhere have proved very useful. If ahigher proportion of small than of large sites may havebeen lost in some areas, for example, then differences inthe proportion of small to large sites are highly suspectas data. Yet in classic central place theory, evaluationsof the relative importance of "market," "transport," and"administrative" considerations depend on those propor-tions in no small part. Again, the use of Thyssen polygontechniques to generate bounded spatial units suggestiveof ancient territorial patterning is also compromised,since the unsuspected omission of a larger center (or itsconversion to a settlement of significantly lower rank)can produce a quite spurious outcome. Similarly, nearest-neighbor analysis, potentially useful in providing coeffi-cients of relative settlement clustering from measurementsof distances between sites, is of doubtful validity whenthe proportion of sites from which such measurementscan be taken varies uncontrollably. These are among theformal locational approaches to the study of archaeologi-cal settlement patterns that in other circumstances havebegun to provide important insights not merely intosystems of land use but into questions of social differentia-

53

oi.uchicago.edu

Urban Origins

tion and administrative hierarchy (cf., e.g., Hodder andOrton 1976; Flannery 1976). But greater hesitation seemsjustified in applying them here.'

Finally, constraints on locational studies of anothersort are imposed by a feature of the settlement history ofthe area. The major early centers tended to be very long-lived, having originated in the fourth millennium B.c.or earlier and in most cases surviving at least into thesecond. Within the limits of a brief reconnaissance, multi-period occupations of this kind make estimates of sizeat any given period frequently unreliable if not wholly.impossible, save for the terminal occupation and in somecases perhaps also an earlier, more widely extended one.Hence many of the data on the sizes of the key centralplaces around which the early urban settlement systemwas organized are impressionistic or speculative. Ob-stacles to completeness or precision in this respect seri-ously undermine efforts to appraise the institutional cor-relates of the settlement system on the basis of its urbanhierarchy.

ONSET OF ALLUVIAL SETTLEMENT

The Arabo-Persian Gulf may be described as a verylarge and long but relatively shallow estuary. Hence theshoreline at its head is peculiarly sensitive to oscillationsin world sea level occasioned by climatic changes. No-where deeper than 100 meters, the Gulf in fact disap-peared entirely as a geographic feature for a time duringthe late Pleistocene. As recently as about fifteen thousandyears ago, the ancestors of the present Tigris andEuphrates emptied directly into the Gulf of Oman, some800 kilometers southeast of the present mouth of theShatt al-Arab. By six or seven thousand years ago, onthe other hand, the melting of ice accompanying irregu-lar but progressive warming trends had brought sea levelsto approximately their present position (Niitzel 1975).

As we saw in an earlier chapter, the position of theGulf shoreline is not so easily established. The earlierargument of Lees and Falcon (1952) that the line of thedepression occupied by the Gulf constituted an unstable,subsiding geosyncline was for a number of years widelyregarded as persuasive. More recently, however, evidencehas begun to accumulate in support of a substantialnorthward marine transgression after the end of thePleistocene. It is not yet possible, to be sure, to define theposition of the shoreline itself at any given period. Buta southward progradation of as much as 150 to 180kilometers during the last five thousand years or so seemsincreasingly likely. The assumptions of an earlier gen-eration of archaeologists thus are finding favor oncemore, even if it is premature to speak of their full con-firmation (Larsen 1975, p. 57; Larsen and Evans 1978,p. 239).

It is important to recognize the effect of the large-scale

physiographic changes just described upon early pat-terns of settlement in the alluvium. In terminal Pleistocenetimes, under climatic conditions that locally may nothave been much less arid than those of today and witha much greater exposed alluvial land surface until theGulf approached its present position, the whole regionwould have offered few attractions to human groups.Vita-Finzi has summarized evidence of dune developmentand the paucity of fluvial sediment on what is now theGulf floor to contradict the thesis that it must have beena fertile, well-watered plain. He suggests instead that itshould be understood as "a generally waterless depressioncontaining a few swampy tracts" (1978, p. 258). Theonly zone of substantial preagricultural potential for acomplementarity of resources permitting year-round sub-

y sistence, in fact, would have been the tidal marshes andlagoons concentrated immediately above the head of theGulf or along its shoreline. Early encampments therewould, of course, now be far out on the bed of the Gulf.Moreover, the rapid retreat of the shoreline as the sealevel rose would have approached an average of 100meters per year, although there were intervals whenconditions temporarily stabilized. This means that mostencampments of hunter-gatherers could have remained inplace only for a relatively short time rather than beingreturned to seasonally.

The retreating Gulf shoreline must have reached itsnorthernmost limits in early post-Pleistocene times, andthe southward progradation that then ensued was surelya much slower process. By around the sixth millenniumB.C., in other words, conditions favoring permanency ofnonagricultural settlement in regions accessible to archae-

> ological inquiry markedly improved. This was roughlythe same time that the advent of irrigation permanentlytransformed subsistence patterns.

Although this reconstruction is based exclusively ongeological data, it accords well with the available archae-ological evidence. Settlements in the alluvium properthat antedate the mid-sixth millennium or so have notyet been found, and the earliest ones appear to clusterin the extreme south of the plain, close to the putativeposition of the shoreline. The terminus post quem forthis study is thus a fairly firm one of about eight millenniaago. And it is noteworthy that the changed conditionsresponsible for the onset of settlement had less to do withsome advance in man's organizational or technologicalcapabilities than with conditions largely or wholly beyondthe control of human societies.

In archaeological terms, the time of onset of humansettlement can be said to coincide roughly with the Ubaid

' period. That period as usually defined is excessively longand somewhat heterogeneous, its identification restinglargely on a succession of painted pottery styles whosecommon elements are not altogether apparent. Hencethe practice followed by some authorities (including

54

oi.uchicago.edu

Urban Origins

Henry Wright in this volume) of referring to Ubaid I andII as the Eridu and the Hajji Muhammad periods maywell be preferable. In any case, most excavated samplesfrom within the alluvium pertaining to the Ubaid I-IVor Eridu-Hajji Muhammad-Early/Late/Terminal Ubaidrange were obtained a generation ago, before importantadvances in excavation technique and the developmentof interest in floral and faunal remains and other non-artifactual materials. It does seem reasonably clear, how-ever, that the material spans most or all of the fifth mil-lennium, the crucial range of time in this regard, with afew of the earliest known settlements apparently showinggreater affinities with the preceding Samarran period andperhaps extending back as far as the middle of the sixthmillennium.

The known sites for this very long span of time areplotted in figures 9 and 10. Most of them seemingly wereoccupied only for part of it, although detailed informationis available on so few that no attempt has been made tospecify subphases at particular sites. Particularly whenwe consider the high probability of prevailingly sequentrather than simultaneous occupancy, the dominant im-pression is of extremely low population and settlementdensity. That impression needs qualification, to be sure,on the basis of regional differences in the available data.

As I described more fully in chapter 2, the data areprobably most deficient in the region of ancient Akkad,the northern part of the alluvial plain between the Tigrisand Euphrates (Gibson 1972; Adams 1972). Much of itwas surveyed before aerial photographs were available.Widespread cultivation interfered with survey coverage,which in any case was rendered less effective by heavy,recent alluviation. Notes and photographs of the originalcollections have been reevaluated in the light of morerecent dating criteria, but what was not originally recog-nized as a potentially significant type for dating and hencenot collected is of course not available for reassessment.A glance at figure 9 conveys a graphic impression of thesparseness of known pre- and protohistoric sites to thenorthwest of Tell Abu Salabikh, in the Akkad region, incomparison with more recently studied areas to the south-east. Since at least in immediately adjacent areas condi-tions must be approximately the same, it is evident thata substantially lower proportion of Ubaid sites have beenrecovered in ancient Akkad.

Northward across the Tigris River are the plains ad-joining the lower course of its last major tributary, theDiyala. Ubaid and other pre- and protohistoric sites thereare shown in figure 10, drawn from slightly later surveysin 1957-58 that were in turn based on darlier OrientalInstitute campaigns in the thirties (Adams 1965). Theoriginal designation of what was regarded as Ubaid inthis area included certain important types, principallyclay sickles, that later were found to be far more numerousand important hallmarks of the succeeding Uruk period. i

What is identified as Ubaid in figure 10 excludes sites forwhich the original dating criteria now seem questionable,again based on a restudy of the original notes and col-lection photographs. Here again the existence of a heavyblanket of more recent alluvial sediments is well docu-mented, and extended areas of cultivation were a de-terrent to systematic coverage.

Excluding recovery procedures employed in theNippur-Isin-Adab region that are discussed in detail inchapter 2, it remains to consider the probable proportionof recovery of Ubaid sites in the region around ancientUruk (modern Warka) farther to the south. In general,as I noted above, reconnaissance techniques were verysimilar to those employed more recently. A significantexception, however, was that somewhat less time was de-voted to systematic collections on the early sites. Whetherfor this reason or for some other, there is greater likeli-hood that small numbers of Ubaid and earlier sherds mayhave been overlooked in the Warka survey. Five Urukperiod sites visited initially in that survey were reex-amined in 1975 using the newer, more intensive pro-cedures; Ubaid occupations that had not been notedinitially were found at two of them (126, 245). Thesample is inadequate to indicate by how much the pre-viously reported number of Ubaid and earlier sites wouldhave been increased if all the early sites had been re-visited, but clearly the modification in the widely dis-persed, sparse pattern that was initially reported (Adamsand Nissen 1972, pp. 9-11) might be substantial.

Taking into account these regional variations in cover-age, what does the known distribution of Ubaid andearlier sites imply? Acknowledging that the initial im-pression of very low density needs correction, especiallyfor the region around ancient Uruk, it is still essentiallysupported by all the available data. To take the Nippur-Isin-Adab area, for example, sparse traces of Ubaid potterywere found only at four widely separated sites (573, 680,1194, and 1416) out of more than eleven hundred thatwere recorded. This was an area, moreover, in muchof which wind erosion had greatly reduced the effectsof subsequent sedimentation, and to all of which therelatively more intensive collecting techniques were ap-plied. When in addition we take into account the verylong span of time represented by the early painted potterytraditions, the settled occupation in southern Mesopo-tamia as a whole at any one time during this span seemslikely to have been far less than it ever was subsequently.

We should also recognize, however, that there is sig-nificant regional and temporal variation within the gen-eral pattern of low population density. First, entirelydifferent conditions obtained along the Zagros piedmont.Significantly higher rainfall there, together with numeroussmall springs and alluvial fans, offered attractive condi-tions for early, continuing, and relatively dense settle-ment. Recent work in the vicinity of Mandali has brought

55

oi.uchicago.edu

1 Pre-Uruk2 Uruk3 Jemdet Nasr4 Early Dynastic I( ) Reduced or doubtful occupation* Trace of occupation

(23)4

(23)4

± (12)

(23A

234

.2

023)4

Ab.1234

+

(32 ) 233 (3 (

.23 +

^-T

2 , (3)4 X23)4

S 234

23)4 )4 23 (04 ) ) 22.4

(43 (3*(3) ('223) 2 3 3 2

323

223 (232342233 ( 223

S 23. 2 232

2

2-,2 3 23-3

U3. 2(3)

K 4

33 ? 2(3 .024 '23 23 2 " 2" ) 2.

S22 34 3 (3)

4 3ý 2- 0 12'-23 23 2ý*

- (23' 2 23"'3 2

(234 * . (23)" 2') 234 2 233\?23 . 3.4 23(334

( 23) - (.2 , 234 3.4(23. \ .3 (23-3

2,"-234 3 3 (23.*234 2

2' . 23 ( 2

0 25 Km.IllOMEN(3)4,

Fig. 9. Pre- and protohistoric sites on the Mesopotamian plain.

56

(23)4

X

234

(23)4

(23)'

23

Jo.ndtN. r 0)22ý

(2 ) (23)

'Amiy(

0KISHt2349

+

IJ

234

2(3)4%

+

(234)

2'

.2

MIPPUR

(34)

12,

oi.uchicago.edu

+ +h

(4)

(3)4.

(3)4-

(34)-

÷(34)"

-(34)

(34)0 (3.4)

(34)

(3)4

.(34)

(34)'

(3)4 -

(3.4)

4- (3)4' .(3)4

"(3)4

1 Ubaid2 Uruk3 Jemdet Nasr4 Early Dynastic I( ) Reduced or doubtful occupation(Dating revised according to 1975Iraq Survey ceramic criteria)

(34) .(34)

(34)" ,(34)4-

4*

,4

.(34)

2(34) * .2

2..2

2(34)*

I-4*

*2(3)4

.1-44)

.(2) 4(34)8

*(34)

2(34)*

0 5 25 Km.L-- a m II

(34).

+- -

0234

2340

Fig. 10. Pre- and protohistoric sites on the lower Diyala plain.

57

.2(34)

(34)"

*(4)

e4

4*

"(34)

Asmo r01-4

34)'.4

(3.4 4 .(34)

(3)4'234 ,-4

2

Agrab1-4

(34)

40

oi.uchicago.edu

Urban Origins

to light clusters of villages that already had come todepend in part on local irrigation systems by the earlysixth millennium B.C., and whose pottery reflects somecontact with the earliest known sites in southern Iraq(Oates 1968, 1973; Oates and Oates 1977). This is animportant chapter in Mesopotamian prehistory, but onethat must be held distinct from developments in thealluvium itself.

There are difficulties of another order in obtaining anaccurate impression of differences in population densityeven within the alluvium. Less comprehensive site re-cording on the lower Diyala plain and in ancient Akkadmay falsely heighten contrasts with the south. But evenwithin the more recently surveyed areas, there is an ob-vious upward gradient in site density as one moves south-ward from the region around ancient Nippur, Isin, andAdab into the environs of Uruk. As adumbrated earlier,the oldest known sites are found in the south. Eridu,the best known of them and the type-site for the earliestphase of Ubaid pottery, is very near the southern marginof the alluvium. Two others, approximately of the sameage or possibly slightly older, are found a short distancenorth of Uruk (298 and 1604).

Recognition of the succeeding Hajji Muhammad orUbaid II phase is complicated because the distinctivepainted ceramic style associated with it at the type siteand elsewhere apparently continues well into or eventhrough the Ubaid III phase (Oates 1960, p. 36). Henceour finding individual surface sherds exhibiting this styleon later sites does not ordinarily permit an assignment toone or the other. But the Hajji Muhammad style, in anycase, is better known and much more widely found thanits predecessor. It is the first to occur in sites along andbehind the Saudi Arabian shoreline, more than 600 kilo-meters southeast of Eridu. Very recent chemical analysesindicate that the painted pottery there was of southernMesopotamian manufacture, implying periodic visits byfishermen from settlements along the Tigris-Euphratesdelta with craft sufficiently well developed for them tomaster regular deep-sea travel (Oates 1976, p. 22; Oates,Davidson, Kamilli, and McKerrell 1977).

The Hajji Muhammad style is also more widely repre-sented in Mesopotamia proper. Within the surveyed areait is most concentrated in the Uruk region (042, 051, 178,247, 267, and 298, as well as Uruk and Hajji Muhammaditself), but it also extends considerably farther north.Definitely attested at a small site directly west of Adab(1416), a single, somewhat doubtful example of the samepainted ceramic style was also found within the surveyarea on a site (1194) at approximately the latitude ofNippur. Fifty kilometers farther north, in the environs ofancient Kish, it is found once more. Hajji Muhammad isan important component of the pottery at the excavatedsite of Ras al-'Amiya (Stronach 1961), also appearingon the surface as a result of continual plowing even

though the mound itself has been completely submergedby a later blanket of alluvium. At least on the basis ofpresent evidence, however, these are the only occurrencesin the northern part of the alluvium as well as on theadjoining lower Diyala plain to the east of the Tigris.Moreover, the northernmost example, Ras al-'Amiya, isprobably to be dated in early phase III rather than phaseII (Oates 1976, p. 25). And phases III and IV of the Ubaidperiod are still more widely represented not only in thesouth, but throughout much of the more northerly re-gion that later became ancient Akkad.

, 2 This does not necessarily imply that the northern partof the alluvial plain was primarily colonized as an out-come of the slow expansion of the southern centers. Manyother sources of settlers lay immediately at hand. Afterall, one of the Near East's earliest agricultural hearthsextended all along the lower flanks of the Zagros moun-tains of western Iran and northeastern Iraq. But the dis-tribution does suggest that adaptation to settled life onthe lower Mesopotamian plains was in an importantsense distinctive and independent. It appears to have beeninitially worked out not far above the Gulf's retreatingshoreline, under conditions starkly contrasting with thosein the rainfall zone along the piedmont, where irrigationwas much less important or even unnecessary.

Several other tentative generalizations emerge from aconsideration of the limited number of early sites, mostof them having already become apparent in the study ofthe more limited region around ancient Uruk. The dis-tribution is on the whole fairly uniformly dispersed ratherthan clustered. Linear alignments suggestive of adjoiningancient watercourse levees are for the most part difficultto elicit. Where they may seem to occur in limited regions,their identification generally depends upon grouping to-gether sites that were not strictly contemporary but wereoccupied during different periods or subphases.

The frequent presence of heavy overburdens of laterdebris makes it difficult to generalize about site size. Itwas noted already in the Uruk region that, excluding Urukitself, the average size of Ubaid sites not obscured byoverlying settlements was almost 4 hectares, considerablylarger than in the following Uruk period. One site (TellAwayli, 460) exceeded 10 hectares in area by late Ubaidtimes. Taking into account the maximum dimensions ofsites with only limited (or deeply buried) Ubaid occupa-tions, on the other hand, it was also clear in the Urukarea that most of them were probably on the order of onehectare or less (Adams and Nissen 1972, p. 11).

The foregoing observations apply to the region coveredin the Warka survey, but most of them can be extendednorthward into the central or upper alluvium with littlechange. As noted earlier, the apparent density of Ubaidremains falls off sharply. Very widely dispersed ratherthan clustered distributions become even more charac-teristic. On the other hand, numerous Ubaid surface cer-

58

oi.uchicago.edu

Urban Origins

amics at sites like Nippur argue that Uruk was not alonein having been a substantial town by that date. Tell'Uqair, whose size in the late Ubaid period is unam-biguous, is slightly larger than Tell Awayli (Lloyd andSafar 1943; Adams 1972, pp. 198-99). Yet, as a group,sites of primarily later date that have a little Ubaidpottery tend to be very small.

These general features present a somewhat anomalouspicture. Both"the size of some Ubaid sites and the sophisti-cation of excavated examples of public architecture atsites like Uruk and Tell 'Uqair argue that a mature, com-plex, and successful adaptation had been made to thedemanding conditions imposed by a semiarid alluvium.

x,,But that is difficult to reconcile with the regularly dis-persed character of the settlement pattern and the ab-sence of linear alignments suggesting a riverine orienta-tion. Small, local networks of canals and even more"primitive" enclaves of flood irrigation tend to favor aclustered distribution of small settlements rather than aregularly dispersed one (see fig. 4).

What would have initiated and maintained the dis-persed pattern in spite of the prevailing ecological con-straints? What would have led to the considerableemphasis on centralization in a few sites, with social in-stitutions sufficiently formal and complex to favor thedevelopment of public architecture, in spite of the pre-vailingly very low density of population? These ques-tions raise the possibility of a major break, a disjunctivestep of some kind, between the Ubaid period and whatfollowed. Yet such a break is belied, at least in the best-studied aspects of material culture, by the apparent grad-ualness of the ceramic transition in the deep Eanna sound-ing in Uruk (von Haller 1932) and elsewhere, and bythe manifest continuities in monumental temple archi-tecture at ancient Eridu as well as Uruk.

y To resolve this apparent contradiction, fundamentalcultural continuity within the major centers of settlementseems beyond question. But they represent only a partof the picture. Their regular dispersion, and in some casesconsiderable size, may indicate that they served centralplace functions-were pilgrimage centers as well as arenasfor exchange-for substantial hinterlands. Yet an out-lying array of smaller, dependent settlements is admit-tedly not in evidence. It appears, therefore, that we havebeen able to identify only the larger, sedentary, agri-culturally based components of an interacting systemwhose members were divergently specialized across thefull spectrum of subsistence resources. Around each cen-

Ster we should visualize smaller, less sedentary groupswho depended primarily on their herds or on fishing whileexchanging some of their specialized produce for thecrops produced by the Ubaid townsmen.

Southern Sumer, as outlined in Henry Wright's ac-companying study of the Ur region, presents a substan-tially different picture. Fairly numerous small settlements,

and some clustering rather than a tendency toward uni-form dispersal, are in evidence there. Maximum site sizedoes not appear to be significantly different, but thereare suggestions of agricultural enclaves in the vicinity ofUr and Eridu whose aggregate populations exceeded any-thing that can yet be identified farther north. Possiblythe pattern of settlement around Warka was somewhatcloser to this than the brief schematic description justgiven seems to imply, particularly if we take account ofthe supplementary Ubaid remains noted on resurvey of afew sites in 1975. But a substantial attenuation in sitedensity does seem reasonably clear as one moves north-ward from the southernmost margin of the alluvium.

One explanation for this contrast might be that theinitial subsistence adaptation based on irrigation agricul-ture was indeed quite localized around the head of thePersian Gulf. The greater population density in southernSumer throughout the Ubaid period would then be merelya continuing reflection of its initial priority of develop-ment. The difficulty is that this fails to account for thepresence of substantial Ubaid towns well to the north,though only at wide intervals and perhaps only in thelatter part of the period. As an alternative, these dispersedtowns with few smaller dependencies may suggest that asubstantially larger proportion of the population aroundthem depended on pastoralism or other semisedentaryactivities. While permanent hamlets and villages for thearchaeologist to find and record would have been fewerin such circumstances, individual towns might under-standably be comparable in size, since the populationsin their immediate hinterlands were roughly equivalent.Another alternative explanation is that a much morecomplete record of the Ubaid occupation happens to beavailable for the Ur-Eridu region. Wright's impressionthat the Ubaid remains there are on the whole fairlydeeply buried, together with general similarities in surveyprocedures and intensity, makes this somewhat unlikely.At least within this group of possible explanations,therefore, the alternative involving a less sedentary andpresumably more pastoral way of life within the centraland northern part of the alluvium is perhaps most rea-sonable.

To summarize briefly, the onset of settlement is firstknown from relatively small numbers of sites that traceout a gradient of declining density northward from thehead of the Gulf. Over most of the region the sedentarycommunities were widely and fairly evenly dispersed.Later overburdens of debris, combined with uncertaintiesas to the rate of alluviation, make size estimates verydifficult. While most sites apparently were relativelysmall, however, at least a few were demonstrably quitelarge-more than 10 hectares-and were characterizedby clear indications of social differentiation and com-plexity. There is some discrepancy between this evidenceof relatively advanced sociocultural institutions, certainly

59

oi.uchicago.edu

Urban Origins

not lagging behind developments in any of the adjacent,longer settled regions, and the sparseness and dispersionof the visible remains of the population responsible forthem. This suggests that the visible remains derive in themain only from the agriculturally specialized part of thepopulation and that other groups of comparable or evenlarger size were less sedentary. If so, we may doubtwhether reliance on irrigation had yet become the basicfeature of subsistence that it was in most later periods.Its more modest role might help to explain the lengthand comparative stability of the Ubaid period, as par-ticularly suggested by the failure of population levels tobegin a dramatic rise for more than a millennium anda half. Only in the succeeding Uruk period, at any rate,can we identify the processes of precocious growth thatwere to lead to the development in southern Mesopotamiaof the world's earliest civilization. The geographic rootsof that transformation provide the unifying theme ofthe section that follows.

MASS APPEARANCE OF SEDENTARYCULTIVATORS AROUND REGIONALLY

DIFFERENTIATED HIERARCHIES OFURBAN CENTERS

The Ubaid period probably drew to a close at or soonafter the end of the fifth millennium B.c. Its absolutechronology is still obscure. There is a single Late Ubaidradiocarbon determination of 4120 ± 160 B.C. (5,570-yearhalf-life) from the base of the deep sounding in the EannaPrecinct at Uruk (Miinnich 1957) that is immediatelyrelevant. With the latest correction factor based on bristle-cone pine dendrochronology, that is equivalent to about5020 ± 170 B.C. (Ralph, Michael, and Han 1973)-anage that seems considerably too high for the archae-ological material involved. In any case, the sample istechnically very questionable. More recent and muchmore extensive archaeological research on the upperElamite plains of southwestern Iran is beginning to pro-duce a more substantial and reliable sequence of de-terminations, but even there the absolute chronology hasrecently been characterized as being "very poorly un-derstood." Susa A, partly contemporaneous with the LateUbaid period, has been relatively securely dated to about4000 B.c., followed by a Terminal Susa A phase. TheUruk period itself "probably begins in the early portionof the 4th millennium and ends somewhat prior to 3000B.C.," with suggested dates for an Early phase of 3750-3500 B.c. for a Middle phase of 3500-3300 B.c., and fora Late phase of 3300-3150 B.c. (Johnson 1976, p. 205).

Unfortunately, these provisional dates from Khuzestancannot simply be extended to the sequence in southernMesopotamia with which we are more immediately con-cerned. Precluding direct comparison, in the first place,is the fact there has been only one radiocarbon determina-

tion from relevant levels of a site in the Mesopotamianheartland. Based on wood associated with Temple Cfrom level IVa of the Eanna precinct at Uruk, this LateUruk sample probably should now be calibrated (as werethe Khuzestan dates) to read 3610 - 95 B.C.2 Althoughthe significance of a single determination must alwaysbe heavily discounted, it at least introduces the possi-bility that developments in southern Mesopotamia slightlypreceded their typological equivalents in southwestern

i'Iran. The temporal difference, if indeed there was any,was probably quite small, since the ceramic corpus inthe two areas is overwhelmingly similar. But the possibil-ity does receive some reinforcement from the absence ofIranian parallels for Early Uruk pointed-base bottles andpainted wares (see Appendix A to this chapter), sug-gesting that some Uruk ceramic styles could have droppedout of use in the Tigris-Euphrates floodplain before acognate tradition had reached its full expression in theplains adjoining the Karun, Dez, and Kharkheh riversfarther east. Susa A, it is also worth noting, has generallybeen regarded as a slightly later cognate of Late Ubaidrather than as a full equivalent (Le Breton 1957, pp. 91,94).

In antecedent periods, to be sure, the locus of primaryachievement may have been the opposite. There is littlereason to doubt that at least the sedentary part of thepopulation was larger and denser in parts of Khuzestanduring much of the Susiana sequence than it was any-where in southern Mesopotamia during roughly the sameinterval. That only dramatizes the rapidly of the processesof immigration or sedentarization or both that occurredin the latter at the very outset of the Uruk period, thepopulation there now rapidly outdistancing that ofKhuzestan and forging ahead in the construction of urbancenters also. But the more important point, in the presentcontext of a discussion of relative dating, is that Mesopo-

Stamia and Khuzestan clearly diverge in a number of re-spects in spite of their proximity. It would be unreason-able, with substantial differences in the scale and contentof the processes involved in the two regions in spite oftheir considerable stylistic homogeneity in other respects,to assume that developments within them were exactlycontemporaneous. All that can be said at this juncture isthat the Uruk period in both probably occupied all orsome large part of the first three-quarters of the fourthmillennium.

The report of survey in the Ur-Eridu region of southern-most Sumer in Henry Wright's appendix indicates a sub-stantial diminution in at least the numbers of individualsettlements after the Ubaid period. In all of the regionbetween the lower Tigris and Euphrates, however, an area

X many times larger, there was a much more striking in-crease in population between the Ubaid and Uruk periods.This can be observed in figure 9 by comparing the verynumerous sites having an Uruk occupation, symbolized

60

oi.uchicago.edu

Urban Origins

by the adjoining number 2, with the quite rare sites forwhich an earlier occupation is indicated by the number 1.As already adumbrated in the preceding section, how-ever, this map combines data obtained under very dif-ferent research conditions and procedures and henceremains very generalized. No attempt is made to dif-ferentiate phases of the Uruk period, for example, sincethat practice could not be extended consistently into theentire northwestern quadrant that was surveyed earlier.Similarly, the lesser density of Uruk sites in the northwestsurely must be regarded as an artifact of different methodsand topographic conditions to a far greater extent thanas a prima facie regional difference requiring explanation.One could make a plausible case for an attenuation ofearly settlement toward the upper, northwest end of thealluvium, linked to greater technical difficulties in intro-ducing simple irrigation systems along more dangerous vchannels and more pronounced levee backslopes there.But there is little point in doing so when the phenomenonitself is so uncertainly perceived because of the contrastivequalities of the data.

The limited number of Uruk sites known from thenorthwestern portion of the plain places another con-straint on interpretation. One can see in figure 9 thatdense clusterings of Uruk sites, probably recoverable onlyunder the conditions of long abandonment and less heavyalluviation obtaining primarily in the southern alluvium,permit us to reconstruct at least portions of the fourthmillennium system of watercourses with considerableconfidence. Portions of that system, moreover, can bedirectly traced on the ground and in aerial photographsas a result of wind erosion. Nothing approaching this isgenerally possible farther to the northwest, where, inconsequence, I have made no attempt to posit even rudi-mentary features of the watercourse system.

There is a tantalizing but poorly understood exception,immediately northwest of the important Ubaid-Uruktown of 'Uqair. The map shows in outline traces of anancient river meander in this region, taken from an airphotograph that I have had an opportunity to examineonly briefly (Adams 1972, p. 197). Considered in isola-tion, there is nothing to connect this feature with 'Uqairor to suggest the period to which it is to be assigned. Thepossibility cannot be ignored, however, of a connectionwith the more extensive but otherwise very similar tracesof ancient meanders that begin again 75 kilometers south-east and continue in the same southeasterly direction. AsI have already indicated in chapters 1 and 2, these lattertraces are attributable to a fourth millennium stream,combining some Euphrates water with some Tigris water,and it is highly probable that all the traces not only arecoeval but reflect the actions of the same system of water-courses. Absence of similar traces in the intervening areaprobably is due to a band of heavy sediments laid downby several gigantic canal systems that are known to have

taken a perpendicular course through this area in Sasanianand Islamic times.

Although present evidence does not allow the con-tinuity of the traces as parts of a single system to beconclusively demonstrated, their interconnection is mademore plausible by more or less simultaneous shifts in thetwo regions that may be linked to this river's decline anddisappearance. 'Uqair itself underwent a substantial de-cline in population soon after Ubaid times, and survey ofits immediate region suggested that "a serious decline inoccupation had commenced all along the northermost ofthe Euphrates (now to be understood as Tigris-Euphrates)branches by no later than the Jemdet Nasr period"(Adams 1972, p. 185). Reexamination of photographs ofthe original Akkad survey collections allows two over-lapping but partly sequent aspects of this process to bedistinguished. The first affected the northernmost lineor group of sites, the one closest to the present positionof the Tigris. There Early-Middle Uruk occupations aremore or less firmly attested at eight sites (Adams 1972,sites 61, 93, 119, 120, 166, 207, 211, and 215), and iso-lated finds of clay sickles at a number of other sites alongroughly the same alignment may indicate additional,smaller settlements of the same date. Only one of theseapparently continues through the late Uruk period (207),and there are also early Dynastic I occupations on thissite and one other (93). A very substantial abandonmentof the alignment after the Early Uruk period thus isindicated.

The second aspect affected the branch serving the twobest-known early sites in the region, 'Uqair and JemdetNasr. In this case we can distinguish nineteen Early-Middle Uruk sites (70, 71, 74, 92, 101, 110, 115, 149['Uqair], 151, 201, 202, 203 [Jemdet Nasr], 204, 213, 216,217, 219, 220, and from Gibson 1972, site 98). In addi-tion to 'Uqair and Jemdet Nasr, which lasted well intothe Early Dynastic period, three of these sites were cer-tainly occupied in Late Uruk times (201, 213, and Gibsonsite 98); it is unclear whether Early Dynastic I sherds atseven others (70, 71, 74, 101, 110, 217, and 220) alsoreflect continuing occupations or only later reoccupations.To summarize briefly this admittedly quite fragmentarydata, the information strongly suggests a process of pro-gressive abandonment that was most extensive and prob-ably began earliest along the northernmost of a series ofroughly parallel channels. Shortly afterward, it also af-fected at least one neighboring channel to the south. Con-ceding the doubtfulness of any interpretation from suchdata, it is striking to find that the same reconstructionapplies to much better documented changes within theboundaries of the present study.

The upper part of figure 11 reproduces all traces of theabandoned channel north and east of Nippur that couldbe identified on the available coverage in air photographs.It also illustrates all known sites of Uruk date in conven-

61

oi.uchicago.edu

Urban Origins

Fig. 11. Fourth millennium channel succession north and east of Nippur.

tional circular form, shown only in outline when there isno more than a trace of an Uruk occupation on a site pri-marily occupied during another period. Sites are classifiedby area into one of five classes: 0.1-4.0, 4.1-10.0, 10.1-20.0, 20.1-40.0, and more than 40 hectares. "E" and "L"designations have been added for those sites that clearlycan be assigned primarily to either the Early (-Middle) orLate phases of the period, rather than being occupied dur-ing both or otherwise being difficult to date unequivocally.

The archaeological and geomorphological informationbrought together in the upper part of the figure is analyzedin the lower part. Two apparently sequent watercoursepatterns are distinguished in the unretouched tracing madefrom aerial photographs before the possibility of such asequence was recognized. As an aid to chronological place-ment of the sequence, only those sites have been trans-ferred from the upper to the lower part of the figure forwhich an "Early" or a "Late" designation is possible.

It appears fairly certain from this reconstruction thatthe channel shifted over time from merely sinuous topronouncedly meandering. The index of sinuousity (chan-nel length/linear length) increases very significantly, from1.36 to 1.77. According to Karl Butzer (pers. comm.), thissuggests a pattern of maturing meanders after an initialstage of more linear channel formation. As I noted inchapter 1, the reduction of gradient associated with thisprocess under certain conditions may be associated withincreased stream flow. No such implication applies in thiscase, however, since the decrease in meander wavelengthis normally associated with a reduction rather than anincrease in discharge. A directional, internally consistentsequence of meander growth thus is indicated, withoutnecessary implications of channel capture, climaticchange, or other upstream factors affecting discharge.Subsequently it appears that flow stopped fairly abruptly.

The looser spacing and reduced amplitude of meanders

in the lower reaches of the channel, in both the earlierand the later phases, argues strongly against an explana-tion for the increasing sinuosity in terms of base levelchanges (e.g., rising or falling sea levels, on the assumptionthat the shoreline of the Gulf was then closely adjacent).Even if one assumes that the head of the Gulf in the fourthmillennium lay not far east of the terminus of the chan-nel that is shown and that sea level there changed sub-stantially during the interval when the channel was inuse, it is difficult to see how this could affect meanderpatterns in the upper reaches without visibly altering thelower channel. Hence the sequence of patterns that isshown must be attributed to the interplay of other fac-tors, such as the aforementioned evolution of a morelinear pattern toward maturity or changes in the hydro-logical regime or sediment load.

While in some respects less decisive than might behoped for, the archaeological evidence contributes sub-stantially to dating this sequence. In the first place, theplacement of many sites, particularly those of the lateUruk period, immediately alongside and even directlyover channel deposits makes it clear that the channelshown here was not in use at any time subsequent to theUruk period. That conclusion also follows, of course, frommuch more extensive evidence of overlying canal andsettlement systems. Additionally, it appears that the sec-ond, mature-meander phase of development was the onethat coincided with-or possibly even led to?-the rapidcolonization of the surrounding plain during the Early(-Middle) Uruk period. If so, the sinuous-channel phasethen must be of Ubaid date. Nothing in the availablerecord suggests that it was necessarily of long duration,which may help to explain why there is so little evidenceof an Ubaid occupation in this vicinity. Finally, to returnonce more to the placement of the Late Uruk sites, theymay afford an indication of a changing regime. Tending

62

oi.uchicago.edu

Urban Origins

to hug the terminal channel more closely that the Earlyones, some of them were in fact situated on bank depositsthat had been newly laid down as the channel moved fromits sinuous phase to its mature-meander phase. This mayargue that discharge during their period of settlement wasless subject to sharp, dangerous peaks, a general character-istic of meandering as distinguished from sinuous streams.It does not suggest, however, that siltation processes wereleading to a gradual reduction of flow. If that were thecase, the powerfully developed later meanders would oc-cupy a progressively narrower rather than wider belt.Hence the cessation of flow again appears likely to havebeen fairly abrupt and unanticipated, probably stemmingfrom a breakout of the course into a new bed that orig-inated much farther upstream.

Hypothetical though it is, this reconstruction suggeststhat natural forces contributed importantly to the dy-namics of Uruk settlement. An extensive abandonment ofthe northern part of the region after the Early (-Middle)Uruk period can be shown, and the cessation of flow of animportant watercourse seems very likely to have been atleast one precipitating agency. The attenuation of settle-ment adjoining the branch that is illustrated in figure11 is subject to misinterpretation, however, if it is viewedtoo narrowly. A comparison of changes in settlement fromEarly to Late Uruk throughout the northern part of theregion (figs. 12 and 13), extending to a much wider belt,most of which must have been watered by other branches,makes clear that only scattered outlying settlements anda few major towns escaped the general decline. This maybe interpreted as indicating that a number of watercoursesfailed simultaneously. On the other hand, the only coursethat does not reemerge as an important artery of settle-ment in the later historical record is the one whose rela-tively brief hydrographic development can be more satis-factorily traced. In all probability the completeness of thepattern shown in figure 11 is a direct consequence of theabsence of later use of this course, preserving its traceswithout overburden or alteration. But the question thatthen must be left open is whether it depicts a more generalcrisis, possibly involving some at least temporary dis-,ruption of the multiple sources of water on which thewhole region depended, or only a more localized failure.It is even conceivable that changing relationships within astill broader geographical framework played a part in theabandonment of this region. Johnson has provided evi-dence, also from surface reconnaissance, of a "major pop-ulation decline" after the Middle Uruk phase on the upperKhuzestan plains of southwestern Iran. By the equiva-lence of site area with resident population employed here,this would have involved some eight thousand persons.Failure or movement of river branches cannot have beena significant factor in the apparently quite rapid displace-ment of the human population from this ecologicallyquite different setting 300 kilometers east. In the absence

of convincing evidence for other environmental factors,Johnson's suggestion that the decrease may have beenrelated in some way to "political instability and eventualhostilities" is certainly plausible (Johnson 1973, pp. 154-55). What is implied, of course, is breakdown of settle-ment and out-migration from the region. Wholesale ex-termination is essentially unknown under conditions ofprimitive warfare.

Political instability is at least as likely to have beena significant source of the breakdown in this case, con-sidering the greater size and number of potentially com-petitive centers and the greater obstacles to stability aswell as to integration presented by a much larger region.But assuming that growing politicomilitary rivalries ineach area may have been at least a contributing factor,we must then ask whether these were wholly separate,parallel processes or whether instead they were to someextent linked. The question remains speculative in theabsence of useful archaeological leads, but nothing ofwhat is known about the advanced state of urban de-velopment in either region seems to preclude occasionalforays at a radius of military action greater than 300kilometers. Included in the repertory of motifs of theglyptic art of the Late Uruk phase are representations ofarmed figures, processions of bound prisoners, and ap-parent casualties. And the sign later identified with womenof dependent status, including slaves and hence possiblyindirectly reflective of hostilities, occurs on some of theearliest pictographic tablets from Uruk at approximatelythe same time (Johnson 1973, p. 152; Falkenstein 1936, p.57).

The general decline of population in the north that hadset in by Late Uruk times in some ways accentuated a fea-ture of settlement there that was already evident earlier.Communities of markedly different size appear to havegravitated to different zones or even river branches fromthe very beginning of the Uruk period. In the north centralpart of the northern region there are clusters of smallhamlets, villages, and towns. Evenly dispersed to thesouth of these are the much larger centers, while beyondthem lay a very large region, much of which was devoidof permanent settlement of any size.

It would be interesting to know more about the spe-cial characteristics of these larger centers, but this is apossibility in the near future only as an outcome of cur-rent British excavations at Tell Abu Salabikh. At Nippurthe pertinent levels are so deeply buried beneath laterlevels of debris that a reliable estimate even of the orderof magnitude of size of the site in Uruk times is very diffi-cult. At Adab, unexcavated since early in this century andnow submerged beneath drifting sand that makes ade-quate surface collection difficult, we cannot yet documentany occupation before the Early Dynastic period even ifone seems inherently very likely. If the hollow outline in-dicating Adab in figures 12 and 13, east-southeast of site

63

oi.uchicago.edu

A

(51

SI.

'*AbtykeSaIab ftc,

*/

N IPP\{R

on /

0

320N.

C,

-\

N

U //

1711944 o .

1172 \ 1

I LP

* - .* \

c

S 320

N

Lo

1 *

/

" N

I0".

*\ *

.. .**1 /\\ */

/ .{/

I/ -/ / N ^ 0

\/02

N .1-

N..

U

p201

'60

1 V.-LU O OCCUPATION C* 20.1-40 @ POSSIBLY LAR(* 40* OCCUPATION

@200., POLITICAL CAPITAL

p .\ ·*URUK

)OUBTFUL3E BUTJNCONFIRMED \

0 15 KM

- 6

II N, N , \

I .. \

Fig. 12. Early-Middle Uruk period settlement patterns.

64

+

1I

,\\

\

o\

H!

!I •

/

4

\1I

,0 \I

*

*0

oi.uchicago.edu

</

/

*00

&-P

<-k

S

C$

'p

-p

L0

** o \

@1237 \ . o\ \ *

\ \

* \

*

0

0

0 0

** 00 0

0 0

S o0

2* * 01o .o

0 °

a ° o°o.

0

0

&

S0.1-4.0 -* 4.1-10 HA.S10.1-20 0 OCCUPATION DOUBTFUL

0 20.1-40 ( POSSIBLY LARGE BUT0 40* OCCUPATION UNCONFIRMED

@200., POLITICAL CAPITAL 0

V°*lo

19pL

1 . .

4-

Fig. 13. Late Uruk period settlement patterns.

65

A

' 1

O

0

0

t

0 o0

o 8

15 KM

oi.uchicago.edu

Urban Origins

1237 and south-southwest of site 1306, were filled in bynew discoveries, the impression of a remarkable uni-formity of spacing among these larger Uruk centers inthe north would be considerably reinforced. Prospectiveexcavators concerned with this problem are also urged togive thought to Tell al-Hayyad, site 1306, perhaps thelargest, best preserved, most accessible site anywherewhose climax of occupation occurred during the Urukperiod.3

What is one to make of this striking zonation of settle-ment? Perhaps we can partly explain the paucity of smallvillages along the great arc between Abu Salabikh andTell al-Hayyad by assuming that large adjacent areaswere claimed by the inhabitants of the major towns fortheir own intensive cultivation. But that does not seementirely adequate; the areas left empty are many timeslarger than could possibly have been cultivated by anyreasonable projection of the population of the major cen-ters. Nor does that explanation assist us in understandingthe absence of a symmetrically placed band of smallersettlements flanking the major centers on their other,southern side.

As another part of the explanation, therefore, we mustassume that conditions did not permit a zone of agricul-tural villages south of Abu Salabikh and Nippur. Thatdoes not necessarily mean a complete absence of water inthe area, but only an absence of reasonably stable chan-nels whose hydrographic regimes and accompanyinglevees encouraged small-scale irrigation. It would be diffi-cult to imagine a set of conditions in the area before anysubstantial human impact, in fact, that could have di-verted the flow of the Euphrates entirely away from acentral segment of its alluvial plain. Instead we can bestvisualize the area as one of seasonal swamps interspersedwith steppe, of insequent, ephemeral channels, and of ex-posure to widespread, periodic flooding. The better-established river channels to the north, and the favorablysituated backslopes of their adjacent levees, would havebeen much more conducive to permanent settlement. Butthis area nonetheless could have represented an importantresource as a natural habitat for wild animals and aboveall as pasturage for flocks.

There was, then, a line of major towns along the inter-face between a region of dense cultivation and a regionnot so employed but instead useful only for pastoral andother nonsedentary activities. Those towns cannot be re-garded merely as independent organisms, explainable onthe basis of subsistence resources drawn from their ownhinterlands, for they were also specialized organs withina larger system of relationships. Specifically, they musthave served to regulate and facilitate reciprocal exchangesof the resources in which the two adjoining regions spe-cialized. Within the institutional frame of early Sumeriancivilization, this did not presuppose the existence of cen-tralized marketplaces. This, it seems, is the context within

which to understand the-admittedly, much later-mashdaria texts in the Bau Temple archive in late EarlyDynastic Girsu. Included among the household consump-tion goods mainly dealt with by these texts, in addition tothe young kids from which the term itself is derived (andfrom which in turn may be derived the later term forinterest on loans), are lambs, fish, bread, butter, oil,cheese, dates and other fruits, onions, garlic, beer, wine,and so forth. Originally interpreted simply as offerings tothe temple within the rigid framework of an all-embrac-ing Tempelwirtschaft (Deimel 1927; 1931, pp. 100-101),they are perhaps better described in more general terms asritual gifts and interchanges between groups of special-ized, primary producers and officials with a variety ofsecular as well as religious functions. Other economicfacilities and administrative agencies were probably alsogeographically diffuse and multifunctional rather than lo-calized and narrowly specialized. Moreover, it cannot be

1 forgotten that towns like these were cult centers, as Jacob-sen has pointed out in connection with his identificationof the Sumerian "high edin" of the third millennium assteppeland devoted to pasturage (1954, p. 54). Reachingout in imagery to herdsmen and fishermen as well as cul-tivators, they encouraged as well as benefited from theintercourse among strangers that could be carried forwardmore peacefully and dependably within the sanctity of agod's domain.

Turning from the major centers of the northern regionto the smaller, more densely clustered settlements north ofthem, it is surely reasonable to regard both the latter andthe districts in which they occur as more uniformly de-voted to primary agricultural activities. For the provisionof at least certain goods and services, therefore, the inhabi-tants of these districts must have been dependent on spe-cialists in the major centers. The forms and extent of ex-change or redistribution need not concern us here, al-though subsequently I will direct some discussion to thesequestions. But the existence of a gradient of specializationin any case suggests a southward movement into the largercenters of comestibles that would have contributed to thesupport of various craft, religious, administrative, andperhaps military specialists who resided there. The re-turn flow is less simply described. Included in it must havebeen certain near-necessities imported from regions fardistant from the Mespotamian alluvium, such as flint,obsidian, copper, stone, and wood (other than the soft,fast-growing poplar that was available locally) for tools.But also perhaps to be counted are urban contributionslike military protection, larger stores of reserves in theevent of famine, and the promise of sympathetic inter-vention by a deity thought to be housed in the city tem-ple-all exasperatingly intangible from the viewpoint ofarchaeological surface reconnaissance. Further compli-cating any attempt to conceptualize the pattern in termsof reciprocal flows is our uncertainty over the essential

66

oi.uchicago.edu

Urban Origins

character of the relations between the larger centers andthe agricultural districts farther north. Insofar as thepattern was already one of the firm, continued subordina-tion of the latter, whether in the fashion of later city-states and their dependencies or merely in a loose buteffective tributary arrangement, there is no reason tothink of an evenly balanced exchange of even symbolicgoods for the southward flow of subsistence products.

This reconstruction is somewhat overgeneralized inthat it fails to take into account the differentiation of set-tlementsize within the zone of clustered, smaller settle-ments. A glance at figure 12 will indicate that numeroustowns, some of them of considerable size, occurred therein addition to still more numerous villages. Note, however,that virtually all those towns are closer to one anotherthan they are to any of the centers farther south. Whilethere is considerable doubt about the size of some of thelatter during the Early-Middle Uruk period, for reasonsalready indicated, certainly some of them (1172, 1237,1306), and in all likelihood Nippur and quite possiblyAbu Salabikh as well, were substantially larger than theirnorthern contemporaries (678, 1166, and 1194, as well asnumerous others in the range of 4 to 10 hectares). Hod-der and Orton have described this kind of distributionas a modification of the classic central place model thatoften provides a somewhat closer fit to the available em-pirical evidence. Citing a variety of archaeological as wellas contemporary examples, they suggest that it is the ag-gregation of services (for our purposes construing servicesvery broadly) in centers of higher order that extends therange of their influence (1976, pp. 63-64). In other words,while specialists who also performed some of those serv-ices were to be found in the districts of generally small,clustered settlement, larger and more diverse groupings ofspecialists in major centers like Nippur and site 1306 re-inforced their individual scales of operations and wereable to attain a disproportionately wider radius ofinfluence.

The density of settlement in the northern region, for-tunately, is great enough to permit at least the approxi-mate lines followed by the major Early-Middle Uruk wa-tercourses to be recognized with little ambiguity. Assistingin their identification, of course, are not only the survivingtraces that have been exposed by wind erosion but alsothe known positions of the major, long-lived channelsthat are even better attested for following periods. It isthese watercourses that constituted the main axial linesto which the settlement pattern as a whole was oriented.)No conceptualization of the pattern can ignore the majorchannels and their accompanying levees as the over-whelmingly most favorable loci for the siting of towns,for routes of intercommunication, and for the most pro-ductive and intensive irrigation agriculture. The region isto be thought of, in short, not as the featureless peneplainof uniform hexagonal service areas envisoned by central

place theorists (Haggett 1966, pp. 119-25) but as a web ofenclosed cells and elongated strips alternately more or lessfavorable for intensive land use, sedentary life, and therapid, efficient interchange of persons, goods, services, andinformation.

This view finds schematic expression in figure 14. Hereunit distances between neighboring settlements alongstraight but evidently branching and rejoining water-courses are arbitrarily assumed, permitting a simplifiedand hence clearer representation of the distribution ofsettlements of different sizes along the network of water-courses. Particularly evident is the contrast betweenwidely spaced towns in the lower part of the diagram, in-terspersed with very few other settlements, and the al-most continuous strings of generally smaller settlementsalong other branches farther north.

Several other important aspects of the pattern are high-lighted by this diagram. In the first place, the significantclusterings of settlement along opposite banks of thelarger branches exhibit a partially alternating distributionor at least do not regularly occur together. The group ofsites including 1159, 1165, 1166, 1194, 1205, and 1306,for example, has no counterpart to the north of the riveralong whose right bank (and right-bank branches) it issituated. Similarly, the right bank grouping below site765 is succeeded by a predominantly left-bank groupingin the vicinity of sites 976 and 1020. A right-bank/left-bank alternation is particularly clear for sites larger than4 hectares, suggesting that towns above this minimum sizetended to be fairly evenly dispersed rather than clustered.The same tendency toward a dispersed distribution of thelarger components of the settlement pattern is apparentif we consider the succession of settlements along eitherbank of any of the watercourses. Sites of 4 hectares orless occur in uninterrupted groups of up to ten, with theinteresting further property that those of 1 hectare or lessare most often found either completely isolated or alongthe margins of a larger enclave. Sites of more than 4 hec-tares, on the other hand, rarely or never adjoin one an-other.4 Finally, sites of more than 10 hectares curiouslyresdmble the very smallest ones in some of their relation-ships. For example, they occur most often either quiteisolated or else adjoined by very small sites rather thanother large ones. Surely, however, different explanationsaccount for the resemblance in overall distribution. Thesmallest hamlets were probably attenuated forms of set-tlement normally associated with marginal conditions andhence with the peripheries of settlement. Large towns, onthe other hand, must have been located in the center ofcultivated regions. Their isolation reflects their success inexcluding competing settlements from terrain utilized (orpotentially utilizable) by the towns' own inhabitants.

Turning once again to figure 12, the asymmetry of thedistribution in the northern region deserves some atten-tion. Insofar as a dependent relationship is posited be-

67

oi.uchicago.edu

Urban Origins

1103f r-I

* 0

//

/

9750 A A E] * *

D101046765

N.

*

*

Abu Salabikh

* 0.1-1.0 ha A 1.1-4.0ha E 4.1-10.0 ha ( 10+ ha

Fig. 14. Schematic representation of arteries of supply and settlement distributionsin the Nippur-Adab region during the Early-Middle Uruk period.

tween the dense zone of agricultural settlement and theline of large, evenly dispersed centers slightly farthersouth, one might imagine that a similar line could haveexisted to the north of the dense zone. Yet this does notoccur. Perhaps a long-term division in control over a sub-ordinate, primarily agricultural region would have beentoo destructive. Somewhat more surprising is the abruptbreakoff in small settlement as well. In fact, a large dis-trict that seems to have been entirely devoid of permanentsettlement of any kind extended all the way northward toat least the present position of the Tigris.

One might assume that much has been irretrievablyburied beneath a heavy blanket of sediments laid downby the Tigris, save that this would hardly explain the ab-sence of even the ubiquitous "stray" clay sickles of Urukmanufacture on the many great mounds in the region thatare of much later date. Perhaps the most reasonable ex-planation of this disparity is the size and destructivenessof the Tigris itself, as described in chapter 1. The risk wasin general too great for substantial towns to be sited withinreach of major Tigris floods. The interface with the pas-turelands to the south, on the other hand, provided aseries of relatively favorable and protected locations. Inany case, we cannot even say at present what course wasfollowed by the major part of the Tigris flow. Some, it has

been argued, must have been included in the meanderingchannel serving the dense band of small agricultural set-tlements in the northern part of the Nippur region. Otherbranches may have occurred at any point within the 40-kilometer band separating the northern limits of settle-ment found in this study and the southern limits of con-temporary settlement on the plains adjoining the lowerDiyala (fig. 10). Because of the dangerous and unpredict-able character of the Tigris, it seems likely that the use ofthis band by pastoralists would have been more sporadicand less intensive than in the otherwise ecologically com-parable zone south of Nippur.

The foregoing reference to "stray" clay sickles promptsa fuller discussion of the interpretive problem they repre-sent. All sites where other contemporary prehistoric re-mains were absent and where only one or two of thesefragments were noted are shown in figures 12 and 13 withcrosses rather than with dots like other sites. Since claysickles were in common use during the Late Ubaid periodas well as in Uruk (and to a rapidly declining extent, alsolater) times, the chronological position of these isolatedoccurrences cannot be exactly assigned. It is also not pos-sible to determine whether they are vestiges of real butsmall settlements that happen to have been heavily maskedby overburdens of later debris, or whether instead they

68

* *

//

//

* S *

N\

a

0 Aý

*

oi.uchicago.edu

Urban Origins

were merely discarded after breakage in outlying fieldsand subsequently became incorporated in the occupa-tional remains of much later periods.

Henry Wright takes the position, in his appendix tothis volume, that clay sickles were most common duringthe Late Ubaid period. The strays he records were inmany instances far from the sharply reduced number ofsites that he can assign to the Uruk period on more se-cure grounds. Hence he suggests that in the Ur-Eridu re-gion they are to be considered as discards of Ubaid datewithin belts of cultivation adjoining natural water-courses or canals. Farther to the north, however, a differ-ent explanation seems more compelling. Except for thesouthernmost portion of the central floodplain (closestto Ur and Eridu), as already noted, sites that could beassigned to the Ubaid period on less equivocal groundsthan the presence of clay sickles was very rare. Uruk siteswere exceedingly common, on the other hand, and thoseof the earlier part of the period sometimes outnumberedall other identifiable ceramic types. And while in someinstances the crosses in figures 12 and 13 occur in mar-ginal locations consistent with defining an outlying beltof cultivation, the vast majority of them are closely in-terspersed among other Uruk sites and, like the latter, re-main in the proximity of ancient watercourses that areoften directly traceable on the air photographs. This sug-gests that "stray" clay sickles-on the central floodplain,although not necessarily along the southern margins ofSumer-probably reflect small Uruk settlements rather ,than Ubaid cultivation. Both explanations are inherentlylikely to have some applicability, if to markedly differingdegrees, in the different parts of what is after all a singlealluvial plain without great intervening distances or othergeographic barriers. But no further resolution of this issueseems possible, at least for the present.5

Patterns and trends of Uruk settlement have heretoforebeen characterized in qualitative and perhaps somewhatimpressionistic terms, although the characterizations aredrawn from maps that systematically reflect the survey'sfindings. A more detailed quantitative assessment offerssome opportunities for further insight into ongoing proc-esses of change. This is possible, of course, only withinthe intensively surveyed area. Table 3 contrasts trendswith respect to the extent of settlement from Early toLate Uruk, separating the northern and southern partsof the surveyed area. A glance at the maps illustrating thesuccession of settlement patterns over this interval makesit clear that these two parts formed cohesive, internallyhomogeneous units that were strikingly dissimilar in theirhistories. Hence an analysis of trends requires that theybe kept distinct rather than conflated.° One observation is in a sense antecedent to this tableand introduces a more general characteristic of the Urukperiod. Some 360 hectares of settlement are recorded inthe northern part as early as the Early-Middle Uruk phase.

TABLE 3 Gross Regional Trends in Uruk Period Settlement

Nippur-Uruk Adab Environs

Environs (North of WS-004)

Early-MiddleUruk periodTotal recordedsettlement 173.1 ha 362.0 ha

Percentage of totalin sites of5 hectares or less 24.7 35.0

Late UrukperiodTotal recordedsettlement 382.5 ha 200.6 ha

Percentage of totalin sites of5 hectares or less 34.8 24.6

Note: The intent of the above tabulation, as well as of figures 14and 15, is to contrast trends from Early to Late Uruk in two regions.In keeping with that objective, sites are not included in which cri-teria of the Early-Middle and/or Late Uruk periods were not spe-cifically identified and differentiated. Thus, sites listed in AppendixA to chapter 3 with the designation "2/3/4" are not included inthe above totals. Sites with the designation "(2 3 4)" also are notincluded. These two categories together comprise approximately39.4 hectares of additional Uruk settlement. There are thirty-twosuch sites, most of them obviously very small.

A reasonable and perhaps conservative estimate of thepopulation involved, based on the standard of 125 personsper hectare of actual site area, or about 100 persons perhectare as calculated only from measurement of maximumlength and width, 6 is 36,000 persons. Yet this area wasalmost devoid of permanent setlement in the Late Ubaidperiod. Nippur and Tell Abu Salabikh can be presumedto have had no more than at most a very few thousandinhabitants. Ubaid traces at other sites are essentiallynegligible. In other words, this was an extraordinarilyrapid, massive process of growth at the very outset of theUruk period.

Population increases under modern conditions candwarf those of any earlier period. Annual increases of 3percent (or even slightly higher) have occurred for somedecades in a number of countries, spurred by the rapid,worldwide movement of foodstocks and by the generalimplementation of public health programs. The doublingtime in these circumstances is only twenty-three years.A tenfold increase would occur, if all other conditions re-mained the same, in seventy-sevsen years. Such rates, how-ever, are absolutely unprecedented. Occasionally in earliertimes there may have been a doubling of population overas short an interval as one generation, but surges of thatkind were surely of very limited duration (Dumond 1975,p. 714). Yet something closer to a tenfold increase than to

69

oi.uchicago.edu

Urban Origins

a doubling seems to have occurred on the central Eu- in the newly emergent pattern there by townsmen andphrates floodplain, and within a period not exceeding settlers already long identified with the region.one or two centuries. Favorable natural conditions, to be In the Late Uruk period the positions neatly reversesure, must have encouraged vigorous natual growth. But themselves. There was only an 8 percent increase in thein addition it is virtually certain that we are witnessing aggregate of recorded area for the two regions, surely aneither an extensive pattern of immigration into the re- amount so low that it requires no further resort to angion, this rapid conversion of large numbers of formerly assumption of substantial immigration or sedentarization.semisedentary folk into settled agriculturalists, or, more But the contribution of the Uruk region more than dou-likely, both together. bles, while that of the north falls by almost half. More-

Greatly complicating the question of immigration is the over, the higher proportion of smaller settlements alsoscale of the phenomenon. Comparing figure 12 with figure shifts to the south, while that in the north declines to the9 not only discloses the increased proportion of recovery same level that had been obtained earlier in the south.in the uncultivated regions surveyed more recently but These figures leave a strong impression that we are deal-strongly implies an extension of the dense Early-Middle -ing with a further population movement, although on aUruk clusterings of settlement far to the northwest of more restricted geographic scale. If we take into accountwhere they are currently known. If so, the estimate of the artificial limitations of the surveyed area from whichsome thirty thousand apparent immigrants within a very\ the only quantitative data are available, literally tens ofshort period can be no more than a fraction of the popula- thousands of small villagers appear to have abandonedtion entering the Mesopotamian alluvium within the space their homes and moved southward.of only a few generations. Moreover, the same trend has Further detail on regional differences and temporalbeen described farther afield. Johnson (1973, p. 90) re- changes in the hierarchy of settlements is provided by theports more than a trebling of the area of settlement (ad- histograms in figure 15. The tendency for a proportion-mittedly, after an earlier decline) and assumes as I do that ately greater number of small sites to occur in the norththis very likely represents an approximate trebling of the in the early part of the Uruk period and later to shift tosedentary population also, on the Khuzestan plains around the south is once again apparent. But in the more refinedancient Susa during the Early Uruk period. Whence, and breakdown given here that shift may be seen to subsumein response to what pressures or incentives, could a stream other changes as well. Sites thought to be exceedinglyof new settlers of this magnitude have come? Lacking an small, 0.1 hectare or even less, are equal in numbers to allobvious source, there is some predisposition to emphasize the remainder in the north during the Early (-Middle)a local process of sedentarization of folk whose presence Uruk period. Sites with only a "trace" of an Uruk occupa-was not previously ascertainable by archaeological means, tion (one or two sherds), not included in this tabulation,But again the apparent rapidity and scale of the process might add half as many sites again of the same small size.create problems for which no convincing answers are im- We do not know, in the absence of excavation, whethermediately apparent. Quite possibly the answer will be they generally consist of isolated rural farmsteads housingfound less in an endless further refinement of archaeologi- at most a family or two, larger villages that left few re-cal techniques (although the importance, and indeed in- mains because they were occupied only briefly, or perhapsevitability, of that can hardly be denied) than in a revision stations devoted to some specialized activity in the coun-of our constructs. The range and rate of movement of tryside that were not regular habitations at all. But in anyprehistoric peoples, under conditions of low population case this large category declined abruptly by the Latedensity and hence limited competition for the use of land, Uruk period. More than nine-tenths of it disappeared inmay often have been much greater than seems "natural" the north, and though there was some increase in the southon a priori grounds. f it failed to match either the increase in the number of sites

Turning back once more from speculation to the (rela- (or the increase in the total population.tively) greater certainty of archaeological survey data, As I have already noted, the left or lower ends of thetable 3 confirms the visual impression from the sequence histograms reinforce the earlier impression of a shift fromof Early (-Middle) and Late Uruk maps that the initial north to south. The environs of Nippur and Adab in thefocus of settlement was preponderantly in the north. The Early (-Middle) Uruik phase and of Uruk in the Latetwo parts are roughly equivalent in size, yet two-thirds of phase are alike in showing a markedly increasing numberthe recorded total site area in the earlier part of the period of sites as one moves to smaller and smaller size categor-is found in the environs of ancient Nippur and Adab rather ies. This tendency is also present, but to a significantlythan in and aound the great southern center of Uruk. It lesser degree, in the earlier part of the period around Urukis also noteworthy that a significantly higher proportion and in the later part around Nippur and Adab. Insofar asof settlement in the north took the form of relatively small there is a "natural" tendency for the number of settlementstowns and villages, occupying an area of less than 5 hec- to be inversely proportional to size, forces appear to havetares. That may well be related to the smaller part played been at work in the latter cases that had a disproportion-

70

oi.uchicago.edu

Urban Origins

20-

12

10

b-

ci

Ez

(n

CD

Uruk EnvironsEARLY MIDDLE URUK PERIOD

1

20-

z LATE URUK PERIOD10-

3 2 2

0.5 1.0 2 3 4 5 6 7 8 9 10 12 14 16 18 20 25 30 35 40 50 60 70 80 80+

Upper Limit of Size Categories in Hectares

Fig. 15. Distribution of Uruk site areas by regions and subperiod.

ately adverse effect on smaller settlements or perhapstended to support population concentrations rather thandispersals.

It is difficult to make a case for any particular set ofdistinct size categories at the lower ends of the histograms.Either aggregating them or examining them individually,there is a more or less regular decline in numbers from thesmallest size categories to ones that are considerablylarger. To be sure, several of the distributions have a long"tail" extending to the right, to upward of 5 or 6 hectares.Those tails suggest that a discrete grouping probably stillawaits unambiguous identification in the realm between3 and 7 hectares, under the rubric of large villages, smalltowns, or whatever other term may be appropriate. Butlet me stress that such a category cannot be clearly dif-ferentiated with the data now available. It is analyticallyconvenient to group sites by size, creating a separation

in the neighborhood of 4 or 5 hectares, but it cannot beargued that the separation into two categories mirrorsa gap in the actual distribution of site sizes.

This blurring does not occur at the right, or upper, endof the histograms. There are wide gaps in all of thembetween the solid grouping of smaller sites and a handfulthat are several to many times larger. In some cases thereis a suggestion of a small but distinct grouping in the 10-14-hectare range. The next group, with no intervening ex-amples, occurs between 20 and 30 hectares. Then thereis another small group, again without intervening exam-ples, between 40 and 50 hectares. Finally there is Urukitself, surely the largest single center throughout the pe-riod. It is suggested here that Uruk grew from an area ofabout 70 hectares in Early Uruk times to about 100 hec-tares by the end of the period.

The size distribution of the clusterings, or the gaps be-

71

I - -

am* - ha A mm

oi.uchicago.edu

Urban Origins

tween them, may be due in part to the random variationexpectable with very small numbers. We will see pres-ently, however, that approximately the same clusteringsand gaps continue to occur through a long succession oflater historical periods. This strongly suggests that by theUruk period at least the larger centers were already fallinginto place at certain steps or intervals that were dis-junctively separated by distance, specialized function, andrange of religious influence or administrative-politicalcontrol. Granting Uruk's extraordinary importance, anexplanation of urban growth in this period that singlesit out as the largest center and then deals exclusively withits unique concentration of ritual activities simply willnot suffice. Uruk was surely at one of the apexes of a(probably fairly diffuse) politicoreligious system, but thehierarchical ordering characteristic of the system musthave been a common feature of many lesser centers. Inorder to generalize about precisely the systemic aspects ofurban and civilizational growth, I shall shortly employ anumber of additional approaches to examine further thehierarchical and distributional aspects of settlement.

Figure 16 applies a different and in general somewhathigher level of aggregation to the same data base used forthe histograms, the estimated sizes of Uruk sites that aregiven in Appendix A to this chapter. The purpose in thiscase is not to display the full array of different site sizes,highlighting gaps and clusterings in the series in order toinfer from them the ascending tiers in the settlement hier-archy. Instead, the more aggregated classification makesclearer the nature of regional and temporal similaritiesand contrasts. Here, for example, the exceptional EarlyUruk emphasis given in the northern part of the area toa range of smaller units of settlement is considerably moreevident, as well as the subsequent, all but complete disap-pearance of the smallest units. Here also can be seen theapparently complete dominance that Uruk already exer-cised over its own hinterlands in Early Uruk times. Later,to be sure, we can see the growth of a few centers of in-termediate size around it, which initially were whollyabsent. It is tempting to suppose that they were not solelya consequence of population growth but were in some wayrelated to unstable conditions and power displacementsresulting from the arrival of large groups of northern im-migrants. But in any case it is not so much the absolutenumbers of sites in each category that is of interest (shownin solid columns) as it is the shape of the curves connect-ing them. Save for the aforementioned deflection to thelow end of the continuum in the case of the northern partof the region in Early and Middle Uruk times, the generalsimilarity of the curves largely overrides the differences be-tween them.

The outline columns adjoining the solid columns indi-cating absolute numbers of settlements convey informa-tion of a different kind. The smallest sites of 0.1 hectarearea or less, whether they were isolated rural farmsteads,

temporary encampments, or whatever, bulk large in thehistograms of numbers and have so far played a consider-able part in the discussion. But on any reasonable esti-mate of their population they were essentially insignificantin comparison with the much smaller numbers of townsand large villages. In a word, it was primarily the rela-tive proportions of the total population living in largerand smaller communities that influenced the nature of thesociety, not the relative numbers of the settlements them-selves. Recognizing that there was surely considerable var-iation caused by other factors, at present we can onlyproceed on the assumption that site area and populationcovary in direct proportion (Adams and Nissen 1972, pp.28-30; cf. Wenke 1975-76, pp. 90-92). The outlinecolumns reflecting proportions of the total settled areascomplement the histograms of site sizes, therefore, withan approximate picture of the proportions of the popula-tion living in settlements of each category.

The first general observation to be made about thesegraphs of population distribution is that they contrastsharply with the accompanying graphs of site distribu-tions. Site distributions classified by size all peak uni-modally at a relatively small size, with a long "tail" ex-tending to the right into the higher categories. Populations,on the other hand, were in each case bimodally distributed.Perhaps we may speak at least figuratively of a broad con-trast between villagers and small townsmen, on the onehand, and urbanites on the other. At any rate, these polarterms suggest the two significant components into whichthe population was divided throughout the Uruk period.

Second, there is an enduring contrast between thenorthern and southern parts of the area with regard to themakeup of the "urban" component. In the south it con-sisted of Uruk alone or virtually alone; the nadir betweenthe two modal maxima lies at 25 hectares or above. Pre-sumably the presence of Uruk in various ways discouragedthe growth of potential competitors above this limit. Inthe north there is a larger group of urban competitors, anumber of them more or less equivalent in size, althoughthere is no way of being certain that size is a reliable in-dex to ranking in terms of power or prestige. The nadirbetween the two modal maxima is deflected leftward inthis case, to about 15 hectares in the Early and Middlephases and to less than 10 hectares in the Late phase. Atleast to judge from the hierarchy of surrounding sitesizes, therefore, Uruk dominated if not directly controlledits immediate hinterlands to a far greater degree than didmore northerly centers.

This contrast between the north and the south withrespect to the dominance of a single center also may beexpressed in terms of a scale of urban "primacy." As wasinitially pointed out many years ago by Zipf (1949), ad-vanced industrial nations tend to be characterized by a"rank-size rule" in which, if the cities of a region orcountry are arranged in order of size, the largest will be

72

oi.uchicago.edu

Urban Origins

Uruk Environs

40-

30-

20-

10-

Nippur-Adab Environs

U.3 1.U 5 1U ZU qU

Upper Limit of Size Categories in HectaresWut 0.5 1.0 5 10 20 40 40+

Upper Limit of Size Categories in Hectares

Fig. 16. Classification of settlement size by region in the Uruk period.

about twice as large as the next largest, ten times as largeas the tenth largest, and so on. Plotting population againstrank on double-log graph paper, systems of cities thatfollow this rule will describe a straight line with a down-ward slope of forty-five degrees. The existence of a har-monic progression of this kind, termed a log-normal dis-tribution, generally is thought to reflect a condition of

regional balance dominated by neither the center nor theperipheries. It can be viewed, as Berry (1964, p. 119) andothers have argued, as the outcome of a stochastic process-the condition toward which many small economic andsocial forces more or less rapidly move a system of cities.

Urban "primacy" is a feature said to obtain when therank-size graph is concave instead of straight, or in par-

73

I

E2 number of sitesI 1 percentage of total area

iiU.JIbn$-_H

L-

oi.uchicago.edu

Urban Origins

ticular when the largest city rises well above the generalslope to which the lesser cities and towns conform. Itspresence reflects the operation of stronger and perhapsfewer forces and suggests that encouragements to popu-lation growth or relocation are not evenly distributedthrough the system but concentrated at its center. Con-trary to earlier expectations, Berry (1961) has shown thatthere is no simple relation between the type of city sizedistribution and the relative degree of economic develop-ment. As Skinner has observed, it "indicates an excessof centrality and suggests either an extraordinary central-ization of regional services or a role for the primate citythat extends beyond its regional hinterland" (1977, p.238).

Thus it appears that primacy is to be thought of as aparticular type of urban size distribution rather than as anecessary stage in the evolution of all urban settlementsystems. It is a type, however, of particular relevance forthis study. Its corollaries, succinctly noted by Crumley(1976, p. 64), at least initially appear to be that "fewer

* I ii - •

cn

WL-

CC

cc

4.C.

forces will affect the urban structure of a country thesmaller the country, the shorter the country's history ofurbanization, and the simpler the economic and politicallife of the country. All of these instances might logicallybe applied in the case of early states."

Figure 17 plots the rank-size distributions of Uruk set-tlements by region as well as by phase. Uruk is clearlyestablished as a primate city in the south throughout theUruk period, although more pronouncedly so in the earlierpart. While there is evidence of substantial growth in set-tlements of all sizes, in other words, the process was notuniform. Instead, there was a relatively more rapid in-crease in the size of smaller towns and villages than inthe size of Uruk itself. In the north, by contrast, the Early-Middle Uruk period saw conditions very nearly the re-verse of those around Uruk. Skinner's characterization ofthe kind of convex or "flattop" distribution occurringthere is that it is "indicative of very imperfect integrationat the level in question" (1976, p. 241). Subsequently therewas a decline not only in the number of settlements but

Rank Rank

Fig. 17. Settlement rank size by region in the Uruk period.

74

oi.uchicago.edu

Urban Origins

in the size of most of them, so that by Late Uruk timessomething approaching a log-normal distribution seemsto have obtained. Starting from markedly different rank-size distributions, in summary, the settlement hieararchiesin the two regions gravitated in time toward each other-and toward a log-normal pattern.

Suggestive as this finding is, its imprecision must alsobe recognized. A degree of permanence probably can beassumed for most of the larger sites in the hierarchy, butthere are corresponding difficulties in estimating the sizeof many of them because of massive overburdens of laterdebris. For the smaller sites, on the other hand, relativelygreater accuracy in assessing the size of most of them isaccompanied by greater uncertainty whether they weresimultaneously occupied and hence can be arranged in asingle hierarchy. A case could be made, if the data wereless defective in these respects, for examining the appar-ent changes in intercept and slope value in the two con-trastive distributions more rigorously to ascertain theirstatistical significance as a function of time (Malecki1975). But while this is unjustifiable in the circumstances,the apparent changes have a bearing on Crumley's formu-lation of urban primacy as a characteristic of early statesettlement systems. Inherent in that formulation, as sheobserves, is "the notion that the primate to rank-sizegrowth model approximates the process of 'urbaniza-tion' " (1976, p. 66). Uruk and its hinterlands obviouslyreinforce the suggested sequence of development. The"flattop" distribution farther north is apparently just asearly as Uruk, however, and is neither primate nor log-normal but something of an antithesis to both. Clearly,this difference is somewhat at variance with the proposalof a unilineal sequence of development, perhaps implyingan early and persistent bifurcation of urban function. On ,the one hand, as Crumley envisions, are cities like Urukthat supply "functions"-the more abstract term ex-plicitly extends outside the economic realm of goodsand services to include organizational, religious, defen-sive, and other domains-to an agricultural hinterland inneed of them. On the other hand, there may also be urbancenters almost as large whose size is a function of theminimum threshold needed for a defensive agglomeration,or of a nonsedentary clientele in the hinterlands.

A further observation concerning the size of the set-tlement component that must be considered in some re-spects "urban" may be drawn from the last two figures.Zipf suggested that the lower threshold of urban sizemight well be set at the point at which the rank-size ruleno longer obtained because the rank-size graph bentsharply downward (1949, p. 424). No consistent point oreven narrow zone of sharp downward curvature is ap-parent in figure 17. A range between 3 and 5 hectares ap-pears most plausible for the four cases, but almost a sixthof the entire number of known sites fall within this range,so that it is of little assistance. Moreover, this threshold

for distinguishing between urban and rural is drawn fromobservation of modern industrial systems, to which dif-ferent regularities may well apply. Implying a definitionof urbanism at population levels of only three to five hun-dred persons by the rough standard of equivalence hereemployed, it seems (on admittedly apriori terms) too in-clusive to be useful in analyzing urban development thatsoon reached a level one hundred times larger. Settingany threshold at a higher population size can perhaps onlybe done arbitrarily at present, especially since it cannotyet be tested or buttressed with data from excavationsillustrating differences in cultural inventory that are re-lated to size of settlement. But it is surely not unreason-able to place at least a provisional line in the neighborhoodof 10 hectares, about a thousand inhabitants. Bifurcatingsites by area along that line, we obtain the pattern shownin table 4.

TABLE 4 Urban and Nonurban Settlement by Region inSuccessive Uruk Subperiods

10 Hectaresor Less More than

("Villagers/ 10 HectaresTownsmen") ("Urbanites")

Early-MiddleUrukUruk area 53% 47%Nippur-Adab area 54 46

LateUrukUruk area 61 39Nippur-Adab area 30 70

This tabulation implies that Uruk's dominance of the ^southern region was somehow linked to the greater dis-persal of the population there into small settlements.Uruk's own growth notwithstanding, there was a trendaway from urban concentration in its environs. In thenorth, by contrast, urban concentration increased eventhough there was simultaneously a general populationdecline. What may be a prototype of the later pattern ofcontending city-states, none of them able to find an as-sured, permanent basis for dominance but all of them to-gether drawing the bulk of the sedentary populationwithin their walls, is to be seen emerging as early as theLate Uruk period.

More generally, however, the above tabulation indicatesthe importance of an agglomerated-urban, by the defi-nition here employed-mode of settlement throughoutthe Uruk period. Already in its Early (-Middle) phase,this can apparently be said of the place of residence ofalmost half the population, and in aggregate that levelwas maintained or even increased slightly in the Late

75

oi.uchicago.edu

Urban Origins

phase. Thus it can reasonably be said that the mostsalient characteristic of the Uruk period that we can yetidentify, certainly in this data but arguably in the rest ofwhat is known, is that the population and hence the so-ciety of the time was well on its way to becoming urban.

Assigning central importance to the thrust toward ur-banism is admittedly primarily a descriptive rather thanan explanatory step. Ruins much more extensive than anypreviously known bulk large in the data on the Uruk pe-riod produced by the survey, and there is a natural but per-haps overcautious predisposition to retain a descriptiveframework that accurately reflects what is found. Thepresence of urban centers, no matter how numerous andrapidly introduced they may have been, does not by it-self tell us much about the structural features and motiveforces behind the rapid development. And ultimately theaim of science must be not merely to describe phenomenabut to raise and answer questions of precisely that kind.

But there is an important distinction between immedi-ate and ultimate objectives. Premature acceptance of anexplanatory framework can focus attention on too lim-ited a set of variables or can lead to recording only a se-lection of the relevant features of the behavior of thosevariables. Particularly in a study like this one, largelydevoted to the primary recording of empirical findings, amore cautiously descriptive framework seems appropriate.

I maintained earlier, for example, that the histogramsof settlement size (fig. 15) do not permit us to distinguishunambiguously the tiers of a hierarchy below somethingon the order of 10 hectares. That tends to direct atten-tion toward urban phenomena that were concentratedexclusively in the larger communities and to imply a polarseparation between them and the smaller ones. I concedethat this is unfortunate. But at this stage it does not seemeither to be more responsive to the data or to providegreater analytical penetration simply to impose gradedhierarchies in spite of existing variance and then to reifythem into bounded categories of behavior.

In a sense, to be sure, science very often proceeds bytentatively establishing categories and then testing theirsignificance in domans independent of the one to whichthey were first applied. That procedure is implicit inGregory Johnson's attempts to demonstrate a correspon-dence between his proposed levels for an Uruk settlementhierarchy in southwestern Iran, the cultural inventory ofsites at different steps in the hierarchy as revealed by ex-cavations, and other regularities of spatial patterning. Butthere are unmistakable overtones of reification and circu-larity in such an effort, particularly when, as in this case,the support rendered by supposedly confirmatory phe-nomena is still sharply limited by their modest scale andfrequency. Working with a much smaller and thereforeless continuously distributed series of site sizes, he hasdistinguished large and small "centers" as well as largeand small "villages," all within the range of less than 10

hectares of total size (Johnson 1973, p. 79). This then istreated as the primary evidence for a three- or four-levelsettlement hierarchy, and the latter in turn is used as sup-porting evidence for a corresponding number of levels ofdecision-making and hence for the existence of the stateas a form of territorially extended political organization(Johnson 1973, p. 141). Attractive as the idea of a "test-able" definition of the state along these lines may be, it ishard to avoid the impression that the evidence so farassembled on its behalf is quite unconvincing (cf. alsoCrumley 1976).

Johnson's position is of broader interest than merelyfor its divergent interpretation of Uruk settlement hier-archies. He defines a state as a "differentiated and inter-nally specialized decision making organization which isstructured in minimally three hierarchical levels" (1973, p.2), and the state's emergence as early as in Early Uruktimes is then equated with the introduction of tripartiteinformation-processing and decision-making. One mightinfer that the primary task of the archaeologist investigat-ing the Uruk period is to test this proposition by seekingout material relics that can be inferentially associated withhierarchically structured tripartite (or, even better, quad-ripartite) distributions of behavior. The objective is indeeda good one, but it is to be regretted that the ancient Urukinhabitants were seldom so obligingly differentiated in therefuse they left behind for us to recover.

While this is in no sense a theoretical treatise, referenceto Johnson's work requires a brief excursus upon thelarger problem-setting to which it is a central contribu-tion. A recent review of studies of the origin of the stateprovides a useful starting point, for its author not onlyshares Johnson's general position but relies in part onthe same body of Uruk period data from Khuzestan.

Uruk society is an example of early state society, ac-cording to Henry Wright also, as compared with the taxo-nomically simpler and less centralized "chiefdoms" out ofwhich it is usually proposed that states emerged more orless independently in a number of different prehistoric andhistoric sequences (Wright 1977). Wright eschews anyconcern for urbanism as it relates to the state. FollowingElman Service, he believes that "urbanism, considered aspopulation nucleation, evidently follows state origin andis correlated with the pattern and intensity of warfare be-tween existing states." Similarly deprecated as fundamen-tal properties of the state that are related to its emergenceare class stratification, a formative phase of largely theo-cratic leadership, militarism, intensification of production,and a number of other factors that need not concern ushere. Instead, and apparently exclusively, at least for defi-nitional purposes, "a state can be recognized as a culturaldevelopment with a centralized decision-making processwhich is both externally specialized vis a vis the localprocesses which it regulates, and internally specialized inthat the central process is divisible into separate activi-

76

oi.uchicago.edu

Urban Origins

ties which can be performed in different places at differ-ent times."

The phenomenon of the state, Wright maintains, mustbe defined in terms of processes through time. That leadshim to "look at processes which control other processes:at the central decision-making or regulatory activity ofthe system of activities, rather than at groups, institutionsor roles." What makes the Uruk period pertinent to thisemphasis is that there is a variety of glyptic and ceramicevidence that can be interpreted as indicating that "anadministration controlled the movement of goods fromproduction points to assembly points and thence to centralpoints for aggregation and subsequent redistribution." Infact, Wright believes, Uruk administrative networks canbe defined. "Each had a major center in an agriculturally ^rich area, within which was a network of smaller admin-istrative centers and production centers. Production inthese networks was differently organized. For example,while some settlements seem to be primarily concernedwith agriculture and moved their products through cen-tral pools into redistribution networks, parts of centralsettlements are concerned with ceramics production andmove their products to agricultural settlements by non-redistributional means" (Wright 1977, pp. 380, 381, 383,386, 387).

This is not the place to develop a comprehensive al-ternative approach to the study of the origin of the state,but an understanding of the Uruk period requires com-ment on certain aspects of Wright's position in theoreticalas well as empirical terms. Its most attractive feature isthat it remains at all times keenly responsive to the po-tentialities of the archaeological record (cf. Wright andJohnson 1975). Wright properly sees seal impressions,ceramic workshops, and site hierarchies not as discreteand unrelated traits but as indicators of social relation-ships that must be systemically related to one another.He further demonstrates that archaeological procedurespermit us to take quantitative account of those traits, andso to approach a quantitative understanding of the be-havioral patterns, social relationships, or both in whichthe traits were originally embodied. Stress on an informa-tion-processing metaphor for the state, viewed in theseoperational terms, should prove a valuable analyticaltool for a considerable time to come.

But let us examine some of the difficulties. AlthoughWright indicates that his interest is in political evolution,he translates this into a central focus on, or even con-fines himself to, certain routinized elements of bureau-cratic administration. In an earlier article, the subor-dination of politics to administration is made moreexplicit: "By 'administrative' we mean 'control,' thusincluding what is commonly termed 'politics' under ad-ministration" (Wright and Johnson 1975, p. 267).Nothing we know of the historic records of any society(wherever there are such records), however would allow

even the full battery of administrative routines (assumingoptimistically that they could ever be known archaeologi-cally!) to stand as surrogate for its political system as awhole. The routines not only constitute a gross over-simplification of politics but also provide a misleadingpicture-one lacking in the pervasive but volatile andusually unexpressed elements of contingency, calculation,and coercion. Similarly missing from the analysis, in theface of overwhelming evidence not only of its importanceas a historic force elsewhere but of incontrovertible ar-chaeological evidence that it was the predominant pre-occupation precisely in the Uruk period (Nissen 1972,pp. 793-94), is any concession of a special role for re-ligion and religious institutions. In spite of its operationalattractions, therefore, Wright's definition of the centralfeatures of Uruk society as a state society is theoreticallytoo narrow to serve as a guiding conceptual frameworkfor us here.

There are, in addition, a number of substantive diffi-culties with evidence adduced in support of this generalposition. Critical to its support is the demonstration inthe archaeological record of hierarchically organizedflows of goods and services as well as (rather more in-tangible) "information." It was originally concluded thatpottery production was an activity centralized in thelargest sites, at least in the Uruk centers of southwesternIran (Wright and Johnson 1975, pp. 279-80). Soon after-ward, new data provided "relatively conclusive evidenceof ceramic production on small settlements," althoughit still appears to have been much more heavily concen-trated on large ones (Johnson 1976, p. 209). It was alsoargued that beveled-rim bowls were not merely con-venient, ubiquitous Late Uruk index fossils with vagueritual connotations, but indications of a centralized, state-administered grain-rationing system like that which canbe unequivocally identified from texts in late Early Dy-nastic times (Wright and Johnson 1975, p. 282). The ob-servations used to support this argument, however, sup-port various other alternatives with equal plausibility(Adams 1975c, pp. 459-60), and now the case for a pri-marily "ritual" identification of the bowls is being vigor-ously argued once more (Beale 1978). Clay cones usedin wall mosaics are another item that has been taken asevidence of specialized administrative functioning (John-son 1975, p. 319), although the distributional evidenceappears to be consistent with a variety of ritual or publicpurposes as well as with their use as status markers in-dependent of an administrative system (Adams 1975c, p.458), and although the excavated examples overwhelm-ingly come from architectural complexes whose centralbuildings have formal features by which they are usuallyidentified as "temples."

Finally, it is worth noting that excavations at smallUruk sites in Iran are beginning to produce "unexpected"data on social status differences. Substantial buildings as

77

oi.uchicago.edu

Urban Origins

well as relatively rare and valuable imports like copperand lapis lazuli are mentioned. Contrary to his own an-ticipations, the investigator has been led to suggest forat least one small Uruk agricultural village that "the rel-atively high status of some of its residents was inde-pendent of administrative function" (Johnson 1976, pp.216-17).

In short, a number of useful avenues of investigationhave been opened-but so far with inconclusive resultsat best. The case for a primarily administrative axis ofdevelopment at the heart of the burgeoning growth andcultural achievements of the Uruk period remains ex-ceedingly weak. I would add my personal view that itremains highly implausible, while also conceding that theSisyphean effort to establish such a case may prove tobe not only one of the most operationally rigorous ofcurrent approaches but also one of the approaches mostproductive of "unexpected" insights. At least for themoment, however, it seems significantly more responsiveto the mass of the available data to remain with the pri-mary "urban" characterization of the Uruk period thatwas applied earlier.

A central part of the strategy employed by Wright,Johnson, and others to investigate developments of theUruk period in southwestern Iran has been to balancethe ongoing excavation programs at major sites like Susaand Chogha Mish with small-scale soundings at a repre-sentative variety of the smaller settlements. Brief, prob-lem-oriented enterprises of this kind have not receivedofficial encouragement in Iraq, so that the empiricalfindings referred to above cannot yet be checked andduplicated with material obtained in stratigraphic con-text from within southern Mesopotamia proper. It isof interest, however, to ask the same questions of surfacecollections obtained from Uruk sites within our region.Table 5 assembles the data currently available for doingso, recording certain categories of imported materialsand other possible indicators of social status or spe-cialized activity by presence or absence. The validity ofattributing surface materials to a particular period isobviously a matter of question, but the table takes note ofsites whose surface collections we can rely on more con-fidently because typologically similar material of later dateis not attested.

The table lends support to the suggestions emergingfrom Johnson's recent work that patterns of distributionare complex and fail to reflect a centralized, hierarchi-cally organized system. Ceramic production clearly wasnot confined to the major sites but was very widespread,occurring on sites as small as 0.2 hectares in area. Copperalso is present on extremely small sites. The use of stonebowls, an item of at least modest luxury, since the rawmaterial was entirely unavailable in the alluvium, ap-parently was all but universal. Carefully ground stonemace-heads, arguably a status indicator or an item of

specialized military equipment, in three cases occur onsites of less than a hectare. Wall cones occur repeatedlyon sites of the smallest size category as well as largerones. Obsidian from the Lake Van region of easternAnatolia (cf. site 1072 in the general site catalog, chap. 7)was perhaps a utilitarian rather than a luxury import, butit came from a sufficient distance so that selective usewould be expected. It was indeed found only sporadically,probably because of difficulties in maintaining uniformstandards for observing small, dark fragments during arapid surface reconnaissance. But there is no obviousrelationship between the sites where its presence wasnoted and hierarchies of site size. Even if we considergroupings of these indicators rather than individual cate-gories, there is at best a very loose positive associationwith increasing site size. At least four of the six categoriesare represented on thirteen sites. While seven of thesesites are 6 hectares or more in area, the remainder aresmaller-in one case less than a hectare.

The lack of a close or obvious relationship betweenthe hierarchy of sites and the patterns of distribution asthey are currently understood is reinforced if we con-sider certain categories of mass production. Clay sicklesand beveled-rim bowls are dealt with in detail in Ap-pendix A because of their relevance for chronologicalquestions, but the essential findings can be briefly re-capitulated. Beveled-rim bowl frequency is not positivelycorrelated with site size, as it probably should be if thesevessels were employed in a centrally administered ration-ing system, and their manufacture occurs on sites as smallas 0.5 hectare. Clay sickles, peaking in frequency some-what earlier, were even more widespread in their manu-facture. Nor is there any evidence that fragments weremore numerous on the smaller sites, as might have beenexpected if there were a significant proportion of non-agricultural specialists in the larger centers (note, how-ever, that frequencies are not available for the largestcategory of centers).

Granting the imprecision of a record of this kind, thedata provide little support for a hypothetical dependenceof the territorial system of settlements on a centrallyorganized flow of goods and services. A better case canbe made that the primary basis for organization was ofa rather more traditional kind: religious allegiance todeities or cults identified with particular localities, politi-cal superordination resting ultimately on the possibilityof military coercion, or a fluid mixture of both.

This is not to imply that the only decisive developmentsof the Uruk period lay in the spreading influence of cults,or in the claim of larger centers to political and economicsuzerainty over weaker neighbors. There was, as we haveseen, an extremely rapid growth of the sedentary popu-lation that sustained these new, territorially extendedrelationships. Behind that must lie important innovationsin agricultural practices, even if at present we can do

78

oi.uchicago.edu

TABLE 5 Possible Indicators of Status and Specialization in Uruk Period Surface Collections

MAX. MAX. MAX.URUK v URUK S I URUK 1S I T E

ITE E SITE E. ARA AREA . AREA RE

(HA.) w (HA.) w S (HA.) °5P4 w H HZ wN(ZA00 U

020 5. + 247 0.1 + 837 1.4 / + +

023 1. / + 260 14. + 845 8. / +

028 2.9 + 262 3.5 + + 853 4. / +

042 3. / + 264 1.9 + 939 2.4 / + + + +

Q48 0.5 + 267 0.3 + 940 1.7 / + + + +

051 3.6 + -272 0.5 + + 975 5.8 / + + +

060 1.8 J + 274 1.1 / + + 1020 8.2 / + + +

082 0.5 + + 276 0.5 + 1046 8.6 / +

087 7.8 + + 282 2.6 + 1072 3.4 / + + + +

107 2.6 + 285 0.9 + 1100 0.5 +

109 5.2 / + + 292 1.9 + 1118 0.6 / +

110 9. / + 293 1.7 + + 1124 6.8 / + + + +

118 5.3 4 + 297 1.6 + 1137 3.8 + +

125 24. + + + + + 305 0.1 + 1154 2.6 / +

126 4. + 309 0.6 + 1159 4.5 / +

127 4. + 310 0.9 + 1163 0.9 +

128 1.8 + 314 0.6 / + + 1164 1. J +

133 0.2 1 + 317 0.1 / + + 1165 5.3 / + +

137 1.5 / + 329 0.1 + 1172 25.5 4 + + + +

144 2. + + 331 0.9 / + 1194 11.5 / + + + +

152 6.6 / + 334 0.4 / + 1198 3.4 + +

153 2. + 338 2.6 / + 1199 1. / +

160 0.4 + 339 0.1 + 1205 7.9 + -+ + +

162 6. + + 376 4.8 / + 1216 4.8 + + + +

163 4. + 386 1.5 / + 1217 0.1 +

164 1. + 406 0.8 + + 1261 0.8 / +

166 4. + 407 6. + 1293 5. / +

168 5. + 573 1. / + + + 1306 50. + +

174 3. + 574 1. / + + + 1312 0.2 J + + +

181 5.8 + + 680 0.1 / + 1315 2.8 + + +

185 4.4 / + + 711 0.1 / + 1383 1. +

191 4. + _765 5.3 / + + + 1394 7.7 + +__

193 3. . + 790 5.5 / + + + 1416 0.6 / + +

201 11. _ + 792 6.8 / + + + + 1432 1.7 J + + + +

219 1.4 + + 798 5. -- - + 1..+1448 0.9 + + + + +

244 0.1 + 804 1.7 J + + 1615 0.6 +

245 6. + + + + + 805 0.5 / + A274 15. + +

Note: Sites indicated as "Uruk only" were unoccupied during the succeeding Jemdet Nasr and Early Dynastic I periods, so that an Urukdating can be assigned to the traits in question with considerable confidence. Kiln wasters were not recorded in the Warka survey (sites001-466), nor generally at sites surveyed later where intensive collections were not made. This applies also to obsidian.

79

oi.uchicago.edu

Urban Origins

little more than speculate about what they may havebeen.

At least on the evidence of surface collections, chippedstone "hoes" (if that is what they were) continuedthroughout the period in undiminished frequency. Likethe other stone utensils and even rough limestone boulders(now badly decomposed by weathering) that are almostalways plentiful on Uruk sites, their transport into theplain as raw material or finished products represented asubstantial basis for interregional contact, as well as aconsiderable investment. How did the use of "hoes" ar-ticulate with the use of ox- or donkey-drawn plows,clearly attested in Uruk pictographic writing even if nonehas yet been encountered archaeologically? It is temptingto suppose that some major agricultural innovation suchas the plow, or perhaps an interrelated series of improve-ments in the technology of irrigation, sustained the pre-cocious growth that is the hallmark of the period. Butthat cannot yet be demonstrated. And the ubiquitous claysickles instead seem to point to a persistent, overwhelm-ing concern with certain critical episodes of the sub-sistence quest, such as the harvest. All that can be saidwith certainty, therefore, is that there was a broad dis-tribution of a considerable variety of artifacts displayingconsistently superior craftsmanship. Verging yet oncemore beyond what can be shown with present evidence,this suggests that the basis for the productive achieve-ments characterizing the impressive numbers of newsettlements of the Uruk period lay in the proliferation ofpart-time specialists in communities of all sizes.

But what of the handful of major centers like Uruk?Status display, ritual, and administrative requirementssurely reached their highest levels of complexity in thesesettings-always remembering that the three are likelyto have been deeply "embedded" in one another ratherthan differentiated. In this respect the difference seemsto have been qualitative and not merely quantitative.Whole classes of objects found in the major centers tran-scend craftsmanship and become art. As Henri Frank-fort has said of the seals and seal impressions, they displaya "creative power . . . such that we meet among [the]astonishingly varied products anticipations of everyschool of glyptic art which subsequently flourished inMesopotamia" (1939, p. 23). Writing is a parallel case.Among the earliest pictographic tablets from Uruk arenot merely crude records of economic and administrativetransactions but formalized lists of gods, professions,geographic names, and classes of objects arranged inconceptual categories (Hans Nissen, pers. comm.). Thelist-keeping, classificatory aspect of cognition that per-manently stamped scribal learning for as long as therewas a cuneiform script thus had made its full-blownappearance here very soon after the script's origins, per-haps reflecting long familiarity with simpler counting

and mnemonic devices (Schmandt-Besserat 1977). Par-ticularly with regard to the crafts and professions, more-over, a hierarchical arrangement of terms is apparentfrom the very outset. Three ascending ranks are separatelylisted (Nissen 1974, pp. 12-14), surely implying that theyfunctioned within a well-defined institutional structure.These examples, and the uniformity with which they canbe extended to other fields like stone sculpture and metal-working, clearly suggest something more than part-timecraftsmen, scribal recorders, priests, and rulers or priest-rulers. In centers like Uruk a highly significant segmentof the population must have been given or won its free-dom from more than a token or symbolic involvement inthe primary processes of food production.

While similar in the sudden appearance of technologi-cal as well as stylistic virtuosity, the example of metallurgyintroduces an important further issue. The availableevidence suggests that copper artifacts were much morewidely employed than the other new arts. Well-made,sophisticated copper jar shapes, for example, can occurby no later than the Late Uruk period on a site of quitemoderate size (e.g., site 185; cf. Adams and Nissen 1972,fig. 82), and unrecognizably corroded fragments are rela-tively common. In general they occur in and aroundshallow surface disturbances accompanied by humanbone, seemingly indicative of illicitly excavated burials.By this time, therefore, copper vessels and other artifactsare likely to have been occasionally employed for per-sonal use as well as for grave goods at least by some ofthe higher-status families residing in towns in the hinter-lands as well as in the major centers. But it is less easythan in the case of stonework to suppose that they couldhave been fabricated locally by part-time specialists. Theskills involved in casting, as well as the difficulties andcosts of extracting the ore and bringing supplies of copperfrom the distant mountains of Iran or Anatolia, arguestrongly for a concentration of the craft. Centers likeUruk not only concentrated demand for finished productsbut were best equipped to sustain the subsistence andraw materials requirements of skilled cadres of crafts-men over long periods.

This brings us to a function of towns like Uruk ex-tending well beyond, although certainly complementaryto, their political and religious roles. They must haveplayed the crucial part in organizing the long-distanceprocurement of certain commodities like metals, woodfor heavy construction, precious stones, and perhaps evenordinary flint and construction stone as well. That meantnot only providing sustenance for craftsmen but organ-izing trading expeditions, even if "down the line" tradefrom one town to its neighbor may have limited the rangeto be covered. And trade demanded further production,since there were no significant Mesopotamian raw ma-terials to be exchanged for needed imports.

80

oi.uchicago.edu

Urban Origins

Here, then, we see the Uruk system of urban centersand their hinterlands in a different light, rather closer towhat Gregory Johnson and Henry Wright envisage thanto the political and religious relationships that have beenstressed above. A flow of resources had to be securedfrom the countryside: grain, domestic animals, otheragricultural products, corvee labor, perhaps some finishedor semifinished commodities like spun wool, woven cloth,hides, dried fish, reed mats, and beer. Other resourcestherefore must have traveled in the opposite direction:stone of various kinds, luxury goods to validate the statusof subordinate local elites, copper tools, and weaponsand vessels for utilitarian purposes as well as for con-spicious consumption. Without a reciprocal movementof goods and services of this approximate form, it isdifficult to see how the surface distributions tabulatedabove could have been brought about. Indeed, it is diffi-cult to see how else a hierarchically differentiated settle-ment system like this could have arisen and flourishedso prodigously.

The question on which this interpretation continues todiverge from Wright's concerns the incentives and instru-mentalities by which the flow of goods and services wasprimarily brought about. Wright apparently feels that itcan best be understood as a centrally administered systemand that its two crucial, defining features were redistribu-tion and a hierarchically organized array of bureaucraticfunctions. The different, though not entirely contrary,view offered here is that we are dealing instead with amuch less simple, less stable mix of relationships betweenthe centers and their peripheries. Among its features were:deities whose cults attracted pilgrimages and voluntaryofferings; intervals of emergent, centralized, militarilybased domination of subordinate centers that had beenreduced to the status of clients, alternating with otherintervals of fragile multicenter coalition or local self-reliance; coercive extraction of rural resources alternatingwith more or less freely balanced exchange of subsistenceproducts of the countryside for status symbols and certainlimited but important categories of utilitarian goods thatcould not be produced locally.

Assuming that something along these fluid, regionallydifferentiated lines obtained during the Uruk period, itseems misleading to characterize relationships betweencommunities of different size in terms of narrowly spe-cialized flows and routinized administrative relationships.State systems of varying size probably existed from timeto time during the Uruk period. But the far more durable,pervasive Uruk characteristic was that it brought a largeand growing population within the compass of an urbanway of life. Urbanism, to be sure, denotes no set of precise,well-understood additional characteristics for societiesso described. But that may even be an advantage in thiscase, when we still know so little.

THE FIRST URBAN CLIMAX

We must now turn from characterizations of urbanismduring the Uruk period to the sequence of even moreimpressive developments that followed soon afterward.The importance of the Uruk period, after all, lies not ininstitutional patterns reaching their culmination at thattime, but in its enormous, diverse, comparatively welldocumented surges of development toward culminationsoccurring throughout the later span of Sumero-Akkadianhistory. Specifically with regard to urbanism, the growthtrends that apparently were initiated in Early Uruk timesreached their fullest consummation in the early thirdmillennium. This is the process of growth that centrallydetermines the somewhat unconventional framework ofperiodization that has been adopted here. Rather thantreating all of the Early Dynastic period as a unity,and linking it most closely with developments that fol-lowed in the later third millennium, it seems impor-tant to integrate the first of its three phases with theforegoing discussion of Uruk's rise to full urbanstature.

Intervening between Late Uruk and Early Dynastic I isthe Jemdet Nasr period, an elusive and probably briefinterval to which the findings of a surface reconnaissancecan contribute little. Its duration cannot be directly es-tablished from dynastic lists or radiocarbon series, but itis very unlikely to be more than a century or two. Asis described more fully in Appendix A to this chapter,it is apparently not characterized by any widely oc-curring "index fossils" that do not also pertain to anearlier or later time. Single-period Jemdet Nasr sites canbe distinguished, as can Jemdet Nasr occupations over-lying Uruk occupations on the same site. But it is oftenimpossible to be certain from surface collections alonewhether Jemdet Nasr levels underlie Early Dynastic Ilevels at the same site, unless the latter occupation wasof limited depth or was sharply reduced in extent fromwhat it had been earlier. At many sites, therefore, thepresence of a Jemdet Nasr occupation must be consid-ered much more problematic than the succeeding andfollowing ones. Statements about individual site sizes,not to speak of aggregate areas of settlement, are cor-respondingly more doubtful.

In spite of these conditions, the presence of an ex-tensive Jemdet Nasr occupation in the hinterlands ofUruk could be relatively well attested. This is largely aresult of subsequent large-scale abandonments accom-panying the urban growth of Uruk in the Early DynasticI period. As noted in the published findings of the Warkasurvey, the Jemdet Nasr period seems to have witnessedthe apogee of relatively dispersed, rural settlement in theUruk environs (Adams and Nissen 1972, p. 27). Sincetrends in settlement in that area extending through the

81

oi.uchicago.edu

Urban Origins

Jemdet Nasr period have already been analyzed in de-tail, they need not be recapitulated here.

Figure 18 reproduces the Jemdet Nasr period settle-ment pattern for the Uruk area but adds to it the exten-sive northern area more recently surveyed. Superficiallyit appears that by Late Uruk times the abandonmentunder way in the Nippur area had already continuedfurther, at least if we consider only the definitely attestedsettlements rather than the doubtful ones. It shouldbe recalled, however, that the Late Uruk map may in-clude some sites that were given only a generalized des-ignation as Uruk ("2/3/4" in table 7 of Appendix A tothis chapter) and that may not have been occupied afterthe Middle Uruk period, accentuating the contrast withthe Jemdet Nasr map. For reasons stated earlier, more-over, confining attention to definitely attested settlementsmay well result in a misleadingly large reduction in thetotal number of occupied sites as compared with that forthe Late Uruk period. In addition, most of the sites thatwere abandoned were small, and at least two that werenewly occupied (site 1032, although its Jemdet Nasrdimensions are uncertain) or reoccupied (site 1237) wereof substantial size. Hence the first impression that therewas a continuing decline in the northern region needsfurther scrutiny.

Let us consider only those sites that are definitely at-tested for the particular phase or period in question. Onthis basis there is a decline from thirty-seven, with anaggregate area of about 200.6 hectares, in Late Uruk timesto thirty-one, with an aggregate area of about 177.3 hec-tares, in the Jemdet Nasr period. But if most of the ques-tionable Uruk sites were Early (-Middle) only-as seemslikely-and if no more than a portion of the ten additionalsites whose doubtful Jemdet Nasr occupations are in-dicated by parentheses in table 7 were actually inhabited,there would have been essentially no further decline eitherin number of sites or in occupied area. Thus the overallpicture of population density is rather similar to thataround Uruk. The principal difference is the one alreadyevident earlier, that in the north the trend was towarda number of coexisting urban centers of moderate sizerather than toward a single major one like Uruk that ex-ercised an influence far beyond its immediate supportingarea.

This emergent regional contrast reaches a climax in theEarly Dynastic I period. Uruk, already the largest centerof its time, underwent a further phase of impressivegrowth. According to the traditions recorded in theSumerian Kinglist, the "builder" of the city wasEnmerkar, an early Uruk dynast whose reign must havefallen either toward the end of the period or at the verybeginning of Early Dynastic II. The tradition is atleast plausible. Several tens of thousands of people musthave been persuaded or compelled to abandon theirformer towns and villages and inhabit the new city. Only

with their participation, again on an uncertain basis ofpersuasion or compulsion, was it possible for Enmerkar'ssuccessor, Gilgamesh, to undertake the construction ofthe massive wall defending it, still to be seen enclosingits ruins. The achievement suggests a strong ruler, whoseconnection with it would have remained alive in memory.The era, however, is on the remote edge of currentlyusable synchronisms and not wholly speculative conjec-tures as to absolute chronology, based on the spans ofyears assigned to following rulers. A dating for his reignin the early twenty-seventh century B.C. is at least in-dicative of where the chronicles, just verging out of legendand into history, seem to point.

Even the relative chronology is contentious and obscureat this juncture. Dynastic synchronisms are neither nu-merous nor unambiguous, and they must be balancedwith partly countervailing considerations based on stylis-tic development, stratigraphy in the royal cemetery at Ur,and orthography. Absolute dating is still more unsatis-factory. Radiocarbon determinations are as yet few andbeset with contradictions. Those currently available forthe Jemdet Nasr period are inconsistent, probably be-cause they are based on unreliable surface samples. Andwhile there are five usable, closely grouped Early DynasticI determinations, there is a considerable discrepancy be-tween alternative calibrations of them. Use of the Suesscalibration curve suggests that the end of Early Dynastic.I and the transition to Early Dynastic II occurred about,or just before, 2900 B.C. The MASCA correction for thesame five samples, on the other hand, indicates a datejust before the mid-twenty-seventh century. Whateverthe relative strengths of the statistical arguments for thetwo schemes, the latter at least has the virtue of con-forming much more closely with the outcome of deadreckoning from the Kinglist (Wright 1973, pp. 200-201;Crawford 1977, pp. 8-9, 14-16).

Potentialities for interpretation are in one crucial re-spect decidedly better than they were in the Jemdet Nasrperiod. So-called solid-footed goblets, very largely limitedin use to this period, were in widespread use. They con-stitute an excellent indicator of Early Dynastic I occupa-tions even at sites with thick overlying levels of laterdebris. But along with this advantage must be mentionedtwo significant drawbacks.

The first is that, except for Uruk and a handful ofother, generally much smaller towns, there are few siteson which the Early Dynastic I levels were either the latestor the most extensive. And while the solid-footed gobletsunambiguously indicate an occupation at this time, thelater remains at most of the larger urban centers oftenmake estimates of size for this period little more thanspeculative.

Second, the greater duration of the period, togetherwith the central fact of Uruk's enormous growth, createsa new series of ambiguities connected with the possibility

82

oi.uchicago.edu

\ \IZ1SI

'91-,4\

C

~P

L

I1 ''

- -'*

IP I R -

I/ 1096

/-~

k·

/

11666-

v -.

01237

c$

J'

C-

P

-. \

N/ -I, Y

/N

1/

- N

130U'

0 U '-U O OCCUPATION D* 20.1-40 POSSIBLY LARG* 40. OCCUPATION I

0 200., POLITICAL CAPITAL

0,

\\

*\306

0 * I

So L0 I0o Io

\ *o

\ \

0

* \ "

o o

S125* o 2 201

* /0

242 /*260

* * /\ 281 *N -038*-

P Ip I URUI(

(it ~ * *35

IOUBTFUL3E BUTJNCONFIRMED

0 15 KM. C -Y

±\

k 2N

0

o

.

oP

30

\

0382

o o

0

\ \\

\"

o /

'000-- ,I //

'^ \

Fig. 18. Jemdet Nasr period settlement patterns.

83

'- '

O-"

-- \ 4-

,p

1 -

T

I

/

^

%

v h,O

Jul

\

Slol0

oi.uchicago.edu

Urban Origins

of sequent rather than contemporary occupancy. Some863 hectares of Early Dynastic I settlement have beenrecorded in the Uruk region, more than double the totalfor the Late Uruk period. But almost half of this is repre-sented by Uruk alone. Since the latter grew by some 300hectares late in the period, probably at the expense ofthe smaller settlements, the aggregate increase may havebeen only moderate.

Where was the new population of Uruk drawn from?We know that there were reduced terminal occupationson many Early Dynastic I sites, and that the somewhatlarger settlements held on longer than the smaller ones(Adams and Nissen 1972, pp. 19-21). That may wellargue against the coercive abandonment of whole dis-tricts and for a more amorphous or fluid process in whichthe new city had a greater appeal for rural peasantrythan for those who were already settled in other towns.We cannot yet determine, however, what proportion ofthe outlying population had been drawn into Uruk as abasis for its greatly enlarged layout. Moreover, thatprocess may have been only a reproduction on a substan-tially larger scale of shifts that occurred much more gen-erally. Several smaller centers upstream from Uruk alongthe same river branch are also thought to have grownprodigiously at this time (sites 230-31, 242, perhaps alsoShuruppak). About several others on a more easterlybranch, the thick overburden of later debris limits us toassigning relatively modest site sizes with a large marginof uncertainty (Zabalam, 169; Umma, 197; Bad-Tibira,451; in addition to sites 168 and 198). There is amplepossibility, in other words, that much of the urbanizationprocess either precipitated or followed abandonments

15

0)EQ

E

C,

S10-

M 5-E

4

6

elsewhere, and that this movement toward succes-sively larger centers climaxed in but was not limited toUruk.

If so, however, it is at least possible that there were twoor more successive phases of abandonment. The first ledto movements from the countryside into many districtcenters. Subsequently there were further movements intothe handful of much larger, militarily more potent urbannuclei. Just as Uruk in its most extended state was inlarge part a substitute for rather than an addition toearlier settlement, so, according to this reconstruction,the area of settlement outside Uruk and before its growthmay be a composite. A misleadingly high total of settle-ment is reached if we neglect the possibility that latercenters combined with and replaced earlier ones ratherthan adding to them.

Problems of this type are, to some extent, unavoidablewhen a surface reconnaissance compels us to work withrelatively long chronological periods that cannot be sub-divided by actual stratigraphy. They are exacerbated inthis case, however, by the profound impact of Uruk'sgrowth. Accordingly, it is difficult to analyze the settle-ment data in the same way as was done for the Urukperiod, in spite of the relatively high degree of certainty asto which sites were occupied and which were not. As withestimates of population density, site hierarchies and rank-orderings depend for a great part of their significanceupon the assumption that the individual componentswere largely contemporary in their occurrence (Hodderand Orton 1976, pp. 72-73).

With this apologetic preamble, figures 19 and 20 sum-marize the available data in a manner essentially similar

4

rI ilppur-14lao rnvrons

3]Si I0.10 3 4 6 8101520 250710203040

0.5 1.0 2 3 4 6 8 10 15 20 25 50 75 100 200 300 400Upper Limit of Size Categories in Hectares

Fig. 19. Distribution of Early Dynastic I site areas by region.

84

15 ki.... A JL C

U.-jA--AW-

Fý4c-15- 1

R I m

oi.uchicago.edu

Urban Origins

Area in Hectares

Fig. 20. Early Dynastic I period settlement rank size.

to that adopted earlier. The size units in figure 19 areslightly different at the upper end of the scale, partly toaccommodate the increased size of Uruk and partly be-cause of the difficulties I referred to above in measuringthe size of some of the larger Early Dynastic I sites. Thosedifficulties perhaps account for some of the irregularityof the rank-size distribution given in figure 20. In spite ofits roughness, the rank-size graph clearly indicates theposition of Uruk as a primate city that must have dom-inated a wide region. If anything, in fact, the diagrameddistribution considerably understates its primacy in thelatter part of the period, at a time when Uruk's enormousgrowth had been made possible only by the progressive,collapse of the remainder of the settlement hierarchy inits surrounding hinterlands (Adams and Nissen 1972,pp. 19-21). Assuming that this process of urban expan-sion was accompanied by a greatly enlarged radius ofdominance of the city over its hinterlands, it no longerseems reasonable to replicate the practice followed forthe Uruk period (fig. 17) of plotting a rank-size distribu-tion for sites within Uruk's immediate area alone. Hencethe graph combines the northern with the southernregions.7

The unprecedented size of Uruk introduces problems

about the size of its sustaining area that heretofore havenot been dealt with, even though the same problemsapply in lesser measure to earlier stages of its growth aswell as to other, smaller centers. Too little is known ofUruk's urban layout at this time for its population to bemore than guessed at. Numerous public buildings ofmonumental size or large open areas within the wall(although some Early Dynastic I ruins extend outside it;cf. Nissen 1972, pp. 797-98) would sharply reduce thetotal. On the other hand, the constraint of living withinthe wall is equally likely to have induced higher popula-tion densities in residential areas than obtained in outly-ing towns and villages. At present, therefore, we can onlyhope to suggest the order of its size by assuming the sameconstant relationship between areal extent and populationsize that has been worked out-with much unexplainedvariance and no wide agreement even on averages-primarily for smaller settlements (Adams and Nissen1972, pp. 28-30; see further discussion above, p. 69, andbelow, p. 144). Covering approximately 400 hectares(about 1.56 square miles), Uruk appears likely on thisbasis to have had a population of certainly no less than40,000 to 50,000.

Further uncertainties appear when we seek to translate

85

lank

oi.uchicago.edu

Urban Origins

the food requirements of a population of that order intothe radius of the cultivated region that presumably sur-rounded the city. Bulk transport of agricultural producefor a considerable distance by boat or barge certainlywould have been feasible, especially from subordinatetowns or regions farther upstream, but the existing rec-ords provide no clue to its actual importance at thisearly period. Nor was agriculture the exclusive source offood; fish and the products of animal husbandry arecertain to have contributed importantly as well. Particu-larly for fishing, one can argue that the very substantialgrowth of the industry was among the conditions thatmade Uruk's ascendancy possible. The term for fisher-man, at least in Old Babylonian times and quite possiblyalready in the period under consideration here, also meanta military auxiliary serving under the crown. As MichaelRowton has perceptively noted, fishermen constituted astrategic reserve of military manpower, since their activi-ties, unlike husbandry or agriculture, could be abruptlyand indefinitely suspended in an emergency (1969, p.309). A hypertrophic growth in their numbers, beyondwhat might be superficially considered adequate to meetUruk's own need for fish as a source of animal protein,would thus significantly strengthen the city's politico-military posture. But products of cultivation are in anycase likely to have been calorically and nutritionally thedominant components in the diet, and they provide theonly available basis for an admittedly rough and provi-sional calculation of the breadth of the cultivated zoneon which Uruk largely depended.

The component activities going on within this zonewere clearly complex. To judge partly from contemporarybut primarily from somewhat later cuneiform sources,the primary crop was six-row barley; its green shoots atan early stage were a source of fodder, and the maturecereal was consumed both as bread and as beer. Othergrains, roughly comparable in productivity per unit area,included emmer, several other varieties of wheat, and mil-let. But in addition dates were surely an important foodsource, and date wine as well as grape wine was in use(Burrows 1935, pp. 10-11). Unlike grapes, however, thedate palm appears to have been "perfectly adapted to theecologic conditions of the region" (Guest 1966, p. 62).Hence the date palm, its fruit, and its by-products traceout a particularly complex, widely ramifying web thatextends from the subsistence economy into many otherbranches of technology as well as into thought and artmore generally (Landsberger 1967). Turning to otherforms of more intensive garden cultivation, vegetables in-cluded onions (the basic term for garden-ki-sum-ma-isliterally "the place of onions"), beans, chick-peas andother pulses, garlic, and a variety of other alliaceous plantsas well as fruits (Deimel 1925 a, b; Jacobsen 1958, pp. 10-11). Higher productivity on these latter plots is suggestedby the more careful and extensive preparation they re-

ceived, as well as by suggestions that they were equivalentin value to areas six times larger in ordinary fields (Edzard1968, pp. 51-52). But on the other hand the areas arecomparatively minor in relation to these devoted to fieldcrops, often accounted for only in individual furrows andseldom amounting to more than a few percent of thelarger cultivated plots in which they were situated (e.g.,Bauer 1967, pp. 74-84). Adding together these divergentconsiderations, probably we can do no better than to con-sider provisionally the productivity of agriculture in gen-eral as only slightly higher than that of the barley cropthat was its major component. In the same admittedlyapproximate terms, it is reasonable to follow Johnson'sestimate, derived from measurements under roughly com-parable dietary conditions among modern Arabo-Persianvillagers in Khuzestan, that some 278 kilograms of barleyrepresented the average individual consumption per year(1973, p. 97). An only slightly higher figure obtained asthe standard subsistence allowance during the Third Dy-nasty of Ur (see below, p. 146).

On the basis of contemporary yields in Khuzestan aver-aging 1,153 kilograms per hectare, Johnson further esti-mates that the sustaining area per individual for barleyalone was about 0.25 hectare, and he goes on to assumethat the same amount in addition would have been suffi-cient to account for "gardens, orchards, other field crops,surplus barley production, fallow land, and so on" (John-son 1973, p. 98). Some refinement in his calculations ispossible by using recorded barley yields from late EarlyDynastic Girsu. Crops in individual fields there are re-ported to have ranged between 742 and 2,794 liters perhectare and to have averaged 2,030 liters per hectare(Jacobsen 1958, pp. 36-37).8 The average yield approxi-mated 1,254 kilogram per hectare, 9 percent higher thancontemporary Khuzestan. This might seem to indicate thateven less than 0.5 hectares of cultivable land per personwould have been sufficient. On the other hand, there isgood evidence that a fallow system was just as integral apart of field cultivation in the third millennium B.c. as itis today. Jacobsen (1958, p. 65) has called attention, forexample, to the fact that in late Early Dynastic harvestrecords for successive years a given field is ordinarily listedonly every other year. With an allowance for fallowing,therefore, at least 0.35 hectare would have been needed tomeet the primary producer's minimal subsistence needs inbarley. Then there were additional grain disbursements ofconsiderable magnitude. At least to judge from contempo-rary experience in the region, for example, 25 percent ofthe crop normally must be allocated to losses in storage.Further provisions of grain, currently averaging about 16percent, must be fed to animals. Reserves for seed wereeven more indispensable, although the 11 percent cur-rently withheld is probably much higher than was thecase in antiquity (Poyck 1962, p. 53; Wright 1969, p. 21),and an almost equivalent amount of barley had to be set

86

oi.uchicago.edu

Urban Origins

aside to feed the plow oxen (Hallo 1976, p. 40). When allthese factors are taken into account, something well over0.5 hectare, perhaps closer to 1 hectare per person, wouldhave been necessary for alternating barley and fallowfields alone. Additional although much lesser amountsof land also must have been set aside for other forms ofcultivation.

But more important still, to an estimate of cultivableland per person we must add land that was not cultivatedeither because of submarginal productivity or because itwas given over to pathways, and also canals, swamps, andsteppelands that were not easy to irrigate. The amount ofsuch land is never directly recorded, but it may be indi-rectly approached, again in the sole case of late Early Dy-nastic Girsu, by dividing what appear to be the totalholdings attributable to the city by the population whoseactivities are listed in its archives. At least in this sourcethe total-uncultivated as well as cultivable-land perperson is on the order of 1.5 hectares per (adult) person(Adams 1965, pp. 23-24). And on this basis in turn wecan estimate that the intermittently cultivated ring aroundUruk would have needed to extend outward from thecity's center for no less than 14 kilometers.

At this point, however, a new set of considerations en-ters, for 14 kilometers is well beyond the limits of com-muting to fields daily. It presupposes temporary farm en-campments that would have been in steady overnight useafter the long days of the two most intense periods ofagricultural labor, the sowing and the harvesting (Adams1965, pp. 14-15). This distance is so great, moreover, thatit may call into question the underlying assumption ofregularity in patterns of land use. Chisholm, surveying awide variety of sources on traditional agrarian societies,has calculated that net product generally falls by morethan half when the distances between settlements exceed8 kilometers, and he maintains that fields are unlikely tobe cultivated at distances greater than 3 or 4 kilometersunless there is "some very powerful constraining reasonwhich prevents the establishment of farmsteads nearerthe land" (1970, pp. 112, 131). In the case of Uruk, theinterest of the state in asserting unchallenged authorityover the large, newly urbanized population it had recentlybrought together might represent just such a constraint.So might a prevailing condition of insecurity in the coun-tryside. But is there anything to suggest a pattern of con-centric zonation in which the seemingly heavy deterrentsto cultivation of the outermost ring led to progressivelymore intensive agriculture as one moved inward towardthe city walls? If not, how and why was the "virtuallyaxiomatic" dependence of agricultural intensity and henceproductivity on the size and proximity of population cen-ters (Skinner 1977, p. 283) somehow avoided in theseparticular circumstances?

Archaeology is not yet of any help whatever on thesematters, and the texts are, as usual, unconcerned with the

queries of the modern analyst and hence very elusive.Nothing locates individual gardens, orchards, or fieldsfor us in relation to the late Early Dynastic city of Girsu,once again our only potentially informative example. Butfurrow-spacing and seeding rates are perhaps worthnoting as at least indirect clues. While there was some vari-ation, both point in the direction of a strikingly nonin-tensive, relatively uniform system. Furrows were posi-tioned from 50 to 75 centimeters apart, for example, andindividual seeds were planted more than 3 centimetersapart within the furrows. Seeding rates generally variedbetween 20 and 27 liters (12.3-16.7 kilograms) per hec-tare. The basis for the difference remains to be explained,but even the larger figure is little more than a third of thecurrent standard in the same region and only a sixth ofthat in the United States (Pettinato and Waetzoldt 1975,pp. 278-81; Jacobsen 1958, pp. 62-63). Perhaps this maybe partly explained as an adaptation to insecure irrigationsupplies, but the very existence of summer-cultivated gar-dens and orchards attests to the capability of assuring ade-quate year-round water in favored locations. Why weresimilar investments of labor not devoted to intensifyingthe cultivation of the major food crops in areas mostaccessible to the city?

One possible answer emerges. It is of considerable im-portance for understanding the character of Uruk as acity, and hence highly regrettable that for the present itmust remain quite speculative. Suppose that a considerableproportion of the lands around Uruk, or at any rate theentire outermost ring, was worked not by its own citizenrybut by dependent laborers from the villages and steppe-lands beyond. Then the inducement to intensificationwould largely disappear, and the standardization of theprocess would be merely the normal impulse of bureau-cratic control. But then we are brought face to face withthe possibility that a correspondingly higher proportionof Uruk's own population was engaged largely in sec-ondary and tertiary economic activities rather than inagriculture. And that in turn would imply a stronger rolefor the city in the economic integration of the countryside,as well as some fairly reliable administrative control overthe necessarily quite distant settlements from which itsvital supply of agricultural labor was recruited.

It must be stressed that this is a fairly speculative foray.The collapsing network of Early Dynastic I towns andvillages in the Uruk hinterlands tends to contradict theidea of a stable, territorially extended base of admin-istrative control. Indeed, the first consequence of stablepolitical control should have been a wide dispersal of therural population, rather than its aggregation in separateurban nuclei that not only reduced agricultural produc-tivity but were at least potentially competitive. Most of theavailable evidence suggests that, even if agricultural laborwas drawn from a ring of surrounding villages within adistance of perhaps 20 kilometers, the demographic and

87

oi.uchicago.edu

Urban Origins

economic impact of Uruk was largely concentrated withinthe same relatively limited area.

Turning northward, we are again thwarted by an in-ability to establish the size of most of the larger centerswith any reliability, Kish, although outside the immediatearea of this study, is likely to have already been of de-cisive political importance. "King of Kish" was a title laterborne by some of the important rulers of southern Sumerduring the Early Dynastic III period, by which time thelarge palace that has been partly excavated there lay inruins. Hence its period of greatest power, if not popu-lation, occurred somewhat earlier. By Gibson's estimates,based largely on systematic surface reconnaissance, it in-creased in size from something less than 7 hectares com-posed of two separate settlements in Uruk or Jemdet Nasrtimes to a twin city of almost 60 hectares in the EarlyDynastic I period. Complicating that estimate, however,are the somewhat more extensive remains of the EarlyDynastic II/III period, which apparently total about 84hectares (Gibson 1972, pp. 118-22, 266-67). Moreover,there is a strong possibility that the mounds he was ableto record have been partly submerged by alluviation andoriginally were connected with one another to form sub-stantially larger settlements. Tell al-Wilaya is another dif-ficult case. Early Dynastic I remains are not reported bythe excavators, probably because the relatively brief cam-paign at the site failed to penetrate below its uppermostlevels. Hence there is no basis for assigning a size to it.Its isolated location north of ancient Adab is shown infigure 21 with the symbol for a small town (4.1-10 hec-tares), but this purely speculative estimate is not employedin any calculations of population distribution or density.Yet if Postgate's recent-highly plausible but admittedlynot definitive-identification of the site with ancient Keshis sustained by further work, this would be a significantomission. As he notes on the basis of seal impressions fromthe archaic levels at Ur, Kesh must have been "a majorcity in the ED I/II period" (1976, p. 81).

Nippur's occupation at this time, although of uncertainarea, is well documented from an excavation sequence inwhich the earliest phase of the Inanna Temple overliesclosely grouped Early Dynastic I private houses. Solid-footed goblets accompanying later debris on Adab's sur-face definitely also attest an occupation of some size.Around Adab a number of new settlements can be seenin the settlement pattern map of the period (fig. 21), theonly example in this area of localized growth. It appearsto argue that Adab had indeed emerged at this time asa center of some importance. Equally significant, it alsosuggests that the ingathering of countrymen into Urukwas no longer a force 65 kilometers to the north. Figure19 therefore provides a separate histogram for the north-ern area, indicating the distribution of sites by size in anarea beyond at least Uruk's demographic influences. Butreference to the cluster of smaller settlements surrounding

Adab only heightens the anomaly of Nippur and perhapsAbu Salabikh. Already centers of very considerable im-portance, they appear to have had no near neighbors andvirtually no dependencies.

The histogram for the Adab-Nippur region, concedingits speculative basis with regard to the major centers, in-dicates a continuing movement out of the countryside andinto those centers. Apart from the Adab area, there wereonly a handful of small outlying settlements left by EarlyDynastic I times; that handful virtually disappeared soonafterward. Even in the Adab area there was only one sub-stantial but clearly subordinate town (site 1421), and thattoo failed to outlast the period. Yet, if the estimates givenfor individual site areas in table 7 are not very substan-tially in error, there was certainly no further decline inpopulation in the north. Occupations are attested withreasonable certainty at thirty-five sites aggregating 212.4hectares of settlement, slightly more than the Late Uruktotal and probably also more than the Jemdet Nasr total.This stability is perhaps to be interpreted as further evi-dence that population movements around and into Urukand those farther north were essentially distinct from oneanother., Urug's links with its smaller northern counterparts are

thus likely to have been either religious, involving largelyvoluntary offerings, or politicomilitary, involving periodicreimpositions of tribute backed by threat of armed force,but not the steady, undramatic, cumulatively heavier costsof a colonial administration. This, at any rate, is the in-terpretation to which the stability of population in thenorth lends support. Had there been a genuinely unifiedregime over the entire region, moreover, the continuingtrend toward urbanization in both south and north wouldbe inexplicable. The entire historic record of Mesopo-tamian settlement makes clear that stable, centralizedregimes promote dispersion of the agricultural popula-tion into the countryside, closer to the fields, rather thanits concentration around towns and military strongpoints(cf., e.g., Adams 1965, p. 22).y We are dealing, therefore, with a discontinuous fabric ofadministration as well as settlement. Many communitieshad come (or been brought) together in a number oflarger, more competitive but coexisting centers, includ-ing an altogether unprecedented conurbation aroundUruk. But these centers remained weakly and perhaps onlysporadically articulated with one another. Characterizedby fluctuating ranges of political and economic influence,they formed neither a stable administrative hierarchy northe core of an attached zone of continuous cultivationthat might have required unified irrigation management.Impressive growth in the maximum size attained by thelarger centers in successive periods presumably reflects,and indeed may have required, corresponding growth inadministrative complexity or "density." That growth waslargely localized within those centers and a few of their

88

oi.uchicago.edu

4, \'9 \I,1 \\ '

\\, \@ \\•

\0 ! \,"- \ J

&(iC

A

0

1421 0-.

o SHURUPPAK 0

0/

/

1010 *\ o

/ f 242 .

I s,& \I

' \ URUK*' - -&, luu~ /

Fig. 21. Early Dynastic I period settlement patterns.

89

\ ((\

IP

U.N

N1 / ~

14

I

Ni

)0

/

3 \

*\I

3820

BAD

/

/

tpC," \

I

1

/

I

L

oi.uchicago.edu

Urban Origins

nearer dependencies, however, rather than extending todistant towns and ramifying into the countryside. In otherwords, there was no regional system with regularly spacedcities as its transport, service, and administrative nodes.In spite of considerable intercommunication and culturalhomogeneity, a tenaciously surviving element behind thekaleidoscopic shifts of population we have traced is thatthe settlement pattern remained a patchwork of con-stituencies.

Figures 22, 23, and 24 summarize the trends in settle-ment during the span of time primarily dealt with by thischapter, from the Early Uruk period through Early Dy-nastic I. They also document the existence of large inter-vening areas with neither settlement nor cultivation-areas whose location changed during an interval ap-proaching a millennium in length, but that remained acharacteristic feature in all successive periods. The ana-lytical units utilized for these representations are culti-vated areas rather than individual sites or the populationsassumed to have inhabited them. To be sure, site loca-tions and populations, the latter calculated on the basis ofour aforementioned assumption of one hundred personsper hectare of site (or more accurately the rectangular areaenclosing the site), constituted the essential input for thecomputer program by which these figures were generated.Taking our further assumption of 1.5 hectares per per-son as the minimum needed to cover subsistence needsin cultivable as well as unavoidably uncultivable land,this program has been designed to stimulate decisions onuse of land involving minimal transport distances andhence the application of a uniform "least effort" princi-ple in agriculture. For this I used an iterative proceduresimulating regular growth in a two-dimensional planearound the individual sites as point locations. The nearestavailable unoccupied land around each site was system-atically searched out by the computer, taking account ofthe simultaneous needs of its neighbors, in incrementalunits of population until the needs of even the largest cit-ies have been met. 9

Also suggested in this sequence of figures are bounda-ries for the major northern and southern enclaves of set-tlement whose similarities and differences have been amatter of concern throughout this analysis. The boundarylines themselves are of course only arbitrarily drawn ab-stractions, forming polygons intended to include the twomajor clusterings of sites and their associated areas ofcultivation. The boundaries shift from period to periodto take account of shifts in the patterns of settlement,but in all three cases they have been framed in essen-tially the same way and exclude only a handful of iso-lated, outlying sites. Hence the demographic character-istics of these polygons can be considered an alternativeform of expression of the principal findings of this chap-ter, as shown in table 6.

Considering table 6 together with the sequence of fig-

TABLE 6 Changing Characteristics of Late Prehistoric andProtohistoric Settlement Enclaves

Period Northern Enclave Southern Enclave

Early-Middle Area: 2,087 km2 Area: 2,010 km 2

Uruk Estimated Estimatedpopulation: 38,540 population: 20,110Density: 18.47/km2 Density: 10.00/km 2

Late Uruk Area: 1,619 km2 Area: 2,231 km2

Estimated Estimatedpopulation: 21,300 population: 41,020Density: 13.16/km2 Density: 18.39/km2

Jemdet Nasr (Ambiguities in data do not permitcomparable estimates)

Early Dynastic I Area: 1,184 km2 Area: 2, 938 km2

Estimated Estimatedpopulation: 20,240 population: 86,300Density: 17.09/km 2 Density: 29.37/km 2

ures, it can be seen in summary that initially the largestpopulation concentrations were in the north. Large andimportant centers, surely to be identified as already urbanin many of their features, made their appearance almostimmediately. A considerable proportion of the population,however, occupied a mosaic of smaller towns, villages,and outlying hamlets. As time went on the northern en-clave at first underwent a sharp reduction, many of itsinhabitants apparently moving south or southeast intohinterlands around Uruk that formerly were very sparselyoccupied. Later the northern enclave tended to stabilize,although concentrating around a smaller number of cen-ters and occupying a considerably reduced area. Mean-while the area as well as the population of the southernenclave continued to grow, although one must bear inmind that a considerable proportion of the inhabitants ofUruk probably have been counted twice in the Early Dy-nastic I population aggregate, since they also are creditedto the smaller settlements from which they originallycame. But the single most striking characterization of thesequence is that relatively small, isolated pockets of landuse coalesced into larger agglomerations. And the latterin turn were either engulfed or replaced by smaller num-bers of still larger, more tightly consolidated zones, cen-tering in most cases on the important towns of the earlyhistoric periods that followed.

Hence it is the continuing urbanization across the en-tire region that is the most striking feature of the EarlyDynastic I period. Around Uruk in the south, at the apogeeof the pattern, some 81 percent of the total of recordedsettlement was in sites larger than 10 hectares. In thenorth, urbanized earlier as a by-product of the massmovement of the rural population south into the Urukarea, the proportion of such settlement held approxi-

90

oi.uchicago.edu

Fig. 22. Simulation of Early-Middle Uruk period cultivated areas.

91

oi.uchicago.edu

NIPPUR

0

Fig. 23. Simulation of Late Uruk period cultivated areas.

92

() ·

oi.uchicago.edu

NIPPUR

0

0

Fig. 24. Simulation of Early Dynastic I period cultivated areas.

93

oi.uchicago.edu

Urban Origins

mately steady at 71 percent. During Early Dynastic Itimes if not earlier, in other words, southern Mesopotamiahad become the world's first predominantly urbanizedsociety.

Why was this so? Is an explanation to be sought pri-marily in the internal dynamics of a few centers likeUruk-for example, in the attempt of urban elites toenhance their power and prestige by enlarging the numberof people over whom they could exert immediate con-trol? To what extent was urban growth instead a sys-temic, relational phenomenon, perhaps linked to thegreater offensive and defensive advantages that largerand larger population concentrations conferred in a con-text of progressively worsening intercommunity rivalries?Was the final impulse that transformed Uruk into whatwas surely a major center of political power (if perhapsnot a capital city in the formal sense) merely an episodein an essentially continuous process that was already un-der way in the Early Uruk period? Or were the forcesresponsible for its initial formation primarily of a re-ligious character and therefore fundamentally different?If the latter, is Enmerkar's much larger city to be under-stood primarily as the relatively sudden, conscious impo-sition of a new form of settlement by a newly emergentpolitical elite? Was it, in short, a social "invention" withenhanced adaptive potential under certain specified con-ditions for those who adopted it, or was it the moregradually emerging outgrowth of smoothly evolving pro-ductive forces and institutional forms?

Questions like these are given prominence and urgencyby the data presented in this chapter, but they are largelyunanswerable in the present state of our knowledge. TheEarly Dynastic I period has as yet received disproportion-ately little archaeological attention, and, except for thosefrom Ur, the texts yet discovered to guide the interpreta-tion of prevailing institutions and social conditions arealso relatively very few and poorly understood. Nor canwe dismiss the possibility that there were ongoing en-vironmental trends that significantly speeded or otherwiseinfluenced what has been described above as a processknown and intelligible in demographic and social termsalone. Recently it has been urgued, for example, that theperiod from 5500 to 3000 B.C. was characterized by "aconsiderable increase in precipitation as compared withtoday," whereas the following period until 500 B.C., aftera moderately abrupt transitional interval, involved "asmall decrease in precipitation as compared with today."1'If this was so, it can hardly be accidental that widespreadtendencies to concentrate in urban centers of unprece-dented size coincided with heightened competition overreduced volumes of irrigation water.

But these are matters for a future research agenda, notto be resolved with the limited evidence of a surface recon-naissance. The achievement of urbanism is indisputable,and at least the basic quantitative dimensions of the proc-ess can be said to have been established. On this founda-tion a much refined and intensified program of furtherwork can and should proceed forthwith.

APPENDIX A

THE SURVEY DATA BASE: URUK-EARLY DYNASTIC I SITES,COLLECTIONS, AND CHRONOLOGICAL INDICATORS

The developmental trends that are the substance ofchapter 3 seem to have their inception in the Early Urukperiod and to culminate in the enormous growth of Urukas an urban center in the Early Dynastic I period. Forthe intervening span of time, the data most relevant toanalyzing those trends have been assembled in this ap-pendix. The reader may wish to refer to the general sitecatalogue.(chap. 7) for additional descriptive materialon individual sites, but the fuller, more systematic bodyof information on their size and dating will be found here.

Table 7 is concerned with three classes of informationfor all pertinent sites. The first is their apparent span ofoccupation within the Uruk-Early Dynastic I interval, asdetermined by ceramic and other dating criteria that aredefined fairly systematically later in this discussion. Alsoincluded for sites occupied during all or part of this inter-val, to provide an element of continuity, are any occupa-

tions for which there is evidence during the precedingstandard and Late Ubaid (Ubaid III and IV) or followingLate Early Dynastic (Early Dynastic II/III) periods.

Second, the table lists specific dating criteria that wereinvolved in making most of the individual site assessments,insofar as they were recorded (or could be identified lateron collection photographs) rather than merely taken intoaccount as an informal basis for dating during site visits.At some forty-seven of the sites listed in table 7, numerous(more than eight) criteria were observed whose presencesuggest an Uruk or Jemdet Nasr period dating (or both).To simplify the lengthy listing, in these cases reference isgiven in the right-hand column to a separate display inTable 8. For thirty-six additional sites of the same date,quantitative (rather than presence-absence) tabulationswere made possible by one or more intensive, localizedsurface collections. These are also indicated in the right-

94

oi.uchicago.edu

Urban Origins

hand column, by reference to table 9, wherein the ob-served frequencies are tabulated. For brevity, the ceramicand other indicators used for dating are in all cases iden-tified by letter codes, whose translations are providedin the text following table 9. The right-hand column alsorelates dating criteria to appropriate illustrations or de-scriptions in the earlier report of the Warka survey (Ad-ams and Nissen 1972) and other publications.

The third category of data in table 7 consists of esti-mates of occupational area for successive periods and sub-periods. All three of these categories involve substantialdifferences in reliability or degree of certainty from site tosite, as well as from period to period within a site, and anattempt has been made to convey these differences as well

as the basic categories of data through the use of symbolsshown in the key to the table.

In most respects, the dating criteria or "index fossils"relied on in this study are similar to those used in previoussurveys. Hence a detailed discussion for each successiveperiod may in some cases be partly redundant. The readermay wish to consult published accounts giving earlierversions of the basic sequence of diagnostic surface ma-terial (Adams 1965, pp. 126-34, figs. 11-16; Adams andNissen 1972, pp. 97-104). Some effort is made below,however, to call attention to regional variations and tonewer publications of material that permit improvementsor modifications in earlier chronological understandings.

TABLE 7 Periods of Occupation, Estimated Areas, and Diagnostic Surface Materialsfor Late Prehistoric and Proto-historic Sites

KEY

Periods1 Late Ubaid (Ubaid IV)2 Early Uruk3 Middle Uruk4 Late Uruk5 Jemdet Nasr6 Early Dynastic I7 Late Early Dynastic (Early Dynastic II/III)/ Periods on either side of slash cannot be differentiated with available evidence.

( ) Parentheses enclose poorly attested or doubtful periods, or periods of sharplyattenuated occupation.

T: "Trace" of occupation (1-2 sherds), possibly attributable to periods whose numberdesignations follow the colon; no attempt has been made to estimate size of occu-pation where this symbol is used.

Dating CriteriaSee pp. 116-27.

Area of Enclosing Rectangle of Occupation in Hectares± Poor collecting conditions, overburden of later debris, or similar problems make

indicated size relatively arbitrary.- Following size estimate, indicates that occupation during this period or subperiod

seemed less well represented (i.e., sparser dating criteria; possibly less densesettlement and/or shorter duration) than another period or subperiod for whichthe same size estimate is given.

Warka Survey (Reanalyzed)004 1Jidr

006

009

012

2/3 4 (5) 6 7- 10.0 ha ± 10.0 ha

(3) 40.3

(3) 44.7

(3) 47.8

AD, GB, HA,LG; many LA

LA

95

oi.uchicago.edu

TABLE 7-Cont.

Warka Survey (Reanalyzed)-Cont.013 (5) 6

2.0

018

020Abu Boga'

022

AD, GB, NA

(3) 4 55.8 ± 1.0

(2 3) 4 (5 6)S5.0 ± 1.0

22.5

21.0

21.0

(3) 42.9

023

024

028

034

041

042

AC, AD

7 See table 8

Adams andNissen 1972,fig. 33; seetable 8

See table 8

FB, LA, NF

AC, AD, EB,GB, JA, KF,LA, NA, NF

Many LA

(5) 62.9

2/34.2

T: 5 (7)1 2/3 4

± 3.0

(3) 4 50.8 0.8

044

047

048

051

053

054

059

060

068

070

071

2/3/4- 0.5

4 5± 0.5 1.4

5 62.0 2.0

6

1 (23) 43.6

T: 2/3/4

T: 2/3/4

6± 0.1

(3) 41.8

54.0

T: 2/3/4

3 46.7

50.5

AD

Adams andNissen 1972,figs. 34-36; seetable 8; also HA

AC, AD, BA,DA, EB, ED,GA, LA

AD, GB, HC,LA

See table 8; alsoEF, GB, KB

See table 8; alsoEF, GB, HA, KB

(5) 63.6 ± 1.0

GB

AC, AD, EA,EB, FE, GA,JA, LA, NA

AD

LA

AC, AG, LA

96

oi.uchicago.edu

TABLE 7-Cont.

Warka Survey (Reanalyzed)-Cont.076 (3) 4 5 6

2.0 2.0 2.0

077 (3 4 5 6)- 0.1

078 (5) 6Abu Dhuba' ± 2.0

079 (5) 69.0

081 (5 6)- 1.0

082 (3) 4 (5) 6- 0.5 6.0

2/3/4± 0.1

(3) 44.2

(3) 47.8

51.0

57.8

52.9

(3) 4 5± 0.1 0.6

5 619.0

5 6± 6. ± 1.0

2/3± 1.0

(3) 4 50.6 0.6

(3) 41.7

3 (4) 52.6 0.1

(3) 41.0

(3) 45.2

AC, AD, EB,FE, GB, GC,JB, LA

AD, GB

AD, FG, GB,MA

AD, GB

AC, AD, EF,GB, JA, KA,KF, LA, LC,LF, NA

JB-2, LA-1

AC, AD, EB,JA, KA, LA

Adams andNissen 1972,figs. 37, 81; seetable 8

AD, CF, NF

AC, AD, GC,KA, LA

AD, GB, NA

LA, many MB

AC, AD, LA

AC, EB, JA, LA

AC, AD, FB, LA,NA

AC, LA

Adams andNissen 1972,figs. 38-39; seetable 8; also FF

See table 82/3 49.0

7

97

083

086

087

091

095

101

102

103

105

106

107

108

109

110

oi.uchicago.edu

Warka Survey (Reanalyzed)-Cont.111 T: 6

112

114

115

118

119

120

123

124

125

AD-1, GB-1

AC, LA

AC, AD, LA

AC, EB, LA

See table 9

LA, LH (LA)

(3) 4- 1.0

(3) 42.9

(3) 41.7

5.32/3/41.3

2/3/42.02/3/41.62/3/42.6(3) 4

24.0

52.9

5 618.0 - 10.0

1 (3) 44.0

(3) 4 54.0 4.0

(3) 4 (5)1.8

2/3 (5)- 1.0 ± 1.0

5 6 712.0

(5)- 1.0

5

(5) 6± 0.1

6- 0.1

6 7

2.6

(3) 4 AC, EB, LA, LC0.2

1 (23) 4 Adams and1.5 Nissen 1972,

figs. 40-43; seetable 8; also HA

(3 4) 5 AC-1, AD± 0.1 -± 1.0

(3) 4 5 See table 82.0 0.5

98

TABLE 7-Cont.

Adams andNissen 1972,figs. 81, 82; seetable 9; alsoEF, GB, KBSee table 9

AC, AD, LA,NA-1

See table 9

AD, BA/B, EB,LA, NFAD, AE, AF,CF, GB, HC,LA-1, NAAD

126

127

128

129

130

131

132

133

137

139

144

oi.uchicago.edu

TABLE 7-Cont.

Warka Survey (Reanalyzed)-Cont.147 T: 6

148 T: 2/3/4 LA-1, MA-1

AC, LA, NB

AC, AD, LA, LC

(3) 46.6

(3) 4 52.0 2.0

2/3/4/5- 1.0

2/3±0.1

1 (23) 4 50.4 0.4

152

153

155

156

160

161

162

163

5 66.0

2/3 4 54.0 1.6

164Jid

166

167

168Smid169Zabalam/Ibzaykh170171

173

174

175

177

AD, LA

LA

AC, AD, EB,HA, HC, LA,LF, NA

AD (?), FG, LA

See table 8;also GB

Adams andNissen 1972,figs. 44-47; seetable 8

AD, FE, GB,LA, NA

AC, AD, EA,EB, FE, FF, GB,HC, LA, NA

AD, LA, LC

(3) 4 (5) 6± 1.0 ± 1.0

(3) 44.0

5 64.0

2/3/40.13 4 (56)± 5.0 ± 5.0

3 (456) 7- 2.0 ± 2.0 ±5.0

T:52/3- 1.0

(23) 4 52.6 2.6

3/4 5 6-- 3.0 - 9.06 7- 1.0

5 64.6

7

AD, LA, LG

Many LA

EF, GB, HC,KA, LA, LC

AD, GC

Adams andNissen 1972,figs. 48, 82; AD,EF, FD, GB,HC, KA, KB,JD, OB

2/3/4/50.1 0.1

(3) 46.0

oi.uchicago.edu

TABLE 7-Cont.

Warka Survey (Reanalyzed)-Cont.178 1 2 3 5

0.7 0.1Adams andNissen 1972,fig. 49; see table8; also HA

AD, EF, GB,KB, LE, NA

See table 8

2/3/4 5 6± 0.1 ± 3.0 18.0

2/3 4 55.8 5.8

(5) 63.6

(3) 44.4

2/3/4- 0.1

(3) 40.6

(5) 61.2

(5) 6± 0.5

(3) 4 (5 67)- 3.0 ± 1.0

AD, GB

Adams andNissen, fig. 82;see table 8

AD, EF, GB,KA, KB, LA

AD, GB, LE, LF

AC, AD, CA,EB, EC, FD,JB, JD, LA

Adams andNissen 1972,figs. 50-51; Seetable 8

See table 8

(2/3) 44.0

193

197Umma Jokha

198Umm al-'Aqarib

199

201

203

206

209

212

(2) 43.0

2/3/4-- 5.0

2/3/4- 5.0

54.8

2 311.0

LA, MB(56) 7

(5) 7

4 511.0

AD, LA, LC, LG

AD, EB, HC,NA

Adams andNissen 1972,fig. 52; seetable 8

AC, EB, JA(3) 40.6

(5) 6-- 0.2 ± 2.0

2/3± 1.0

5 66.8

AD, GB

AA, CE, IA,JA, JE, LA

AD, EF, GB,HC, KA, KB,LC

100

179

181

183Hammam

185

186

187

189

190Abu Bott

191

oi.uchicago.edu

TABLE 7-Cont.

Warka Survey (Reanalyzed)-Cont.215 2

0.1

1 26.5

3 46.5

LA, LB, LC, LE

Adams andNissen 1972,fig. 53; see table8; also HA

Adams andNissen 1972,fig. 54; seetable 8

AD

(3) 4 51.4 ± 0.5

220

229Qal'a Maltus

230Umm al-'Ajjaj

231

232

233

234

236

237

242Suheri

244

245Mismar

52.6

2/3± 0.1

2/3/4± 1.0

(5)24.0

LA

5 653.0

6

5 67.4

513.6

6

5 64.8 ± 1.0

(3) 44.8

2/3 40.9

(3) 4 5 6 7± 10.0 65.0

T: 2/3/4

1 2 36.0

4 5 617.0 -±0.5

T: 2/3/4

T: 2/3/4

(4)S0.1

258

260Abu 'Ilba

5 61.9

2/3- 0.5

(13) 4 5 614.0 14.0 ±3.0

101

See table 8; alsoAE, EF, GB

AD, EF, GB,KB, NA

Similar to 233

AD, EF, GB,HC, NA

Similar to 233;also CF, FD, JD

See table 8

AA, CA, IA,IB, JAAC, AD, CF,LA, LG, NA, OB

LA, LC

See table 9; alsoGB, GC;Schmidt 1978

HA-1, LA-1,LB-1, OA

Adams andNissen 1972, fig.55; AD, DA, ED,FD, GB, HC,LA, LE, ND

LA, MA-1

AD, GB, LA,LC, LF; HA-1

218

219

247

248

256

oi.uchicago.edu

TABLE 7--Cont.

Warka Survey (Reanalyzed)-Cont.261 (5) 6Abu Khawa ±2.0

2/3/43.5

2/3

7 AD, GB, LG

See table 85±2.0

4 5 61.9 1.9 ±0.2

(3) 4 50,3 -0.1

(3) 4-±0.5

50.1

(3)

5 61.7

262

264

41.1

(3) 40.5

5 63,0

LC

See table 8;also EE

See table 8;also KB

AD, GB, HC,JB, LC, NA

See table 8

AC, AD, CE,CG, EB, FD,GB, LA, LC, NA

Adams andNissen 1972,figs. 56-59; seetable 8; alsoGB, HCAdams andNissen 1972,fig. 60;

AC, AD, EA,FD, HA, MA,NA

See table 8; alsoEF, GB

CG, FD, LC

Adams andNissen 1972,figs. 61-62; seetable 8Adams andNissen 1972,figs. 64-65; seetable 8; alsoGB, KF

267

272

273

274

276

(6)±0.5

277

281

282

285

286

288

289

292

293

52.6

5+-0.3

50.6523.0

(3) 42.6

(3) 40.9

(3) 40.5

5 66.0 ±1.0

5S0.6

(3) 41.9

(3) 41.7

5±1.0

5 61.7 1.0

102

oi.uchicago.edu

TABLE 7-Cont.

Warka Survey (Reanalyzed)-Cont.297 2/3/4 5

1.6 1.6

305 T: 2/3/4

306 (5) 60.8

309 (3) 40.6

310 (3) 4 50.9 0.9

312 5 (6)4.1 -2.0

(3) 40.6

T: 2/3/4

2/30.1

2/31.1

50.1

4 (5)±0.1

52.2

314

315

317

318

321

325

327

328

329

330

331

334

338

339

347

350

See table 8

LC-1

AD, GB, KB,NA

CC, CF, EC,ED, FD, JC,JD, KC, NA

Adams andNissen 1972, fig.68; see table 8

Adams andNissen 1972, fig.69; see table 8;also EF, GB

See table 8;also FF

LA-1

LA, ME,NA, NE

EA, EB, LA

AD, NA, OA

AC-1, GB, HC

AD

62.0

50.4

2/3/4±0.1

2/31.2

(3) 40.9

(3) 40.4

42.6

T: 2/3/4

51.0

(3) 40.8

A few LC, MB

LA

See table 8

AC, CF, LA, NA

AC, LC

LA-1, LC-1

AD, KA

AC, LA

103

oi.uchicago.edu

TABLE 7-Cont.

Warka Survey (Reanalyzed)-Cont.356 T: 2/3/4

50.1516.0

(5) 6-0.5

(3) 4-2.0

5

7

AD-1,LA(?)-1

AD, NA

AD, GB (?),KA, LA

AD, GB-1,LG, NA

AC, LA

AD, HC

AD, LA

AD, CG, EA,FD, GB, HC,NA

AD, FE-1, LA

52.6

2/3/4 5-0.5 -0.5

5 62.5

(3) 4 5±0.5 -0.5

(2) 44.8

(5) 64.2

(5) 61.0

T: 2/3/4

2/3/4 51.6 1.6

5 612.6

(5)22.0

6

(45) 6±0.1 6.42/3 4

1.5

(3) 4 5-1.0 11.0

See table 8

AD, GB, KF

LA-1AD, LA, NE

Adams andNissen 1972, fig.70; AD, BA, CF,EE, FB, FD, GB,HC, KA, KB,KF, MD, NA

AD, EF, KA, NA

AD, ED, EF, GB,KF, LA-1

Adams andNissen 1972,figs. 71-72; seetable 8

Adams andNissen 1972, fig.73; AC-1, AD,EA, ED, EF, FD,GB, HC, LA,NA

6

104

357

358

365

367Abu Zumal

369Umqtaif'

370

372

373

376

377

378

379

380

382Baydha

383Mansuriya

384Twaimi386

387

oi.uchicago.edu

Warka Survey (Reanalyzed)-Cont.390 5 6

1.7

T: 2/3/4 6

51.0

2/3-0.5

(5)4.0

6

Adams andNissen 1972, fig.74; BB, CA, EE,EF, FD, GB,HC, NA

LA-1, GB-1

AD, FG, GB(?)-i, GC-1LA, NA-1

LA

AD, EF, GB,GC, JB, KA,KB, KE, KF, LF

AC, AD, EA,EB, JB, KA, LA,LC-1, OB

AC, AD, GB,LA, LC, LF,HA-1

AD, EF, GB, IA,KA, NA

LA, FG

(3) 4 50.8 ±0.1

(3) 4 5 66.0 6.0 -2.0

2/3S0.1

2/3±0.2

(5) 61.3

(3) 40.3

(3) 40.4

(5) 6±0.5 2.9

Adams andNissen 1972, fig.77; EE, EF, FD,GB, IA, JA, KB,KF, NA

AD, GB(5) 60.5

T: 2/3/4 LA-2

AD

AC, AD, EA,EB, JB, LA

HA

Adams andNissen 1972, fig.78-80; see table8; also HA

TABLE 7-Cont.

400

401

402

404Awayli

406

407

409

410

417

418

422

442

444

7451Bad-Tibira/

Medain

453

459

460Awayli

(56)-25.0

(3) 44.9

(1)1 2

11.0

53.0

3 4±1.0

105

oi.uchicago.edu

TABLE 7-Cont.

Warka Survey (Reanalyzed)-Cont.462 5 (6)

2.8 ±1.0

465

Raidu Sharqi

Shuruppak/Fara

Uruk/Warka

Deheshiya

Bed of Shatt al-'Ajauwiya

Nippur Survey

517

539

562

573

574

639

653

655

662

667

671

673

677

678

680

691

(5) 62.9

1 (23) 4±0.5

2/3±1.0

5 6 725.0

1 2 3 4 5 6 7-70.0 ±100.0 -100.0 400.0

2/3-0.5

(3) 4-0.5

T: 2/3/4

42.0

T: 2/3/4

T: 1 2 3 4±1.0 ±0.2

3 41.0

2/3±0.1

T: 2/3/4

2/3±0.1

50.2

2/3±0.1

2/3±0.5

T: 2/3/4

2/30.3

2 313.5

1 20.1

T: 2/3/4

AD, EF

AD, GB, JB,KB, KF

von Haller 1932,pp. 35-37

Schmidt 1931,pp. 200-202,211-214; alsoseveral LA

von Haller 1932,p. 3 4

von Haller 1932,p. 3 3

AC, CC, CD,CG, EA, FG,JD, LA

LA-1

See table 9

See table 9

A few LA;NE-1

LA-1

LA-5, widelyscattered

LA-6

A few LA

LA-1

LA, HB

BC/D, CD, CG,HA, LA, NB, NE

NE-1

4

106

oi.uchicago.edu

TABLE 7-Cont.

Nippur Survey-Cont.692

706

711

714

720

722

737

738

740

743

744

745

748

749

765

768

778

781

782

783

786

789

790

T:6

2/3±0.1

2/3±0.1

T: 2/3/4

T: 2/3/4

2/3 50.2 ±1.0

2/3/41.1

2/3/41.2

2/3/4±4.0

22.0

2/33.6

2/3/40.6

(4) 50.1 0.4

2/3±0.1

2/35.3

2/3/4±0.1

2/3/4±0.1

2/3 5±5.0 ±0.5

2/3-0.1

2/3±0.1

(3) 40.6

T: 2/3/4

3 45.5 5..

GB-1

LA-3

LA, NA-several

LA-1

LA-1

AD?(67)

Numerous LA;FE-2

Mostly LA

See table 8

CA, FB, HB, LA

AD-many;CF?; LA-few

LA-some

See table 8

LA-several

LA, NE-several

AA; AD-few;CA; GB?; LA,NA, NE. NoAC, EB/C

LA-several;MA-1

LA

(6)

T:5 AC, EA, EB,AD-1

LA-1

See table 95-

107

oi.uchicago.edu

TABLE 7-Cont.

Nippur Survey-Cont.792

793

797

798

801

802

803

804

805

818

821

824

826al-Arsan

829

831

832

835

837

838

845

853

854

858

26.8

20.4

T: 2/3/4

(5) 6±5.0

6±0.1

2/3±0.2

2/3/4±0.1

21.7

40.5

2/3-0.5

2/3±0.1

2/31.4

2/30.5

2/30.3

2±5.0

2±3.0

2/32.9

21.4

2/3±0.1

28.0

2/3±4.0

2/3±0.2

T: 2/3/4

7

See table 9

AA, CA, LA

LA-1

AD, BA, CC,FG, GB, GC,KA, KB, LG, NA

GB-several

LA-numerous

FG, LA-several

See table 9

See table 9

LA-numerous

LA-several

AA, LA, NF

LA, NE

AA, CA, LA; noAC, EB/C

FG-1; LA-many

See table 9

LA-several

AA, LC, NF;LA-many

LA; NA-1,NE-1

LA-several

LA-2

108

oi.uchicago.edu

TABLE 7-Cont.

Nippur Survey-Cont.876

912

924

928

933

935

936

937

939

940

944

945

946

947

950

951

952

953

954

959

960

961

964

975

976

977

T: 2/3/4

23.0

T: 6

(23 4)

T: 2/3/4

20.8

20.2

2/3/40.6

22.4

2/3 41.7 1.7

T: 2/3/4

50.1

T: 2/3/4

2/3/4±0.1

(5) 60.5

(23 45)±0.1

2 50.2 0.2

(5) 60.2

T: 2/3/4

T: 2/3/4

T: 2/3/4

T: 2/3/4

2/30.7

2 (3)5.8

2/3±0.2

2 (3)0.6

LA-1

No AC, EB/C

GB-1

LA-1

No AC, EB/C

AA, LA; no AC,EB/C

See table 8

See table 9

LA-1

AD

LA-1

LA-several,widely scattered

AD-many; GB

AD-1, HB-1,LA-several

AA, AD, CA,FF, IA, LE;LA-many

LA-1

LA-1

LA-1

LA-2

See table 9

LA-many

40.6

4 50.2 0.2

AA, BC/D, CA,CC, CG, FB,GA, IA, LA

109

oi.uchicago.edu

TABLE 7-Cont.

Nippur Survey-Cont.979

980

981

982

1002

1005

1019

1020

1021

1024

1027

1031

1032

1034

1036

1044

1046

1049

1050

1054

1056

1059

(4) 50.8 0.8

2/30.8

2-±1.5

2/30.1

T: 2/3/4

T: 2/3/4

21.4

2 (34)8.2 ±0.2

2/3±1.0

21.3

2/3±0.1

T:4 7

2/3 5 6±3.0 29.0? ±3.0

AD, HC, LA

No AC, EB/C,HA/B

LA-1

LA-2

AA, EB/C,LA, LB, NF;no AC

See table 9

LA-many

AA, CA, LA,NE, NF; no AC,EB/C

LA-several,NE-1

AF-1, FF-1

AD-many,widespread;GB/C, LA, LE,LG, NE,localized

LA-11localized

AD-few

LA-4, NE-1

2/32.6

(56)±0.1

2/3±1.0

28.6

(5 67)±0.5

T: 2/3/4

2/3±0.1

(5) 6±5.0

53.7

7

See table 8

AD-few

LA-1

LA--few

AD, GB

110

oi.uchicago.edu

TABLE 7-Cont.

Nippur Survey-Cont.1067

1069

1070

1071

1072

1087

1095

1096Umm al-Fugas

1099

1100

1103

1108

1109

1112

1113

1114

1115

1118

1119

1124

1129

1130

1131

1134

2/3±0.1

2/3-0.5

2/3±0.1

2/3-0.1

2 3 43.4

T: 6

T: 2/3/4

3 4 5 6±2.0 11.5 ±2.0

(2 3 4)

2/3/4 5±0.5 1.8

2/3±5.0

T: 2/3/4

2/3±0.5

2/3±0.1

24.0

2/3±4.0

2/3±0.5

2 (3)0.6

T: 2/3/4

2 3 46.8 ±1.0

2 3 4 52.4 2.4

T: 2/3/4

(3) 4±3.0

T: 2/3/4

111

LA-few

LA

LA-few

LA-few

See table 9

GB-1

LA-1

AA, BA, LA,NF; AD-many;GB-few

AD-many,LA-few,NA-1

LA-many

LA-1

LA-some,NF-1

LA-3

Many AA, CA,LA, NE, NF

LA-many

LA

See table 9

LA-1

See table 9

LA-1

LA-1

oi.uchicago.edu

TABLE 7-Cont.

Nippur Survey-Cont.1135Rubahiyat

al-Torra

1137

1152Hamayma

1154

1159

1163

1164

1165

1166

1168

1169

1170

1172

1174

1178

1179

1180

1185

1194

1195

1196

1197

LA-few2/3-0.1

2 3 4 5 63.8 3.8 0.1

2/3±0.1

(23) 42.6

24.5

40.9

2 3 41.0 0.2

3 45.3

2/3 5 610.6 10.6 ±2.0

50.1

2/3-- 0.5

2/3±0.5

2 3 425.5 25.5 -

2/3 T: 6+-0.1

2/3 (6 7)±0.1 ±0.1

T: 2/3/4

2/3±0.1

T: 2/3/4

2 3 411.5 11.5-

2/3+0.1

2 3 4±0.5 ±0.5

(3) 4 51.1 0.1

See table 9;also GB

LA-3

See table 9

See table 9

See table 9

See table 9

See table 9

AD, LA-many;GB-few;LD, LE

AD-many; FG,GC, LA, NA

LA

LA-many

See table 9

GB-2, LA-3

AD-many; AE,CC, CG, GC, LA

LA-1

LA-few

LA-1

See table 9

LA-3

LA-many; AC

AC, LA-many;ED, FB, JB,LE, NC

6

112

oi.uchicago.edu

TABLE 7-Cont.

Nippur Survey-Cont.1198

1199

1205

1207

1208

1209

1210

1215

1216

1217Dowayhis

1219

1221

1230

1233

1237Dlehim

1247

1261

1271

1272

1278

1284

1291

1293Abu Dhaba'

1294

1303

2 53.4 0.1

2/3±1.0

2 3 (4) 5 67.9 ±2.0

T: 2/3/4

2/3±-0.1

T: 2/3/4

2/3-0.1

(5) 64.8

(3) 4 5±0.1 4.8 ±0.5

2/3±0.1

T: 2/3/4

T: 2/3/4

2/3±0.5

T: 2/3/4

2/342.0

(4) 542.0

2/3 (5) 6-0.1 ±0.1

7

See table 9

AA, LA-many;CA, FG, IA/B,MB, LH (LA)

See table 9; alsoGB, GC, KF

LA-1

LA-several

LA-1

LA-several

AD, GB-many;ED, EF, FG,GC, KB, KF

See table 9

LA-2, NA, NE

LA-1

LA-1

LA-numerous

LA-1

AD, LA-many;AA, CA, JA

AD-many, GB,LA-few; AE,CF, FG, OB

See table 9

LA-many; AD,AE, AF

FE-1, LA-1

GB, LA

(3) 40.8

2/3-0.5

7

T: 2/3/4

2/3 610.1 +0.1

2/3±0.5

T: 2/3/4

(3) 4+5.0

2/30.8

T: 2/3/4

LA-14

LH (LA)-1

LA-many; AC,EC, FE, NA

LA-many

LA-1

113

oi.uchicago.edu

TABLE 7-Cont.

Nippur Survey-Cont.1304

1306al-Hayyad

1312

1315

1316

1318Hayyad

al-'Alwi

1337Tell Khathale

1353

1354

1355

1356

1357

1358

1372

1375

1383Bayt

1386

1394

1399

1405

1410

1413

1416

2/3-0.1

2

2

LA-few

3 450.0 50.0-

3 (4)0.2

(3) 4 52.8 0.5

2/3 5/6-0.1 ±0.1

2/3±0.1

2/3±0.1

3

5 6-20.0 ±5.0

See table 9; alsoGB, GC, KF

See table 9

See table 9

AD, LA

LA

BC-1, FB-1,LA-4

AD, LA-many;CA, CD, EA

AD-few

7

4 54.2 4.2

(5)±0.1

3 4 53.0 3.0

5±0.1

4 51.2 1.2

T: 2/3/4

T: 2/3/4

3 40.2

2/3±1.0

6

7

AD-few

AC-1, AD-several

GC-1, LA-1

LA-1

See table 9

AD, GC, KF-many; LA-6,NA-1, NC-1;AE

AD-numerous;LA-rare;AE, GC

See table 8

LA-2

LA-1

GB-1

LA-1

See table 9

2/3 5 (6)±0.1 ±0.5

4 57.7 7.7

T: 2/3/4

T: 2/3/4

T:6

T: 2/3/4

2 (3)0.6

114

oi.uchicago.edu

TABLE 7-Cont.

Nippur Survey-Cont.1421

1428

1430

1432

1434

1437

(5)18.0

6

2/3 (±0.5

(5) 6±4.0

2 31.7

2/3-0.1

2/3-0.1

1440

1442

1443Barghasha

1445

1448

2/3 (±-1.0

T: 2/3/4

2/3±0,1

T: 2/3/4

3 4

1450

1451

1454

1456

1459

1460

1465

1471

1571

1592

1602

1612

1615

0.9

(5)-2.0

2/31.4

6

5)±0.5

AD, GB-many;FG, GC, KA,KB, KF, NA

AD, GA, LA-several

6

See table 941.7

(5) 6±0.1

(5) 6±6.2

5)-1.0

6

AD-3, GB-2,LA-4

AD, GB-many;CC, EE, FG, GC,KA, KB, KF, LA

AD, GB, LAfairly common

LA-1

LA-few

LA-1

See table 950.9

7

(5) 61.4

(5) 62.1

(5) 61.1

(5) 60.1

2/3 (5)0.1 0.1

2/3±0.1

2/3 5±0.1 ±0.1

T: 2/3/4

T: 2/3/4

T: 2/3/4

T: 2/3/4

6

AD, FG, GB,GC, KB, LA, NC

AD, GB

AD, FG, GB,LG, NF

AD, AE, GB-rare

AD, GB, GC,LA-rare

LA-3, NE-1

AD, LA

LA-1

LA-1, NE-1

LA-1, NE-1

NE-1

See table 83 4 5 60.6 0.2

115

oi.uchicago.edu

Urban Origins

TABLE 7-Cont.

Nippur Survey-Cont.1618 T: 2/3/4 LA-2, MB-1,

NE-1

Adab/Bismaya

Isin/Bahriyat

Nippur/Nuffar

Tell al-Wilaya

(2 3 4 5) 6±50.0

(234 5) 6-10.0

1 2 3±25.0

(2 3 4 5) 6-±-10.0

7

7

4 5 6 7-25.0 ±50.0

7

Hrouda 1977,p. 55

Akkad Survey (Reanalyzed)

A 221 2/3-- 1.0

A 259 2/3±0.2

A 261 2/3±0.1

A 264 2/3±0.1

A 275 (1)Abu Salabikh

Harris andAdams 1957LA-3, HA/B

LA-3

LA

2/3 4 5 65.0 15.0 15.0 25.0

7 See table 8

Ubaid and Pre-Ubaid Periods

Very little information is available with which to elab-orate upon Nissen's discussion of ceramic indicators forthese periods in The Uruk Countryside (Adams andNissen 1972, pp. 98-99). It will be noted that virtuallyno new Ubaid sites were located after the survey movednorthward and away from the Uruk/Warka area. Theone significant exception is site 1604, situated less than20 kilometers north-northwest of Warka, although itwas not reached by the survey until 1975. Apparentlythis is a single-period site directly comparable to WS-298and is assignable to the pre-Ubaid and Ubaid I periods(Adams 1975a.)

Uruk Period

Before turning to a detailed chronological ordering ofthe collections, it is necessary to describe the typologicalsystem that was applied to the late pre- and protohistoricsurface materials and to indicate at least the roughchronological categories into which individual types ap-parently fall. To facilitate recording, a number of ceramic,stone, and metal vessel and tool types were given letterdesignations at the outset of the 1975 season. Some ofthese types, such as beveled-rim bowls (AC) and solid-footed goblets (GB) were of known common occurrence

and chronological significance. Others, such as clay sickles(LA) and stone bowls (NA), were known to be commonbut were regarded as probably having remained in verywidespread use over a long span of time. Recording oftheir frequency was planned as a clue to patterns of re-gional or functional differentiation more than to chrono-logical sequence, although in the sequel some of themalso were found to be of at least supplementary signifi-cance for chronological ordering. Still others, such aslow, flat-bottomed basins (DA) and triangular shoulderlugs on jars (EE), had served as useful indicators duringthe Warka survey but occurred too infrequently to beof much service in the area around ancient Adab, Isin,and Nippur farther north.

Once adopted, there were obvious incentives to retainthe coding of types consistently throughout the recon-naissance that followed. It gradually became apparent,however, that the normal range of variation of one type(CG) included examples that were only arbitrarily dis-tinguishable from two others (CC, CD); hence this wasultimately dropped. Similarly, four types were addedonly in time for the concluding weeks of quantitative re-cording in December 1975 (AG, BB, CB, CE), their utilityhaving gradually become apparent during the course ofthe main reconnaissance of the previous spring.

It was clear from the outset that the primary objective

116

- I I

oi.uchicago.edu

+

w.1 .pa

<

0'

z u

O

IWQM

Or

EOp

(((H

P-H

um

qr

P-l

,-1 .°

01-

HU

< P

Q

<,

u4m-

WC

I4UW

.

+++

++

+

4

S++

++++"4".

+++++ +t

++.

<uM c,

"a<L>

04-Jz4-J-

u7:1 0 U(-)

r.1

cri

'-5 oCým 0 CC0 .S0.V-4.

r;o'-jCS

00'-T c-H

00 ID30^Q0 (a

)C0000 o H<u

-

VI

c(D

CO

ooP^

(C

++++++++-++++++++<M

++++++++c

+++++++++++++++++++++0

+

+++++++++++++00O

+++++++++++++++++++,.-10

++.++++++++++++++0•

+

++++++.

+++++++++++++++o

++++++++++++++0

4444444444444444444 ~+444444+ +-+"*

4-4- ++++ +++++++ +

+++4-444'44 4

--

i 4+I+444++++4+

+++++++++++++++.+4++++ 4o4-,

4-4-

++++++ ++++++++ 4o

~++4-

,4444444444+4444+++ +c

+++++++4

+++++++++++ 4o

++4

++4

+44++++4++++++++++4

++++++4-++ +

++++++++, 4-++ 4.

+ 4-4-

+4-

++++++++++++++i

4'

+

+

+4-

+++++++++++±+ 4+

+++++

++++++++++++++ 4"

+ +

I-

44 4 44

\-4 4+++

• 4+e4

++ ++++4++

4+++.

4÷+4.+ 4-l

4+

+

+

+

4

+4++

-+4++

++-

+

++4

++

ii+

+4-

,4 e-

+

++

4++ +4++

++

++

+

4 ++++

+4

+

+

S+ 4+

+

4

++4+

-4-

4- 4-

4-1cc

4++++++++++ 4~

4444444444++++

4++++++++4-++-

+ +4r4c

+

.+

4 +

++

+

4+

4+

+

44+

+

+

4+

+

+

+

+

+

+

+

+

+

+++

4+

4

+

+

++

+

+

+4

+

+

+4

+

+

+

9T

:I 91 9

+

+

++

+

+

+

+

++

+

+

4

+

4 4

+.+

+

. +

+

+

Ns

N\o

N

N

CM

(V csl

CM

4 -

+4

+

4

+44-

+

4

4 444

++4

4-

+

+

+4

4-

4- 4 i

+4 +

+l

+4

+4

4+

4

\o

+

r- or"+

e

+

+

++

+

+

4+44+

+

4+

+1 +

~ +

+

+

++

++

+

+

44 4

++

+

+

+

+

+•

+

+

4- +

+

+

+

+

+

+

4+

4 ++

+

++

+

4

++

+

+

4-

+

+4

+44+

+

+

+ 4

.

+

++

1 +

+

+

+

+

+ ~II I

-7

4 +

4-

+

+4+

++

+

+4

4 +

+

+

+

4 +

+

+ 4-19

+

+4+

+

4+

+1

+

++

1 I

I ÷

÷

!

+

+4

+

+

+

++

+

+

+14

++

÷

+

++

4 +

+

++

+

4+

+

4+

4 +

+

+

»0

Ht

Ho

Ho

H3

H

H

\ U

H

Hd

r-

N-

H~

0

Hi

Ht

H

NH

1

r _

<

91

,_l

H

<

4i+

1 +

I

+1

+I

+

ON

r-4 vC

14 C

147

I .

0CM

0I I

CM

4

oi.uchicago.edu

ClI

Cl

E4- U0 (SCtCd

>11.)

c,CU

.2F4

C.)

CU

4,

118

zW 00 f I :

1- .14 4

,0 -

4- -

4C993 r-4

1 I Il ý4'V

-'

~ I I I I~ llrl

1 41

*I I I I I I r

4t ('4

Ic * ! I I* I Ir*

i I I

r-41-4

1-4 -4

r-4 1-4

C4 C-41

I I I I

I I ~r ~ 1 lr~~l~l1

1 I~

1-

94- - *

1z

U~I I II

*I I *1 ~ 1~1 1 v-4

C4 1-4

-4 C4 C

4 m

m

e4 I Irl I rl I

*I I I

1- -

1 4 -

Ad

1-4 1-4 -4

1-4~ I I'1 5 ;iI'

' I Irb 1

I I

lc I I

VL4 m I I q m

I4

M

C4 '4

1-4 4 00 r 'T

4 C

N

M (N

-4 q

+ +

) *

rt~

rl I I I~I C

-44c

491 I

M

-.7 I

r -00 'M -4 ~ 00 01 01 M

co0 r4 co 0%

%

D

C, r-4 M cn 0 4 LM

h r-I ýl C

n .4 r, Cq _j o L

% 0% .4 00 eq cn C

m nr Qo

U, 0, rl0

D en1

m ýq

P, wl C

o) .1I

I rl V

m

~ rý f-I Ln 0 ý? CN

cli 1-1 rl

I NI

q (n Ml &

M

M C4 I

m 17 r-4

mvcJI

fr-f

U4

r-0 -.4

-4~ I~r

I l~ ~11

1 I ilrl I I I I I

* r ~

I I I Irda

re rl

I I·-1J I I4 C

14 rl

I I rl

i I I I I I I I

r

r4 (n

rl V

4 14 -1 M -1 -4 -4 M

U

i 4

-4I I

1-4 r I I r4 in -4 r

Tl I I I I 04 m

%0

IC

I I n

I 4 r-4

I In

'I C, 1

-4Iu I I I *

I i~-l I rl

Ic'W

-JI Irc

I en

1-4 eq C

4 1-4 -1 e4 14 1-0lhl~

I I I Inla

~t I~~l

II I 1 I I I~rl~1-4 (n

l ulli -4

Ir

r-4 41 1 1I

I I

1-4 (sa

%D

r-4 -4 cn m

en IT

I I I(V

U (I I

-4 1-4 m

C

4 r-4

r-4 ý4 tI !r tl

Irl I I

I I -4 -4

-4lIrl I Ih

rn l -41 " 11 "1 11 1 0-41

MI I LM

I U

" I

r 1-1 -4

m

m

m

4 n

in -4 N

C11 4

In 4 N

%D

lh

rr-

rl "

LL

n %

0 V

) w N

%D

%

I m %

a 0%

'T

cn .4

11 W

r4 r4 -1 I~I Ir101

rI-I I L rl 1 -41

%"1

I 1 11 f 171

-1 -1 1 '. I I

ic4 1.1Ir

~ IN

. lc *

I I I I I I I I I

ý. I

-1 -1 J J J -1 -1 1

I J-

I J -1- Irfr

0

O0%

4 -00

4-J

Cfs

0 -e

00

0

4-J

u FCd

0 0.

"0 0

s-4) o

0 -0

bD 0 Q

0 C

0o0

0u

00

;-, U,

1-4

0

0 :30

41

bo

.4 -J

4 -J 0

w CU

4-0

00

CU

U

C.)

a0

4

, 4

,o

CSC

U6

C0.

C

CU

BC

Jbf oC

4,4

,~4,,

CU

"0

3-4

"

"0 .~

05

U4

- 0~

SCU

·C

00

3 04

H-

u H

;-

*1'-

oi.uchicago.edu

Urban Origins

of quantitative recording would be a better understandingof sites of the Uruk period. Preliminary work in 1968 and1973, as well as the earlier findings of the Warka survey,had already disclosed that these sites would be numerous.There was every reason to expect, on the basis of previousfindings, that the settlement patterns of the period wouldreflect important shifts over time. Only with a systematic,detailed chronological framework that was generally ap-plicable to surface collections could these shifts in pat-tern be unambiguously perceived. Moreover, the surfaceaspect of Uruk sites is one that invites an extensive listingof traits. In striking contrast to the drab uniformity ofJemdet Nasr and Early Dynastic I sites, on which one ortwo utilitarian categories are endlessly repeated to thevirtual exclusion of all other types, most Uruk sites havea vigorous proliferation of types. In the coding of cate-gories that was finally adopted, therefore, the overwhelm-ing majority of types are only (or at least preponderantly)of Uruk date. The exceptions are the following:

Pre-Uruk: HA (all types of Ubaid painted ware)Jemdet Nasr: AD, HC; also, at least some of the follow-ing:Early Dynastic I: AD, EE, EF, GB, GC, KB, KFEarly Dynastic II/III: (AD), AE, AF, GC, LG

Leaving all but the first of this small group of typesfor discussion in connection with subsequent periods,brief definitions follow for each of the Uruk types towhich a letter code has been assigned. Reference is madeto only a few standard sources of illustration and chrono-logical assessment, since exhaustive treatment of distribu-tion and typology would contribute little to placement ofthe types for dating purposes. Dating spans are provision-ally supplied only in the few cases for which, based onmore or less adequate bodies of stratigraphic evidence,the available literature indicates a reasonable consensusor explicit disagreements.

A BowlsAA Large-diameter bowls with thickened rims, usually

with external collar or flange (Adams and Nissen1972, figs. 49:15, 53:16; Johnson 1973, pl. 2f-h; vonHaller 1932, Taf. 18A: s-z,a'-f'). Early Uruk.

AB Proto-beveled-rim bowls, similar in size, shape, andcrude, straw-tempered irregularity to type AC, thehallmark of the Late Uruk period, but with a signifi-cantly less pronounced bevel or taper (Johnson 1973,pl. Ib). Early Uruk.

AC Beveled-rim bowls, heavily straw-tempered fabric.Widely noted and described (e.g., Nissen 1970, pp.132-38), they need little further reference here. Theonset of the type is variously attributed to a MiddleUruk phase (Hansen 1965, p. 202) and to the LateUruk component of a two-part subdivision of the

period (Adams and Nissen 1972, pp. 97-98). Themajor concentration is in any event coeval withlevels VII through IV in the Eanna precinct at Urukand levels XX through XV in the Inanna Temple atNippur. A perhaps slightly later date for the generalappearance of these bowls at most other sites issuggested later in this appendix.

AG Thumb-impressed bases of crudely made, straw-tempered cups or small bowls (Johnson 1973, pl. If).

B BottlesBA Narrow-necked bottles with tapering-band or

folded-over rims (Hansen 1965, fig. 7; Adams andNissen 1972, figs. 54:1; 59:34; von Haller 1932, Taf.18D:a', 19B:n',p'). Middle-Late Uruk.

BB Same with low outcurling or ledge rims (Adams andNissen 1972, figs. 44:6, 56:6, 61:13).

BC Rims of small, slender, pointed-base bottles withsinuous sides. The rim form most commonly asso-ciated with this shape is a simple, gently outflaringone (von Haller 1932, Taf. 17D:r; Adams and Nissen1972, figs. 53:10, 71:3), but surface examples alsooccur with a more everted rim and one or moresharply edged horizontal ribs. Early Uruk.

BD Bases of the above. On many sites rims and baseswere counted together, and the two types are groupedtogether in most of the analysis that follows.

C JarsCA Neckless, "hole-mouthed" jars with thickened rims

(Adams and Nissen 1972, figs. 52:8, 53:8, 61:11, 12;Johnson 1973, p. 31).

CB Low, out-turned thickened jar rims (Adams andNissen 1972, fig. 43:33, 34, 58:30; Johnson 1973, pl.3b).

CC Thin outflaring or outcurling jar rims (Adams andNissen 1972, figs. 36:36-38, 43:38, 49:7). Early Uruk.

CD Small globular jars with short vertical or slightlyflaring rims. This designation is applied to rim sherdswithout evidence of handle attachments, but in factmost such jars seem to have been equipped withone or two handles. See particularly handle type EA,also EB. Late Uruk.

CE Tapering-band or folded-over jar rims (Adams andNissen 1972, fig. 39:15).

CF Minature jars of various shapes, none commonenough to justify a separate typological category(Adams and Nissen 1972, figs. 39:12, 42:27, 47:4,58:23, 62:20, 68:11).

CG "Simple," vertical to slightly flaring jar rims. Prov-ing to be somewhat of a catchall category in whichotherwise quite dissimilar vessel forms were arbi-trarily grouped, this was dropped before the conclu-sion of the survey.

119

oi.uchicago.edu

Urban Origins

D BasinsDA Large, flat-bottomed basins or trays, generally with

vertical or slightly flaring sides and occasionally withthickened and/or everted rims. Examples of thiswere numerous in the region around Uruk, wherethey occurred fairly regularly with deep interior scor-ing on Late Uruk sites (Adams and Nissen 1972, p.100), and they are also reported at Nippur (Hansen1965, p. 202). However, they proved less useful as aperiod diagnostic elsewhere. Also common in theUruk area at about the same period were smaller,flat-bottomed plates with vertical sides, often withcrosshatched rims (Adams and Nissen 1972, figs.37:13, 54:6-7, 56:7, 61:8; von Haller 1932, Taf.18D:a-c). These were so rare in the Adab-Isin-Nippur area that they were not recorded as a sep-arate type.

E Handles, LugsEA Strap handles, normally applied singly to globular

jars or cups of small to medium size. Rims are low,vertical, or slightly flaring (see type CD), and fre-quently there are a series of shallow concentricgrooves (perhaps the impressions of the teeth of acomb) on the shoulder. A very common Middle andLate Uruk type, probably continuing into JemdetNasr (Hansen 1965, p. 207, fig. 6; Adams and Nissen1972, p. 100, figs. 37:11, 41:16, 21, 44:163b/1, 3,50:5, 9, 58:26, 27, 68:9, 10; Johnson 1973, pl. 7c, d).

EB Twisted rope handles, applied singly to approxi-mately the same range of vessel forms as type EA(von Haller 1932, Taf. 18C:p; Hansen 1965, pp. 202,207; Adams and Nissen 1972, p. 100, figs. 37:9,41:18, 20, 45:4, 6, 50:8, 58:25, 64:14; Johnson 1973,pls. 7e, 8a). Also Middle and Late Uruk, lastingslightly longer.

EC Horizontal twisted rope handles or lugs, generallyapplied in pairs to a variety of jar rims including(but not limited to) those commonly having type EAor EB handles (Adams and Nissen 1972, p. 100, figs.41:22, 44:163b/4, 45:1, 47:2, 58:28-30; Johnson1973, pl. 8b). Middle or Late Uruk.

ED Nose lugs, usually pierced horizontally, applied tothe shoulders of a variety of jar shapes. To judgefrom relatively complete specimens, either two orfour were normally attached to each vessel, possiblyfor suspension (von Haller 1932, Taf. 19D:b; Hansen1965, figs. 11, 12, 14, 21; Adams and Nissen 1972,figs. 37:4, 38:2, 42:27-28, 47:5, 7, 51:12, 58:23). Along-lived embellishment of limited utility for chron-ological placement of surface materials, it apparentlyspans most or all of the Uruk period and continueswell into Early Dynastic times.

F SpoutsFA Split spouts, placed at rims of bowls and open jars

for pouring (Hansen 1965, fig. 3; Adams and Nissen1972, p. 100; figs. 33:4, 7, 36:30). Early and perhapsMiddle Uruk.

FB Tall, straight spouts, freestanding and generallyplaced at an angle on the shoulders of jars (von Hal-ler 1932, Tafs. 18A:f, 18C:b', 18D:i; Adams and Nis-sen 1972, p. 100, figs. 33:5, 36:31, 52:4). Early Uruk.

FC Short spouts directly attached to low vertical oroutflaring jar rims. Apparently two types of some-what different chronological significance are involvedin this single characterization (Adams and Nissen1972, cf. fig. 30:c, n, and fig. 31), but it is generallydifficult to distinguish between them on the basisof fragmentary surface material. One is of unequivo-cally Early to Middle Uruk date (Adams and Nissen1972, fig. 49:7, 72:20; von Haller 1932, Taf. 18D:m);the other, equally clearly, is Late Uruk (Hansen 1965,p. 205, fig. 20; Adams and Nissen 1972, fig. 59:33,36-39). The two could not be consistently separatedin this study, but almost all the examples found ap-peared to belong to the earlier subtype.

FD Short conical spouts, wide in diameter at their junc-tion with jar shoulders but tapering rapidly to asmall orifice (von Haller 1932, Taf. 19B:a; Adamsand Nissen 1972, figs. 41:15, 45:1). Apparently aLate Uruk form, this continues at least throughJemdet Nasr times.

FE Drooping spouts, a common and characteristic lateMiddle-Late Uruk form (Hansen 1965, p. 204, figs.17-18; Adams and Nissen 1972, p. 100, fig. 44:6,54:1, 59:34; Johnson 1973, pl. 6a).

FF False spouts, lacking an opening through the vesselwall. Roughly contemporary with type FE in theUruk region and at Nippur (Hansen 1965, p. 204;Adams and Nissen 1972, p. 100), not a single ex-ample was found during subsequent survey.

FG Unclassifiable fragments or variants.

G Vessel Bases Other Than Those Specified AboveGA Ring bases, various profiles and sizes. This category

was recorded in a frankly speculative attempt to seewhether it could be linked with a particular temporalhorizon. From surface evidence (see table 9), an Earlyor perhaps Early-Middle Uruk dating seems mostlikely.

H Painted WareHB Uruk painted styles. Most commonly applied to

spouted jars, patterns consist of horizontal bandssometimes connected by crosshatching and separatedby one or more horizontal bands of triangles orlozenges that are frequently filled in by stippling or

120

oi.uchicago.edu

Urban Origins

crosshatching. (Adams and Nissen 1972, p. 100, figs.52:1-3, 6, 71:11, 72:16, 80:43-44). Small fragmentsare of course difficult or impossible to distinguishunambiguously from earlier or later styles.

I Appliqued Strips or RidgesIA Horizontal, fingertip-impressed strips, usually one or

more placed below large bowl rims or on jarshoulders (von Haller 1932, Taf. 18A:y-z, b'-f';Adams and Nissen 1972, figs. 33:1, 42:31-32, 49:15,53:8, 16, 57:15, 16, 21, 72:13, 14; Johnson 1973, pl.2a, d). Early Uruk.

IB Roughly contemporary with the preceding, but thesehorizontal are thinner, shallower, and accented withfingernail or shell impressions (Adams and Nissen1972, figs. 42:30, 43:37,44:2; Johnson 1973, pl. 3m).While contrasting at the extremes, types IA and IBoften intergrade with one another and hence couldnot always be recorded separately.

IC Sinous appliqued strips, generally thin and of lowrelief. Some are fingertip- or fingernail-impressed, asin types IA and IB, but more often these strips areunembellished. Fragmentary surface findings do notpermit the larger designs presumably traced out bythese curvilinear strips to be discerned in most cases(Adams and Nissen 1972, figs. 33:1, 49:18, 72:17).

] Incised DecorationJA Crosshatching within triangular borders, generally

framed in horizontal bands around jar shoulders(Hansen 1965, p. 205, fig. 21; Adams and Nissen1972, figs. 47:7-8, 54:11). Late Uruk.

JB Other crosshatching, mostly also in horizontal bands(von Haller 1932, Tafs. 19D:b, 20A:n'; Adams andNissen 1972, figs. 49:12, 62:21, 72:14; Johnson 1973,pl. 4b, c). Late Uruk.

JC Meander pattern, usually consisting of single broad,shallow groove.

JD A form of decoration applied to globular and otherjar shoulders, predominantly of types EA and EB,consisting of a band of parallel grooves or incisionsthat are cut obliquely by parallel, somewhat morepronounced slashes (Adams and Nissen 1972, fig.37:9, 41:20, 44:163a3, 50:10, 62:17, 64:14, 68:9).Late Uruk.

JE Irregular interior scoring on large shallow bowls orbasins.

K Other Surface DecorationKA Reserved slip, a technique in which "the body (or

parts of the body) of a vessel is covered with a thinslip which is then partially wiped off so that the bodyclay shows in a pattern against the darker color ofthe slip" (Perkins 1949, p. 109). At one time thought

to be characteristic of the early part of the EarlyDynastic period, excavations in the Diyala regiontraced its origins there to at least Jemdet Nasr times.In southern Mesopotamia, however, the techniquewas in substantial use through much of the Uruk aswell as Jemdet Nasr period and seems to have beenless extensively employed thereafter (Hansen 1965,pp. 202-3, 207; Adams and Nissen 1972, p. 100,figs. 37:1, 45:1, 49:7, 52:12, 53:7, 59:36, 68:1, 70:2,77:10, 11). Oblique patterns are the most common,but vertical banding also occurs.

KC Punctate impressions, perhaps reed-tips, applied tojar exteriors (other than the specific pattern of usementioned for the Early Dynastic period under typeKB).

KD Roughly combed or raked finish on jar exteriors(Adams and Nissen 1972, figs. 37:5, 50:1, 52:13;Johnson 1973, pl. 4f).

KE Chevron-rocker impressions, perhaps obtained byadvancing the edge of a shell alternately at one endand then the other (Hansen 1965, figs. 1, 2; Adamsand Nissen 1972, fig. 41:16). At Eridu these werefelt to be assignable to early Uruk (Perkins 1949, p.103), but elsewhere they apparently continue intoMiddle and even Late Uruk times (Hansen 1965, p.201).

L Other Ceramic ArtifactsLA Clay sickles. Originally thought to be a hallmark of

the Ubaid period (Lloyd and Safar 1943, p. 155;Adams 1965, p. 127), they were regarded during theWarka survey as an undifferentiated indicator of theUbaid, Uruk, and Jemdet Nasr periods (Adams andNissen 1972, pp. 208-9). While some apparently dooccur on single-period Jemdet Nasr sites (e.g., 130,256), surface collections made during the course ofthis study argue that the Early and Middle Uruksubperiods (as well as perhaps the Ubaid period)were the time of their most intensive production, andthat by Late Uruk times their use was apparentlydiminishing rapidly.

LB Celts or axes, usually found with badly batterededges and hence apparently serving a utilitarian endin spite of the relatively soft material. Perhaps theywould serve for cutting down shok, camel thorn,and other leguminous shrubs for fuel?

LC Cones, frequently with pigmented and/or indentedheads, for use in wall mosaics of public buildings(Adams and Nissen 1972, p. 211).

LD Hammers, with splayed ends and a generally ratherslender shaft hole. Like celts, these are usually foundin battered condition and in small fragments. A utili-tarian purpose seems likely but remains obscure(Adams and Nissen 1972, p. 213).

121

oi.uchicago.edu

Urban Origins

LE Net weights of various shapes (Adams and Nissen1972, p. 213).

LF Spindle whorls (Adams and Nissen 1972, p. 212).LH Kiln debris. A record was kept of the distribution of

pottery kiln wasters in order to trace the extent ofproducing activity in various periods. Where kilnwasters can be identified as fused clumps of sicklesor badly deformed vessels of one of the types classi-fied above, this is indicated in table 9. Not in Warkasurvey, nor generally noted except in intensive col-lections.

M Chipped Stone ToolsMA Bifacially worked flint or chert axes, celts, or hoes,

generally in the shape of an elongated trapezoid.MB Small (although characteristically not microlithic)

flint blades and rectangular blade sections.MC Small, rectangular flint blade sections with denticu-

lar retouching along one or both edges.MD Prepared flint blade cores, always having been used

until blades of adequate size could no longer bestruck from them.

ME Small obsidian blades or rectangular blade sections(rarely noted in Warka survey).

N Ground Stone UtensilsNA Stone vessels, more often shallow rectangular plates

than deeper bowls. The skilled craftsmanship andcomplex forms that were common around Urukduring the Late Uruk period, even on smaller sites(Adams and Nissen 1972, pp. 206-8), were never inevidence in the region farther north.

NB Mace-heads (Adams and Nissen 1972, p. 211).NC Large crude, pierced, doughnut-shaped "digging

stick" weights.ND Mortar or metate fragments.NE Small cubes with rounded edges, 5 centimeters or

less on a side.NF Hourglass-shaped weights, crudely shaped of porous

white stone.NG Axes/celts.

O Copper or Bronze UtensilsOA "Tool" fragments; rods of circular or rectangular

cross section, sometimes tapering.OB "Vessel" fragments; sheet or plate, generally curved.OC Small, unclassifiable metal objects.

The foregoing listing is in some respects a concatena-tion of dissimilar units of observation, not all of themnecessarily of any chronological significance. A numberof whole vessel types are included, distinctive not onlyin basic form and size, but often in associated appendagesand modes of surface treatment. Not a few of the latteralso have been listed as independent types, since they are

not confined to one vessel category but can occur onseveral. Particularly the stone and metal classificationscontain tool types that tend to be more functionally thanstylistically defined. The adoption and spread of thesetypes conceivably may be more cumulative, an index tobroad, long-range patterns of technological or economicachievement and activity, in comparison with the pre-sumably more abrupt cycles of popularity of stylisticallyvariable but functionally equivalent pots. And finallythere are residual categories, lumping undefinable frag-ments and minor stylistic variants to provide no morethan a slight extension of our knowledge of the distribu-tion of, for example, jar spouts or metallurgical products.Yet also contained in this list are a number of reasonablyconsistent and reliable indicators with which to assignindividual sites to one or more of a sequence of Uruksubperiods.

The traits in question are those used as ordering criteriain table 10, drawn from quantitatively recorded Urukcollections that are shown in table 9. In most cases theirchronological sensitivity has been previously established,although the declining frequency of clay sickles (typeLA)-on some Early Uruk sites more common than allother recognizable types added together-does not appearto have been noted previously. Typologically, the transi-tion from Early Uruk to Late Uruk times can be viewedin this region as involving the decline or disappearance oftwo vessel categories in addition to clay sickles: large di-ameter, thick-walled, thickened or flange-rim bowls (typeAA), frequently with appliqued, notched or impressedstrips (types IA and IB); and small-diameter, pointed-basebottles (types BC and BD). It may also be viewed as in-volving the introduction and substantial rise in popu-larity of more globular bottles with different rim profiles(types BA and BB), small globular jars with strap ortwisted-rope handles (types CD, EA, and EB), and thecrudely made beveled-rim bowls that are sometimes re-garded as the most ubiquitous diagnostic criterion forthe Uruk period as a whole. Simultaneously, there is atendency for certain types of bowl and jar spouts (typesFA and FB) to be more or less completely replaced byothers (types FD and FE). Using these diagnostics, astable 10 indicates, an unambiguous assignment of a ma-jority of the intensive collections to either Early or LateUruk subperiods can be readily accomplished.

However, this procedure leaves a substantial residualcategory of sites that cannot be so easily or unambigu-ously dated. Included in it must be some sites occupied-or periodically reoccupied-during varying proportionsof both the Early and the Late subperiods. In other casesdifficulties of assignment may arise from unnoticed fea-tures of ancient intra- or intersite specialization. But instill others it is reasonable to expect that the collectionsreflect "single period" sites that were occupied during atransitional or full-fledged Middle Uruk subperiod.

122

oi.uchicago.edu

Urban Origins

TABLE 10 Provisional Grouping of Intensive Collections at Uruk Period Sites by Chronological Subperiod, Based on CeramicIndicators Most Clearly Reflecting Change in Frequency.

SITES GROUPED BYSUB-PERIOD, INMTTMV.TrAT Tn'DTh

118792804837

1118s 1124

S115911981416

SITEAREA/( A )IIC"

5.36.81.71.40.66.84.30.6

rnnM AT.. .p TU TAL :PERCENT:

E 128 1.8245 6.0573 1.0790 5.5

S9404-#1 1.7S 975 5.8S1020 8.2

ýE 1072 3.4* 1137 3.8

S 1164 1.01172 25.51194 11. 5

Z 1205 7.9H 1312 0.2o 1375 0.2S1432 1.7

H TOTALS:PERCENT:

125 24.0126 4.0574 1.0805 0.5

E-4940-#2 1.71154 2.61163 0.9

S1165 5.31216 4.81261 0.8

S1315 2.8S1448 0.9

TOTALS:PERCENT:

TOTALURUKSHERDmTnO

EARLIER URUK TYPESBC+ FA+ IA+

AA Rnh TI T A

93101124118120147519075

919

,1441358694

11261

247279136

1916533

192783867

7268736133397870458349

1)6807

LATER URUK TYPESBA+ EB+

ArC BB Cn A C 'FD+PF

8 * 4 7311 6 2 15 4315 4 14 7816 5 2 7 78

3 2 2 7211 7 * 9 101

6 * 1 5 2710 1 11 5121 1 18 20

101 26 5 85 54311.0 2.8 0.5 9.2 59.1

2 1 752 1 3 11 342 3 4 38

10 2 1 203 * 1 2 593 * 2 39

20 11 2 12 1146 * 11 848 5 1 4 671 * 11

19 * * 5 1006 1 1 2

18 2 * 4 1207 1 6 452 2 1 152 1 1 * 27

111 22 17 63 8505.9 1.2 0.9 3.3 45.1

1 1 1 ** 1 143 * 2 101 1 11

61 * 1

2* *

3 51 * 2 381 4 1 9

9 398 4 2 19 135

0.9 0.5 0.2 2.4 16.7

* Rare sherds of this type noted elsewhere than in area(s) of intensive collection.

Working with surface collections of the Uruk periodfrom southwestern Iran, Johnson has developed a strongif still not wholly convincing case for the existence of anidentifiable Middle Uruk subperiod. What makes hisviews especially pertinent is that they are based on ma-terial that not only is typologically very similar but wasmore extensively and systematically recorded than waspossible here. He summarizes the grounds for his judg-ment while candidly acknowledging ambiguities that havenot yet been eliminated, in terms possibly applicable tosouth central Iraq as well to Khuzestan:

the absence of clear phase diagnostics for Middle Uruk

renders identification of Middle Uruk occupations fromsurface collections a problematic issue. Our experienceindicates, however, that Late and Early assemblages con-tain a sufficient number of phase diagnostics that theirabsence from an Uruk surface collection of reasonablesize is a reliable indicator of the presence of a Middle Urukoccupation. Middle Uruk contains a sufficient number oftypes not present in Early to allow identification of anEarly-Middle occupation (in the absence of Late Urukdiagnostics). A more serious problem is posed by theclose similarity of much of the Middle and Late Uruk as-semblages. Thus it may be difficult to distinguish a sitewith Middle and Late occupations from a site with a Lateoccupation only. [Johnson 1976, p. 204]

123

* *

1* 1 *

1

0.2 0.1

11 * 1 410 * 4 1 1 ** * 14 2 6 1 * *1 * * *

* 2 *1 12 * * *

114 4 5 114 2 * 1

1 111 1 *4 14 2 *1 1 4 14 17 1 2 2 *

187 22 12 15 12 49.9 1.2 0.6 0.8 0.6 0.2

9 3 8 3 15 8 6 2

16 2 4 *23 1 4 3 6 314 * 1 233 *52 2 3 636 1 6 3 410 2 2 2 117 1 5 3 111 1 2 3 *

6 1 2 1 13 10232 14 33 24 42 1928.7 1.7 4.1 3.0 5.2 2.4

oi.uchicago.edu

Urban Origins

It remains to be seen how useful it is to define a tem-poral unit on this fairly tenuous basis-although in somerespects the basis for the widely accepted Jemdet Nasrperiod is not very different. Insofar as we deal with surfacecollections rather than stratigraphic excavations, we mustremain conscious of unstated assumptions about occupa-tions falling conveniently into "Early-Middle-Late"blocks rather than generally being confused by repeatedreoccupations and abandonments. But a Middle Urukcategory is satisfying to good trinitarian principles andcertainly is not an unreasonable chronological refinementtoward which to work. While the field data available fromthis study in general do not permit its recognition evenwith the qualifications that Johnson describes, I have em-ployed the term provisionally as an admittedly mixed andsomewhat amorphous grouping of sites for which moredefinitive criteria cannot yet be assigned.

Concern for chronological subdivisions is obviously im-portant if we are to understand the dynamics of settle-ment change within the Uruk period. Even a partial andprovisional assignment of individual sites to one or moreof these categories contributes to the identification of se-quent patterns of distribution and shifts in size anddensity. But the variation in individual collections thatis suppressed by any system of categories is itself signifi-cant for cultural understandings of a different order.Within the framework of a discussion focusing principallyupon chronology, let us consider a supplemental methodof ordering the collections that highlights variation andseeks to test its relationship to successive chronologicalintervals.

The Uruk period was a time when certain artifacts en-tered into genuine mass production, even if the producingunits in many cases seem to have been small handfuls ofkilns in numerous individual settlements of all sizes ratherthan large, centralized workshops. The unprecedentedconcentration on a few seemingly utilitarian types givesevery appearance of being a sociological as well as astylistic fact, a decisive shift in behavior patterns that in-dicates an alteration in the traditional framework of socialrelationships governing production, distribution, andconsumption. Although we are only dimly aware of whatthe shift may have entailed, the development of mass pro-duction deserves to be treated on a different plane from amere succession of stylistic shifts. It is at least arguablethat the onset of this new pattern was a cumulativeprocess that ramified widely into the social organizationof towns and villages of the time and was in this sensequite distinct from the waxing and waning of traditionalstyles through atomistic individual decision-making. Atany rate, this interpretation suggests a rank ordering of thecollections on the basis of certain frequencies of occur-rence as a means of assessing the uniformity of incidenceand spread of some of the most common, widely circu-lated items.

Clay sickles and beveled-rim bowls are the two mostuseful cases, at least within the framework of types em-ployed in this study. As already noted, the formerreached an apparent peak of popularity in Early Uruktimes and thereafter declined, whereas the latter were es-sentially absent in Early Uruk times but subsequently be-came very popular. Table 11 rank orders collections onthe basis of the frequency of these traits, providing in addi-tion certain other chronological information. How doesit complement or modify the chronological ordering givenearlier?

Several pertinent observations may be made with re-gard to the distribution of clay sickles. First, it seems clearthat a substantial, continued decline in the frequency ofsickles began during or immediately after the early phaseof the Uruk period. At eight of the nine sites attributableprimarily to that phase, sickles constituted from 42.6 to78.5 percent of all identifiable Uruk sherds, and the lowerfrequency at the ninth (1416) is surely due to special limi-tations on collecting procedures that were necessitatedthere by an overburden of post-Uruk debris. Uruk collec-tions at the twelve sites of primarily Late Uruk date, onthe other hand, ranged in sickle frequency from 0 to 45.8percent, with only one of them rising above 29 percent. Itmust be added, however, that actual production of sickles-as distinguished from mere use and discard-seems tohave continued until at least the end of the Uruk periodif not later (cf. Adams and Nissen 1972, p. 208). This isconfirmed by the presence of sickle kiln wasters at a sitelike 1163 where there is essentially no evidence of anEarly or even Middle Uruk occupation.

It can also be observed that sickle production and dis-card is not linked in any close or persuasive way with sitesize. One might have hypothesized, for example, thatsickle production was concentrated in the larger settle-ments and that the smaller ones were merely subsistence-oriented villages. However, this seems not to be the case.Kiln wasters of sickles were found at only one of the fourEarly Uruk sites larger than 4 hectares, but at three of thefive smaller ones. Similarly, there is nothing to suggest thatthe frequency of ordinary sickle fragments was inverselycorrelated with Early Uruk site size.

Overshadowing these particulars of sickle distributionis a remarkably common feature: their overwhelming fre-quency. They constitute more than three-fifths of allidentifiable Uruk sherds from all the Early Uruk sites con-sidered as a group, and more than 45 percent even at sitesassigned to the Middle Uruk phase. To be sure, manybody sherds from a broken pot may not be assignable toone of the above categories, whereas virtually all sicklefragments are immediately recognizable as such. But theprofusion is still noteworthy and should be kept in mind,even though no obvious explanation is available. Grant-ing their fragility in the face of hard use, what accountsfor the presence of such vast numbers of these implements

124

oi.uchicago.edu

Urban Origins

TABLE 11 Rank Orderings of Uruk Sites in Terms of Typological-Chronological Criteria Listed in Table 10

DECREASING RATIOOF TYPES PRESENT:EARLY/LATE URUK

SITENO.

83711880411591416111811987921194137511241172120511541164573975

1312940-#1113710201432

574107224512611315126128125790805116512161448940-#21163

CHRON.CLASS. RATIO

EarlyEarlyEarlyEarlyEarlyEarlyEarlyEarlyMiddleMiddleEarlyMiddleMiddleLateMiddleMiddleMiddle

MiddleMiddleMiddleMiddleMiddleLateMiddleMiddleLateLateLateMiddleLateMiddleLateLateLateLateLateLate

5/04/04/04/04/04/14/15/24/24/25/35/35/33/23/24/34/34/35/45/45/55/54/44/45/64/54/53/4

4/64/63/62/52/52/61/41/4

DECREASING EARLYPROPORTION OF TOTALURUK SHERDS

SITE CHRON.NO. CLASS.

837 Early118 Early804 Early1124 Early1198 Early1416 Early1159 Early792 Early1312 Middle1172 Middle1205 Middle975 Middle1118 Early1020 Middle1164 Middle1137 Middle940-#lMiddle

573 Middle128 Middle1375 Middle1261 Late1432 Middle245 Middle1072 Middle1448 Late790 Middle1315 Late1194 Middle126 Late805 Late574 Late940-#2 Late1216 Late1154 Late125 Late1163 Late1165 Late

91.591.489.587.181.180.076.576.275.675.275.072.165.864.463.262.558.054.754.252.649.446.337.836.235.335.130.630.322.121.320.618.217.85.14.22.60*

INCREASING LATEPROPORTION OF TOTALURUK SHERDS

SITE CHRON.NO. CLASS. %

1188048371159141679211241118940-#1119857312059751020117213121164128245113713757901194143214485741261261125131512161072940-#2805116511631154

EarlyEarlyEarlyEarlyEarlyEarlyEarlyEarlyMiddleEarlyMiddleMiddleMiddleMiddleMiddleMiddleMiddle

MiddleMiddleMiddleMiddleMiddleMiddleMiddleLateLateLateLateLateLateLateLateLateLateLateLateLate

000000*

0.70.80.91.11.23.13.35.37.39.010.6

11.111.912.513.213.815.217.924.330.130.932.533.334.737.844.451.563.971.480.884.6

DECREASING PROPORTION OFCLAY SICKLES (TYPE LA)

SITENO.

:110

11248379758041205

'117211181164131211981159940-#1128113710201261

573792143213751072144814162457901261315940-#280557412161194116311541165125

CHRON.CLASS.

EarlyEarlyEarlyMiddleEarlyMiddleMiddleEarlyMiddleMiddleEarlyEarlyMiddleMiddleMiddleMiddleLate

MiddleEarlyMiddleMiddleMiddleLateEarlyMiddleMiddleLateLateLateLateLateLateMiddleLateLateLateLate

LOCAL MFG(KILN

S WASTERS)

78.568.766.163.962.962.560.660.057.957.756.752.952.752.149.346.245.8

44.242.640.339.530.128.726.725.221.320.618.418.218.013.711.16.12.62.60*0*

x

xxxxxxxx

x

xx

x

x

x

xx

x

INCREASING PROPORTION OFBEV.-RIM BOWLS (,TYPE AC)

SITE CHRON.NO. CLASS.

118 Early804 Early837 Early1118 Early1159 Early1198 Early1416 Early573 Middle792 Early975 Middle1124 Early1020 Middle940-#1Middle1312 Middle1205 Middle790 Middle1448 Middle

1164 Middle1172 Middle126 Late245 Middle128 Middle1137 Middle1432 Middle1375 Middle1194 Middle125 Late1261 Late574 Late1216 Late1315 Late805 Late1072 Middle940-#2 Late1165 Late1163 Late1154 Late

LOCAL MFG(KILN

S WAPSTERS)

00000000*0*0*0*

0.40.91.32.14.34.4

5.36.77.47.47.610.310.410.612.112.520.521.922.222.437.740.942.451.466.784.6

x

x

xxx

* Rare sherds of this type or types noted outside area(s) of intensive collection.

in occupational debris rather than in the surroundingfields for which they were presumably destined? It appearsthat they must have been supplementarily employedwithin the settlements. The conversion of low desert brushto some more condensed form of fuel is a modern house-hold task for which they would have been at least margin-ally suitable.

Turning to beveled-rim bowls, tables 10 and 11 estab-lish that their rise in frequency more or less paralleled thedecline of the sickles. Examples are extremely rare on thegroup of Early Uruk sites and hence perhaps suspect aslater strays. The Middle Uruk group-emphasizing oncemore that this may not be a chronologically well-definedphase-ranges from equally rare occurrences up to fre-quencies of 10 to 12 percent (the much higher frequencyat site 1072 is due to one of the three intensive collectionsthere having been made in an area of exceptional, highlylocalized concentration around an apparent kiln). LateUruk frequencies of beveled-rim bowls can be almost as

overwhelming as was the case with sickles during theEarly Uruk period, although the median is only 22.3 per-cent and the variance is greater. The significance of thegreater variance is that beveled-rim bowl frequenciesprobably reflect more clustered patterns of manufacture,use or discard than those of sickles. Hence the generalchronological trend toward increased use may often bemasked by localized variables that are not yet well under-stood." Suggestions that beveled-rim bowls were stan-dardized measures functioning in connection with a rationsystem, most systematically advanced by Johnson (1973,pp. 129-39), find little support in the distributions re-corded in tables 10 and 11. Rationing carries a presump-tion of specialization and of a differential concentrationon larger sites where agricultural surpluses were morelikely to be stored and administrative specialists weremore likely to congregate. Yet in these data beveled-rim-bowl frequencies are if anything negatively correlatedwith site size. Only two of the eleven Middle and Late

125

----- -- r- ~~ --- ---'-I i 11 * %"4-& / ·U -IT. . . . . .f . . . . . . . . . II "i"1l0|--l I -I r)v

I-----------

I

I

oi.uchicago.edu

Urban Origins

Uruk sites of more than 4 hectares have Uruk ceramic in-dustries in which beveled-rim-bowl frequencies are higherthan 12.5 percent, while eight of the seventeen remaining,smaller sites have frequencies of 20.5 percent or higher.Manufacture, to be sure, may be marginally more concen-trated. Kiln wasters of beveled-rim bowls were found onthree of the larger sites but only two of the smaller ones.That was clearly the case in Uruk, where Nissen reports60 to 79 percent frequencies in occupation levels (1970,pp. 129-31).

It must be kept in mind, of course, that the entire fore-going discussion rests upon distributions of surface mate-rials rather than on excavated, stratigraphically securesamples. The possibility of subsequent disturbance cannever be entirely excluded. It is only a matter of conjec-ture, moreover, what the depth and character of overlyingdebris may have been that provided a matrix for the pres-ent assemblages of surface artifacts before it was scouredaway by various erosive processes. Still, the results of thisanalysis are closely congruent with what is known fromreasonably secure and well-documented Uruk strati-graphic sequences. If future excavations or more refinedsurveys indicate the need for corrections in this frame-work, we can at least hope that the data in table 9 willprovide a basis for refining the site chronology as it issummarized in table 7.

There may be need also to reiterate a second caveat orqualification. As I suggested earlier, a variety of sequencesof occupation at individual sites could account for similartype frequencies among the surface materials. Differencesin duration, periodic changes in size or concentration, andeven intervals of temporary abandonment all may lie be-hind sites whose surface aspect is very uniform. Individualsettlement histories almost certainly were complex andnot at all uniform, and the presentation of rank orderingslike those given in table 11 should not be taken to implythat the demographic history of the region as a whole canbe conceptualized as an orderly sequence of "single-period" occupations arranged along a continuum fromEarly to Late. To test the extent of variability in occupa-tional sequences, as well as to advance our understandingof the Uruk period in almost any other way, we mustawait an opportunity to examine considerable numbersof both smaller and larger Uruk sites with programs ofstratigraphic testing and perhaps large-scale excavationas well.

Finally, the Uruk chronology that emerges from thisanalysis should be regarded not merely as a somewhatprovisional sequence but as a strictly relative one. Thetripartite divisions within it are not clearly established,although the considerable contrasts between the twohalves of an alternative bipartite division make this furtherrefinement plausible. Additionally, the numbers of sitesfalling into the various divisions of the sequence implyvery little about the duration of those divisions. Depths of

occupational debris observed in excavations can also be amisleading guide to duration of settlement, but in mostcases they are a far better index to chronology than whatcan be offered here. To take the giant next step, onceagain, we must await a substantial program of excava-tions directed toward the long list of promising candi-dates identified in the site catalog. But during the un-known and surely considerable interval until that ispossible, this discussion has proceeded on the convictionthat it is better to light a small candle than to curse thedarkness.

Jemdet Nasr Period

As was pointed out at the time of the Warka survey(Adams and Nissen 1972, pp. 100-103), diagnostic cri-teria for the Jemdet Nasr period are extremely difficult tocharacterize without ambiguity. Perhaps because of itsapparent brevity, surface collections from sites occupiedduring the period have very few distinguishing features.

Conical cups (type AD), produced crudely but in greatprofusion, are in every respect the dominant feature.There may be little difficulty with their origins, sincethey appear to replace beveled-rim bowls more or less ab-ruptly in a shift that is equated here with the Uruk-JemdetNasr transition. Similar, continuously intergrading formsseemingly continue for the greater part of the succeedingEarly Dynastic period, however, so that a site abandonedafter Jemdet Nasr times often can be distinguished fromone that continued only by the presence of conical cupsand the simultaneous absence of characteristic Early Dy-nastic I forms like solid-footed goblets (type GB).

Perhaps the only specific features are certain styles ofboldly linear painted decoration in red, black, or purple(HC) that are typically applied to the upper surfaces oflarge jars. But these, unfortunately, do not occur fre-quently enough for their absence at any particular site toprovide a conclusive indication that it was not occupiedduring the Jemdet Nasr period. Hence a large proportionof sites that are cataloged in table 7 as probably havingbeen occupied during the Jemdet Nasr period can beshown only parenthetically, implying a considerable ele-ment of doubt about their existence, not to speak of theirextent, at this period.

Perusal of finds cataloged in table 7, both from single-period Jemdet Nasr settlements and from others in whichEarly Dynastic I remains also were present but could bedistinguished because their distribution was obviouslylocalized, permits a number of additions to be made to theabove-mentioned two types. While they also occur onLate Uruk sites as indicated earlier, the following traitsapparently continue into or even through the Jemdet Nasrperiod: bottles (with folded-over or ledge rims (BA, BB),which are attested also in Jemdet Nasr levels at Nippur(Hansen 1965, p. 207); miniature jars (CF); strap andtwisted-rope handles for globular jars (EA, EB), again

126

oi.uchicago.edu

Urban Origins

lasting into Jemdet Nasr levels at Nippur, nose lugs on jarshoulders (ED), conical spouts (FD), reserved-slip ware(KA), clay sickles (LA) in severely declining numbers,clay cones for mosaic wall decorations (LC), and stonebowls (NA).

Early Dynastic I Period

While it remains "orthodox" to consider the Early Dy-nastic period a cohesive phase of archaeological develop-ment, the attainment by Uruk of its maximal urban sizein the Early Dynastic I period has led in this chapter to ascheme of periodization concluding only after that time.However, there are at least some additional considera-tions of a more traditional kind that lend support to thealternative division followed here. As Crawford has notedon the basis of architectural as well as ceramic criteria, "itis becoming increasingly clear that there is a strong casefor a continuous development from the Jemdet Nasrphase into Early Dynastic I." Moreover, Early DynasticII, at one time "merely an awkward transition betweenEarly Dynastic I and Early Dynastic III," has increasinglytaken on developmental as well as archaeological char-acteristics of its own (Crawford 1977, p. 9).

Diagnostic criteria used throughout this study remainessentially as they were in the Warka survey (Adams andNissen 1972, p. 103). Conical cups (AD) continued inheavy use, but they were supplemented by the mass pro-duction of taller conical goblets or chalices with a low,solid foot that took the form of a truncated cone. As slop-pily made as the conical cups, these goblets are generallysomewhat irregular in shape and often have slumped dur-ing the firing process. Hence it is unlikely that the feet, inany case of very modest size in proportion to height,could have been intended as supports that would normallypermit the goblets to stand on a flat surface. At least atNippur this type seems to have been of very brief dura-tion in spite of its popularity, for it is noticeably more fre-quent in the middle than in the earlier or late levels asso-ciated with the period (Hansen 1965, p. 209).

Other dating criteria, much less frequently found thanthe former two, include: a vertical lug handle placed on

jar shoulders, taking the form of a triangular slab (EE),apparently the forerunner of the later, so-called goddesshandles (Adams and Nissen 1972, p. 103, fig. 74: 10); tabsor short ledge handles placed at intervals (normally fourto a jar?) around low vertical or slightly flaring jar rims(EF), "characteristic of Early Dynastic I" at Nippur (Han-sen 1965, p. 208); curved upper surfaces of jars withreserved-slip decoration (usually in a continuous obliquepattern) and a line of punctate impressions at the junc-tion between the shoulder and the neck of the jar (KB);and "cut ware" (KF), tall vessels, stands, or pedestalswith groups of triangular or circular excisions and oftenalso with incised crosshatching, notched horizontalridges, and other forms of surface decoration. So-calledfruit stands (GC)-tall, inward-sloping pedestals, ac-cented with (frequently notched) horizontal ridges, thatpresumably supported open bowls-were thought at thetime of the Warka survey to date only from the laterEarly Dynastic period. To judge from surface materials,however, some examples must go back at least intoEarly Dynastic I times. It will be noted in Appendix Bthat there are several examples on Jemdet Nasr sites withno other recognizable Early Dynastic material at all, sothat the beginning of the use of these stands may be evenslightly older. Other traits that continue onward fromJemdet Nasr times include nose lugs on jar shoulders(ED); spouts of various form, mostly easily distinguish-able from earlier ones (classified as type FG, a residual,indeterminate category); and reserved slip ware (KA).If stone bowls (NA) continued in use at all-rememberingthat we are concerned with primarily secular use in gen-erally small, outlying settlements, not with their manu-facture for temple service or urban elites in the majorcenters-it appears to have been on a scale that was atbest very small and questionable. Clay sickles (LA),similarly, surely all but disappeared from general use bythis time. Probably the rare examples that do occur inEarly Dynastic levels are better interpreted as straysbrought up from underlying levels by various forms oflater disturbance rather than as objects of contemporaryuse or manufacture.

APPENDIX B

A GROWTH-SIMULATION PROGRAMRobert G. Hassert

Technical AspectsThe program used to generate the maps in figure 23,

24, 25, and 36 was written in FORTRAN and was runon the University of Chicago's IBM 370/168. The pro-

gram itself consists of only 164 lines of code, but it re-quires an additional 226K bytes of memory, mostly forformating the entire map image in a large two-dimen-sional array. Thus the program requires either a sub-

127

oi.uchicago.edu

Urban Origins

stantial core memory or an efficient paging environ-ment-both of which were available on our machine.The generation of the largest map, which included 645sites, took just under three and one-half minutes of CPUtime. This could probably be reduced somewhat by usinga more efficient compiler than the University of Water-loo's WATFIV.

The Problem

The problem we set out to solve with this programwas a very specific one-what was the probable con-figuration of cultivated land during various periods ofancient Mesopotamian history, based on the estimatedpopulation at all known sites in each period. In fact, how-ever, the program is applicable to a variety of relatedproblems-for example, the growth and encounter ofcell cultures in a Petrie dish, or the spread of competingplant species in a homogeneous environment. The keyelement in all of these situations is the steady expansionof a population from a point source, an expansion thattakes the form of a circle whose radius gradually in-creases until it encounters some obstacle. When the ad-vance is blocked in any direction it continues in whateverdirection is open until the total area occupied reachessome limit (in the present simulation, the amount of agri-cultural territory needed to support the population of thecity at its center).

The Algorithm

The present program is a simulation. It attempts toarrive at the results of the growth process by replicatingthat process, as described below.

First, the data for each site-horizontal and verticalmap coordinates and estimated population-were readinto the program. These data were then sorted by the es-timated population size.

A scale of 1:100,000 was chosen for the finished map.The site coordinates and distances were scaled to the mapsize and converted to print positions. The large scalewas chosen because each print position measures 0.1 inch(0.254 cm) horizontally by 0.125 inch (0.3175 cm) ver-tically, and in order to achieve the necessary detail withso coarse a grid size we had to print a very large map-500 print positions vertically by 430 horizontally. Thistranslates into a map 43 inches across by 62.5 inches fromtop to bottom-a size that required printing the map infour separate vertical strips to be joined by hand. Theresolution in real dimensions, therefore, is approximately317.5 meters vertically and 254 meters horizontally; eachprint position represents an area of slightly more than8.06 hectares.

A print value (a single letter or numeral) was assigned toeach site. Territory belonging to that site would be filledin with that character.

The main part of the program filled in the territory

for each site by circling around each site in turn, lookingfor unoccupied space. As the program found free terri-tory, it assigned it to the site it was examining, one printposition (8.06 hectares) at a time, and subtracted thatamount from the total remaining area required by thatsite. When all the territory around a given site had beenoccupied out to a particular radius, the program went tothe next site; when all sites had been considered at thatradius, the program increased the radius and repeated thewhole process for all sites that still required territory.

If we could watch the assignment of territory as itactually took place in the computer's memory, it wouldlook like this: each settlement is marked as a point onthe map; all these points begin to expand simultaneously,at the same rate, forming gradually expanding circles ofterritory; the smaller sites, and those with no near neigh-bors, soon reach the limit of their territorial needs withno interference and stop expanding; other sites, however,continue to grow and their territories collide; when thishappens, each such site ceases to grow where it abutsoccupied territory, but it continues to expand at the samerate in other directions, still forming a circular arc whosecenter is the settlement to which the territory belongs. Ifa site is completely surrounded, it continues to search foropen land beyond its borders, leapfrogging over the oc-cupied areas enclosing it. Only when the full territorialneeds of each site have been met is the map complete.

One of the technical problems involved in this searchprocedure was the choice of the radius increment andangle increment. As the program searches for unoccupiedcells at a given radius from a site, it must circle the sitein small jumps whose size is determined by the angle in-crement. If the jumps are too large, some available cellsmight be missed; if the jumps are too small, the samecell might be examined several times, resulting in slowerand less efficient execution of the program. The sameholds true of the radius increment, which determines thenew search radius once all cells have been filled withinthe current radius. After some experimentation we de-termined that workable values were a radius incrementof 0.085 inch (slightly less than the finest resolution ofthe map, the 0.1 inch horizontal spacing) and an angleincrement calculated as ARCSIN (0.8/RADIUS) (thejumps, and therefore the angle increment, must decreasein size as the radius increases). Each time the radius isincreased, a new angle increment is calculated. Thesevalues left no holes in the coverage of the occupied areasand caused each cell to be examined 1.8 times, on theaverage; this is probably close to the theoretical optimum.

The final steps in the creation of the map were theaddition of distinctive signs to mark the location of eachsite and others at the vertexes of the boundaries of thesurvey area. The finished map was then stored on a diskmemory unit, from which it could be retrieved and ex-amined at a terminal or printed on a high-speed printer.

128

oi.uchicago.edu

Urban Origins

The Results

Two of the major assumptions that underlie this simu-lation-that all sites begin at zero territory and that allexpand at the same rate-are patently suspect, and thisshould caution us against injudicious attempts to analyzethe configurations of individual territories. Nevertheless,the appearance of the finished maps is probably a fairlyaccurate reflection of agricultural requirements at theheight of the periods they represent.

The program could be modified to remove those

assumptions. For example, we might begin with theterritorial configuration of one period and expand (orcontract) the territories on the basis of data from thesucceeding period. Similarly, we might have individualterritories expand not at the same rate, but at a rate pro-portional to their size. These approaches would, ofcourse, involve their own assumptions.

Researchers are invited to use or adapt the programpresented here for their own purposes. We hope that itwill prove a stimulating tool for further experimentation.

129

oi.uchicago.edu

4Integration and

Fragmentation under Successive,Contending Dynasties

(Early Dynastic II-Middle Babylonian Periods)

Uruk's attainment of maximum size toward the end ofthe Early Dynastic I period signaled the end of a cumula-tive process of growth and change. Within less than amillennium an extensive, potentially fertile plain withonly small isolated pockets of permanent settlement hadbeen unevenly but permanently and in some cases denselyoccupied. Urban centers had appeared throughout theregion, attracting their new inhabitants from surroundingtowns and villages through what was probably a com-plex, continuing series of smaller-scale multidirectionalmovements. Superimposed upon the general process ofexpansion in numbers and localized concentration in andaround urban centers, however, was a major internalshift of population toward the south and southeast.

Growth, in other words, was neither universal nor uni-form. Some important towns had not only flourished butbeen abandoned, and the increasing predominance ofcities was everywhere accomplished at the cost of anaccelerating abandonment of the countryside. But intrends of settlement, just as in a host of institutional in-novations and cultural accomplishments, the sense ofcumulative development in societal scale and complexityoutweighs the deep and surely somewhat disruptive fluxthat accompanied it.

No similar accumulative trends characterize the twomillennia that follow. There is ample textual evidencefor the periodic emergence upon the scene of new ethnicand linguistic elements. But the settlement record, dividedinto fairly gross intervals of three centuries or so, is noteasy to correlate with geographically unspecific attesta-tions of new arrivals. Instead, the maps of successivesettlement patterns seem to show an ebb and flow ofpopulation into and away from outlying regions, ex-panding and contracting the nuclei of settlement and cul-tivation in response to fairly transitory political stimuli.Individual urban centers also experienced alternately ris-

ing and falling fortunes, a few of them briefly advancinginto the status of political capitals but all more commonlyexperiencing repeated, often prolonged, interruptions intheir power and prosperity. Urban institutions and ameni-ties flickered on in only a handful of the largest andlongest-lived among them as the two millennia with whichthis chapter deals drew to a close.

The alternating conditions of growth and decline, sta-bility and upheaval require detailed analysis in their ownright. Awaiting more systematic, securely dated study,for example, are the potentially significant effects ofvariations in precipitation and river discharge patterns.These could play a part in destabilizing settlements de-pendent on irrigation, and in addition they might wellbe responsible for inducing large-scale movements ofnomadic or seminomadic elements that at times werecapable of gravely disrupting urban life. Apart fromclimatic change in the strict sense, of course, human ac-tivities like deforestation and overgrazing also can exer-cise a large influence on runoff. However, quite apartfrom shifts linked to environmental changes, we mustrecognize that much instability can be essentially socialin origin. In any case, all these geographic and demo-graphic, and in an important sense also societal, param-eters are at least as crucial to an understanding of thehistoric record as the "dynasties, wars and religions" onwhich a far greater proportion of scholarly attention hastraditionally been lavished. But it must be conceded at theoutset that cyclical shifts as well as other conditionsduring the two-millennium span of early Mesopotamianhistory militate against a primarily archaeological studyas sequential and systematic as that made in chapter3 with respect to the late pre- and protohistoricphases.

One particularly adverse change involves a diminutionin the quantity and reliability of survey data, a conse-

130

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

quence of the increased locational stability of the majorcenters. Long, multiperiod occupations or reoccupations,or both, became much commoner. Particularly in view ofthe preponderantly urban character of settlement afterlate prehistoric times, only a much reduced proportion ofthe total occupied area during the succeeding periods isrepresented by either the latest or the sole occupationsof settlements recorded in a surface reconnaissance. Ac-cordingly, there is a steep decline in the accuracy-some-times even in the possibility-of estimates of aggregateas well as individual site areas. The approach conse-quently taken with respect to the sites dealt with in thischapter, in full acknowledgment of the increased uncer-tainty, is to deal with their areas only in terms of fairlybroad size categories rather than specific size estimates.Unfortunately, this precludes a number of approachesto the data that played an important part in the preced-ing analysis of the pre- and protohistoric periods, but toconvey a spurious impression of precision would be evenmore unfortunate.

Only with much more elaborate (and time-consuming)survey methods, based on intensive sampling of surfacematerials from all parts of each site, would reasonablyreliable estimates of occupied area during successiveperiods become feasible. Even then it must be kept inmind that the physical removal of occupational debrisand its presumed dispersal over surrounding areas is anancient practice that is very widely attested, particularlyin connection with preparing the foundations of newmonumental buildings. It seems quite likely, therefore,that only an extensive program of stratigraphic testingcoupled with careful assessments of spans of occupationthat are attested in ancient documents (e.g., Jones 1976,pp. 43-44; Stone 1977, fig. 2) can fully repair the de-ficiencies of surface collecting as a means of estimatingsite areas for the greater part of the third, second, andfirst millennia B.C.

Reference to the documentary sources reminds us thatthe availability of massive numbers of texts completelytransforms the character and potentialities of the ancientrecord. Authorities may differ on the relative potentialcontributions of archaeological and historical approachesto the subject matter of the fully historical portion of thisbook, but there is no doubt that heretofore the archae-ological contribution has been strictly secondary andancillary. Work has been focused primarily on large-scale,relatively uncontrolled excavations of public buildingsand tombs and has largely eschewed the quantitativelybased, interdisciplinary themes of investigation pioneeredby the prehistorians to whom no information from textsis available. It is certainly no longer true that Mesopota-mian archaeology for the historic ranges of time can befairly described (as to some extent was the case in theearly days) as a mining operation in search of texts.But the dominant strategies of study-the avenues of

investigation followed, the priorities, the questions asked-are still very largely those geared toward narrowlycorroborating and supplementing texts as well as maxi-mizing the chances for further textual recovery. Theirony is that in the long run such strategies are preciselythe wrong ones with which to complement and extendthe textual testimony most effectively.

Archaeological survey procedures, for all the unques-tionable deficiencies in this particular application ofthem, represent at least a small step toward rectifyingthe balance. A grid of comparable findings, more or lesssystematically obtained, is thrown over prehistoric andhistoric periods together. Questions become urgent andinescapable that otherwise could be postponed or avoidedaltogether: What was the changing nature of settlementand land use, not merely in the immediate hinterlandsof major cities but in the countryside as a whole? Towhat extent did the frequently asserted institutional con-tinuity and a textual stream of tradition reflect moregeneral social patterns of flux and continuity? How canwe begin to scrutinize the question whether the volu-minously recorded attitudes of the literate urban elitesare representative of those of the great, silent mass of thepopulation? How do we overcome the constraining effectsof successive textual-and specialized philological-genres in order to obtain a comparative and develop-mental perspective on the whole five-thousand-year sweepof Middle Eastern history?

Here we must undertake to ask some of those questions,however crudely and tentatively. Ideally, this should bedone as a substantial, long-continuing, collaborative en-terprise in which archaeological findings are continuallymatched with the judgments of historians and philologistsspecializing in the whole range of textual corpora fromearly Sumerian to medieval Arabic. Again, this can onlybe a very modest beginning, the expression of a hope asto what may one day develop on a vastly firmer footing.The textual sources employed, often uncritically andperhaps mistakenly, are those available in publishedtranslations. Large numbers of additional texts still awaitstudy and publication in any form, not to speak of im-proved translation. But it is time to begin asking a differ-ent set of questions of the available material, not onlyto help in interpreting the survey record but to exercisea reciprocal influence on the premises and priorities em-ployed in the study of the texts themselves.

The field of investigation of this and the followingchapter is necessarily limited. Insofar as possible, it skirtsthe often controversial questions of institutional structureand function that are as vital for a broad synthesis ofsocial and economic history as they are for political his-tory. Emphatically, this does not pretend to be any suchsynthesis. What it seeks to contribute to the eventualwriting of such a synthesis is primarily an understandingof the ebb and flow, as well as the underlying, cumulative

131

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

trends of development, of those patterns of land use andsettlement to which the findings of an archaeologicalsurvey are most immediately relevant.

An impression of ceaseless ebb and flow is particularlystrong in the lengthy time period dealt with in this chap-ter. For our subject matter it greatly outweighs most ofthe differentiating features of the long succession of dy-nasties. Hence I adopt a form of discussion that placesless emphasis on chronological sequence than was pro-vided in chapter 3. Instead, at least as a first step andeven as a predominant orientation, it seems more worth-while to highlight the apparent poles between which allthe contending dynasties tended to oscillate. Only in theconcluding section of the chapter will I consider fairlybriefly the actual sequence of Euphrates branches andthe settlements that accompanied them.

References to ebb and flow, or to flux and stability,provide an important qualification to my earlier com-ment that increased locational stability made archaeologi-cal survey techniques less effective. As these terms imply,the larger towns continued to exert an attractive forceon outlying populations over long periods. Agriculturalimprovements in their vicinity, and sometimes the ad-vantages of enclosing defensive walls, must account forpart of this. The occurrence in many cases of precinctsdevoted to particular divinities was surely also a factor,accounting for repeated reconstructions of temples asthe pious acts of individual rulers as well as some inwardflow of common people seeking the god's largesse orprotection. But these attractions emphatically do notalso imply stability in an individual settlement's size oreven its uninterrupted occupation.

Little systematic attention has yet been devoted tothe volatility of urban residence in early Mesopotamia,although this is at least beginning to be recognized as asignificant theme for study (Limet 1972; Renger 1972;Adams 1975d, 1978). One aspect of what was probablya prevailingly oscillatory pattern with regard to sizeand prosperity is documented by numerous referencesto the sojourn of individuals in cities other than the onesthey are identified as natives of. Merely to cite the EarlyDynastic IIIa example of Shuruppak, by no means a townof major importance, Jacobsen has noted that "'visitorsto the city' (uru(-se)gin) from almost all of the majorcities of Sumer appear in the accounts as working forthe palace and receiving rations." He also cites a difficulttext and its variant that apparently list more than sixhundred gurus (conscripted soldiers or workers), theShuruppak contingent being smaller than those fromUruk, Adab, Nippur, Lagash, and Umma in spite ofShuruppak itself having been the provenience of thetablets (1952, pp. 121, 122 n.70). Movements into andout of the major centers by individuals primarily identi-fied with smaller, less well known localities in their im-mediate hinterlands are likely to have been even more

numerous, but they were seldom or never noted as suchin the available records.

Physical destruction and ensuing decline of populationwere certain to be particularly severe in the case of citiesthat joined unsuccessful rebellions, or whose ruling dy-nasts were overcome by others in battle. The traditionallamentations provide eloquently stylized literary accountsof this, while in other cases the combination of archae-ological evidence with the testimony of a city like Ur'svictorious opponent as to its destruction grounds theworld of metaphor in harsh reality (Brinkman 1969, pp.311-12).

Uruk may have suffered similar vicissitudes after itsgreat florescence in the Early Dynastic I period, althoughin that case both literary and historical records are silent.Excavated remains of the later Early Dynastic, Akkadian,and even Ur III (excluding only the high terrace of UrNammu's ziggurat) periods are extremely sparse, if oneconsiders the scale and duration of the archaeologicalcampaigns in the vast ancient ruins. Surface ceramics inthe northern part of the site suggest that the later partof the lacuna will presently be filled there (H. J. Nissen,pers. comm.), as indeed it should if there is any substanceto the confusing references in the Sumerian Kinglist toa series of dynasts who both preceded and followed theGutian interregnum (ca. 2200-2116 B.c.) (Cassin, Bottero,and Vercoutter 1965, pp. 96-97). However, the EarlyDynastic II-III hiatus in material remains is still morepuzzling. Perhaps it implies that the last Early Dynasticruler, Lugalzagesi (ca. 2350 B.C.), assumed the title ofking of Uruk after a series of conquests elsewhere withoutfinding many living representations of the past glory ofUruk either to oppose or to support him (Sollberger andKupper 1971, pp. 91-95).

For the present it seems safer to assume that excava-tions at major urban centers like Uruk in most cases willone day reveal a relatively continuous record of occu-pancy. Uruk, after all, was still able to furnish one ofthe larger contingents of men listed in the Shuruppakaccounts mentioned earlier, well after the beginning ofthe putative hiatus. But it is perhaps worth noting thatparallel, archaeologically grounded arguments have beenadvanced for the abandonment even of the great sanctu-ary of Nippur for as long as three centuries in late OldBabylonian and early Cassite times (Stone 1977, p. 270,citing McG. Gibson). Yet Nippur at the time was in nosense a political capital and so could not have been ex-posed, as was Uruk, to political reprisals., As this suggests, flux and stability are persistent, closelyintertwined, and yet obviously also mutually opposingthemes. The heavy preponderance of towns and citiesis as good an illustration as any. City walls and granariesreduced external risks and smoothed harvest fluctuations,which was surely conducive to stabilization. Yet, on theother hand, urban dwellers were more easily impressed

132

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

and disciplined into superior army detachments, as usefulfor predatory expansion as for passive defense. Preciselythe existence of strong internal hierarchies and localizedloyalties reinforced city-states as the primary polities ofthe time. Larger, more centralized groupings were periodi-cally superimposed upon them, but both loose alliancesand protoimperial formations were in most cases ex-tremely transitory. As they disintegrated, it was the peren-nial, destructive rivalries between city-states that alwaysreemerged.

SThe essential point is that sequential or even simulta-neous tendencies toward flux and stabilization are equallycharacteristic of the historic record of settlement andland use. Excessive water diversions for irrigation improvethe short-turn prospects for adequate or above-normalharvests, as we have seen, and in that sense enhance thestability and security of the society whose members re-sort to them. Within a very few seasons, however, theconcomitant rise of saline groundwater seriously erodesor destroys agricultural productivity and thus has a highlydestabilizing outcome. Excessive upstream diversions sowlasting seeds of discord with downstream irrigators, more-over, that may lead to the short-run subjugation of thelatter but that can hardly be conducive to long-run socialstability. Decisions made in the light of considerationslike these naturally were modulated by many other factorsas well. Hence the local outcome varied from time totime and place to place. There is not, and never was, asingle, uniformly prevailing set of administrative andeconomic priorities with which the historical develop-ment of Mesopotamian society can be apprehended.&The obvious political corollary of these remarks con:cerns parallel, ongoing tendencies toward centralizationand fragmentation. With much wider application thanits specific reference to the Old Babylonian period (1894-1594 B.c.),rYoffee has recently outlined a cyclical par-adigm of imperial growth and disintegration that per-suasively articulates these tendencies. In the vigorous,early years of a dynasty, he argues, an efficient, highlycentralized "patrimonial bureaucracy" is largely recruitedfrom among the kinsmen and dependents of the royallineage. Its military and economic effectiveness is suchthat it consistently appropriates for its own ends a dis-proportionate share of the deployable resources pro-duced by its conquered constituencies. That highly op-pressive process engenders the formation of opposingalliances, especially since local elites, firmly rooted in theirown communities, are by no means swept away or en-gulfed by the new political system. As constituencies andwhole regions in time begin to break away, the resourceswith which the dynasty can impose its will progressivelyshrink. Local authorities are reinvested with powers toreplace them, siphoning away resources for their ownlocal ends and reasserting their hereditary rights. Oftenthere is a proliferation of official titles as the dynasty

bargains away additional prerogatives in order to meetincreasingly compelling short-term needs. The final rulersof a dynasty thus are left in the end as little more thanfigureheads in a largely disarticulated system. Hence thespecific forces and events leading to their overthrow canbe quite minor and almost accidental (Yoffee 1977, pp.147-49). 4

Yoffee emphasizes that this is only a generalized, ex-planatory model, not a historical narrative in which everyfeature can be correlated with textually documentedevents and relationships. It is intended to introduce someorder and coherence into the understanding of writtenrecords that in any case can never be expected to providemore than an exceedingly fragmentary account of ancientinstitutions from a very limited set of perspectives. Butit is sufficiently generalized so as to apply with virtuallyequal relevance across the broad range of socioeconomicand managerial categories-"temple," "palace," "state,""manorial," and "private"-that have been variouslythought to have been dominant during successive dynasticsegments of the two-millenium span here under review.It places principal stress not on institutional forms ap-pearing in particular communities as a result of a con-catenation of circumstances, but rather on the "complexrelationship between the political system and the localsocial structures, with no one overriding power com-pletely co-opting the others" (Yoffee 1977, p. 149). Theresult is a picture of endlessly renewed struggle in thename of stability, of a ceaseless renegotiation of conflict-ing central and particularistic claims to power.

SIn political regimes as in patterns of settlement andland use, it is thus clear that stability almost alwaysmay have been the objective, but flux was the prevailingoutcome. It is also clear that the ordered hierarchies-andtraditions of individual towns do not provide an adequatebasis for understanding this dynamism. The major sourcefor the latter lay in the relations between a wide arrayof differentiated communities-differently constituted interms of social organization as well as differently strati-fied, specialized, and endowed with resources. Through-out this chapter, in other words, we must strive to avoidthe impression that the individual city-state and its hinter-lands, or even a group of immediately neighboring citiesand towns, constitutes the elemental, seemingly "natural"unit of study.

Arguably, individual towns or groups of towns werea more nearly adequate frame of interpretation in pre-and protohistoric times. Chapter 3 placed considerablestress on them, although it also made reference to massivepopulation movements that could not have been strictlylocal in origin. But archaeology alone is at best somewhatambiguous in this respect, and the circumscribed, poorlyunderstood textual sources of Early Dynastic I and earliertimes could provide little additional insight into longer-range patterns of interaction. For the later third and

133

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

second millennia, on the other hand, textual sources make from third millennium Lagash, on the other hand, are byclear that it was often the interactions between fairly all odds the richest and most informative of any withdistant centers, as well as the broadest, most extensive respect to a broad range of social and economic ques-interregional relationships, that proved decisive. This tions of the kind relevant to this study. There is a tan-also implies, of course, that the geographically con- talizing mismatch of textual and archaeological data, instrained and somewhat arbitrary limits within which other words, underlining the inadequacy of the surveyedsystematic data on ancient settlement patterns are so far boundaries as they exist at present. Since heavy depositsavailable may considerably distort certain aspects of in- of silt have been laid down by the present-day Shatt al-terpretation. Gharraf and its effluents along what must have been the

Consider a few fairly gross, well-understood patterns western borders of Lagash, moreover, it is not clear howof interaction in which the central Euphrates floodplain, much surface reconnaissance in that direction can everthe primary focus of this study, was only part of a larger, contribute to an improved understanding of geographicalintegral unit. Kish, the references to which in royal titu- interrelationships.lary strongly suggest that it must have attained political Textual sources leave no doubt of recurrent, sanguinarysupremacy for a time in the early third millennium, lies struggles between Lagash and Umma over a contestedwell to the northwest of the region where an absence of tract of fields some 41 square kilometers in extent (Pet-contemporary cultivation makes intensive surveys possi- tinato 1970-71, p. 306). Once more it appears, therefore,ble. Babylon played an even more decisive role after the' that purposeful interruptions in the supply of water byearly second millennium, probably manipulating water " Umma to Lagash, the downstream user, were both ansupplies to gain military and economic advantage. Renger expression of the underlying conflict over a potentiallyhas suggested that the canals supplying Larsa in the scarce resource and an immediate pretext for the resump-south, Babylon's principal contender, were blocked or tion of overt hostilities (Jacobsen 1969, p. 106). A furtherdiverted already in the time of Sinmuballit (1812-1793 measure of response by Lagash rulers, just as by Rim-SinB.C.) and that his successor Hammurabi (1792-50 B.c.) of Larsa (1822-1763 B.c.) during the last desperate phaseslater relied on the same measure in his successful quest of his resistance to the advance of Hammurabi (Rengerfor political dominance over all the city-states of the 1970, p. 78), was the construction of an alternative feederalluvium (1970, p. 78). With such policies in effect, canal from the Tigris. While not a viable long-term solu-Babylon's fortunes were a harshly imposed reciprocal of tion to water supply interruptions, for hydrological rea-the well-being of cities situated below it in the floodplain: sons outlined earlier, repeated recourse to this step con-Babylon's own growth, not to speak of the widespread firms that the natural regime of the rivers was no longercolonization and canalization around it in the Old Baby- only passively accommodated to through localized, small-lonian period (Adams 1972, p. 186), either led to or was scale irrigation along the backslopes of adjacent naturalmade possible by equally widespread abandonments levees. Fairly lengthy canal construction was now a prac-farther southeast. After achieving political ascendancy, of tical possibility; indeed, essentially artificial canals morecourse, Hammurabi's interests shift to the restoration of than 15 kilometers long have been identified on purelythe now submissive countryside. Under those altered con- archaeological grounds as having already been in exis-ditions we find him taking credit for having dug a canal tence before the end of the fourth millennium (Adams andnamed "Hammurabi Is the Prosperity of the People" to Nissen 1972, p. 12). Purposive actions on a new scale thusfurnish water to the major southern cities (Reallexikon were becoming an increasingly significant factor in modi-der Assyriologie 2:180, no. 135, s.v. Datenlisten). Both fying the food supply and habitat either for better or forKish and Babylon, in their changing relationships with worse. Stabilization or the enhancement of local ad-cities like Nippur, Isin, Adab, Umma and Uruk, Larsa vantage must have been the immediate objective in mostand Ur, illustrate the potential conflict of interest between cases, but as often as not the outcome was a destructiveupstream and downstream irrigators for supplies of water oscillation between water surplus and water shortage.that were not infallibly adequate to meet both sets of \ Thus far we have dealt with "boundary problems" asneeds. The same source of rivalry, generally to the ulti- if the primary difficulties associated with a circumscribedmate detriment of the downstream consumer, not only geographical frame of analysis arise from the exclusionoccurs among distant political capitals but can also be of particular realms or city-states essentially similar tofound among relatively small, adjacent communities those that are included. This is, indeed, one problem. At(Adams 1975d, p. 4). least in certain periods urban interrelations can be visual-

A related set of problems arises if we turn to the eastern ized as a zero-sum game in which the gains of one or aperipheries of the surveyed area. The kingdom of Lagash few centers in southern Mesopotamia are more or lesslies in its entirety beyond the intensively surveyed limits, evenly counterbalanced by the losses of many others.and so far it has received only brief and limited topo- Hence, the omission of particular contenders from thegraphic scrutiny (Jacobsen 1969). The textual sources discussion, especially a city of central importance like

134

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

Babylon, would involve a serious distortion of condi-tions on the Mesopotamian plain as a whole. But theapplication of the zero-sum game metaphor to the Meso-potamian alluvium encounters further dangers in that itrests on an acceptance of the physiographic boundariesof the plain as a factor of constant and decisive import-ance. Insofar as the radius of interaction from time totime extended beyond this zone, use of the metaphor isat least partly contradicted by more distant patterns ofrelationships that yet help to explain cycles of growthand decline affecting the whole region.

Assyrian ascendancy during the last century or sobefore the fall of Nineveh furnishes an instance of this,although unfortunately it is one that cannot yet be il-luminated by the findings of an archaeological recon-naissance. Massive deportations at that time must havesubstantially, if perhaps only temporarily, reduced thepopulation of tribal regions in southern Babylonia. Thebrunt may have been borne by semisedentary groupingsthat left little tangible debris, but many of the smallerMiddle Babylonian settlements that are recorded in figure35 also must have been adversely affected. That cannotbe shown in the poorly dated surface materials attributedto this period, however, since not less than three or fourearlier centuries are conflated with the period of un-questioned Assyrian dominance. Note that the periodas a whole was one of marked reduction in sedentarypopulation. Clearly, therefore, that was a trend that con-siderably antedated the successive Assyrian deportations,having its origins at a time when political relations be-tween Assyria and Babylonia fluctuated repeatedly buton the whole were fairly evenly balanced. In any case,population transfers of this kind exemplify the short-comings of the zero-sum game metaphor. And one shouldremember that such transfers had occurred in the oppositedirection during earlier intervals of southern ascendancy,including the Akkadian, Ur III, Old Babylonian, andCassite periods.

The same need for a flexible, periodically widenedframework of analysis is further illustrated by the balanceof relations between Mesopotamian cities and the Elamiterealm in southwestern Iran. An imposed flow of peopleas well as resources can once more be documented in bothdirections. The frequency, although perhaps not the scaleand decisiveness, of movement between Susa and its Meso-potamian counterparts bound these two geographicallydetached regions together in spite of the considerable dis-tance separating them. Especially during episodes of ex-parision and contraction, in other words, an assessment ofthe balance of gains and losses between neighboring city-states must be supplemented by recognition of less sym-metrical, much more widely extended patterns of reci-procity reaching well outside the Euphrates floodplain.

There is a final broad class of interrelationships to bementioned, between Mesopotamian towns and the semi-

settled or wholly pastoral peoples who periodically movedaround and among them. The primary cultural signifi-cance of the alluvium, that is defined a semiarid but po-tentially irrigable zone, was at least initially meaninglessto these peoples. Their concern was with rangelands fortheir flocks, whether dependent on rainfall, seasonal flood-ing, or the residues of cultivation. Their movements ac-cordingly cannot be understood within a circumscribedgeographical framework but instead trace a continuumalong the whole of the Fertile Crescent, as well as intoadjacent mountainous regions. Urban relationships withthem transcend the scope of the present study in still an-other respect, for ordinary archaeological means are sel-dom sufficient to detect the presence of even substantialnumbers of herdsmen who were widely dispersed in tem-porary campsites.

Periodic attempts were made to interdict the naturalpaths of movement of these semisedentary folk, some-times on a scale as massive as that of Shu-Sin's (2036-28B.C.) Wall against the Martu, which is said to have run forsome 280 kilometers from the Euphrates to the Tigris andeven farther eastward (Wilcke 1969-70, p. 9). This couldhave been effective to a degree in slowing the penetrationof large, cohesive tribal units, like those often threateningto dominate the middle Euphrates around Mari. However,similar formations were for the most part not in evidencein Babylonia (Moran, comment following Adams 1975d).Fortifications and urban-based armies there would havebeen much less of a deterrent to seminomadic herdsmenorganized into smaller, more splintered groupings. In gen-eral, even the best-organized efforts were only briefly andpartially capable of resisting large-scale population move-ments that took the form of small, mobile, fluidly com-posed groupings.

This is a theme extensively dealt with by Rowton in anongoing series of studies on nomads and the "dimorphicstructure" of ancient Near Eastern societies. He has calledattention to the relative narrowness of the alluvial zoneoccupied by ancient cultivators, a point greatly strength-ened by the new archaeological evidence presented in thisvolume, and to its resultant openness to nomadic incur-sions. To the north and east of the main band-or, better,intertwining series of ribbons-of ancient settlement lay azone of gradually increasing but characteristically unde-pendable rainfall as one approached the Taurus-Zagrospiedmont. Grazing lands generally preponderated overlands devoted to dry agriculture there, and the extent ofthe latter was in any case at the mercy of numerous minorclimatic fluctuations. Hence this "dimorphic zone" con-stituted a kind of "pastoral corridor" that served repeat-edly to channel new groupings of nomads and semino-mads into close proximity and hostile interaction with thegreat urban centers of the Mesopotamian alluvium andtheir outlying dependencies (Rowton 1973, pp. 252-53;1976, pp. 20-24).

135

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

There are two respects in which I would slightly modifyRowton's reconstruction. A first and quite minor one con-cerns his emphasis on the dependence of rangeland on ad-equate local precipitation. This unduly narrows the po-tentialities of nomad movement and even dominance. Weneed to take fuller account of the effects of seasonal flood-ing along the Euphrates and of the importance as fodderof deep-rooted perennial weeds whose roots reach the sub-surface water table in irrigated fields left fallow duringalternate years. On the alluvial plain itself, therefore, con-ditions for pastoral movement were virtually as propitiousalong the Euphrates as along the Tigris, in spite of rela-tively modest differences in rainfall favoring the latter. Butthis is only a slight shift, to be sure. In a sense it rein-forces Rowton's primary emphasis on the vulnerability ofthe urban zone, adding penetration along its Euphratesflank to that along the Tigris. The existence of an irregu-larly oscillatory pattern, involving major as well as minorshifts in the power relations and relative proportions ofnomads and sedentary folk, seems impossible to deny asa recurrent, fundamental feature of the entire historicrecord (Rowton 1976, pp. 24-27; cf. Adams 1975d).

The second, more substantial difference involves thenature of the distinction between pastoralists and culti-vators in early Mesopotamia. Rowton's repeated refer-ences to dimorphism tend to delineate two sharply con-trasting, essentially antagonistic life-styles, pursued byethnically as well as structurally distinctive groups. In hisview, interaction, apart from hostile encounters of vari-able outcome, apparently takes the unidirectional form ofsedentarization of the nomadic or seminomadic elements.As I have already adumbrated in chapter 1 and elsewhere,I regard the groups immediately concerned as having beenmore in flux and less polarized. Ecological relationshipsdocumented in recent contemporary Mesopotamia may,of course, be qualitatively different from ancient ones. Buton the whole this seems less likely than that the accountsof ancient scribes, officials, and literati do not supply uswith entirely balanced and comprehensive testimony onmatters from which their authors were socially remoteand of which they were technically ignorant. As is thecase now, predominant emphasis on husbandry or cultiva-tion frequently must have been a shifting, pragmatic de-cision. Across the frontiers of cultivation there usuallymust have extended a structural and ethnic continuum,with the acculturation of particular groups proceedingbackward and forward between nomadization and sed-entarization according to circumstances. If so, the maineffect of semisedentary groups upon the predominantlyurbanized body politic of the lower Euphrates core landsis not likely to have stemmed from their alien backgroundand direct military potential. Instead, it was their em-bodiment of a practical and at times even preferablealternative for an oppressed rural peasantry and its coun-terparts in the semiurbanized working force, upon whose

continuing, docile productivity the whole edifice of power, .,/privilege, tradition, and ceremony that was lodged in cit-ies ultimately depended.

URBAN HIERARCHIES AND CONTINUITY

What can be said of the size and agglomerations of func-tions that characterized Sumerian and Akkadian cities?The actual hierarchy of town and city sizes, insofar as itcan be derived from survey data whose limitations havealready been indicated, is a strictly delimited empiricalquestion to be dealt with separately. The range of activi-ties and functions carried on in cities, on the other hand,is largely beyond the scope of this study, since it wouldrequire a tour d'horizon of virtually the entire corpus ofcuneiform and archaeological sources. Insofar as cognitivecategories are suggested by nomenclature, however, thereis nothing to indicate that the use of the term for city wastied either to a minimal size of settlement or to the pres-ence of specific urban institutions. The Greek city, clearlyidentified with an autonomous group of citizens and thepoliticoreligious institutions through which this groupfound its corporate expression (Martin 1956, pp. 30-32),clearly belongs in a different tradition. Perhaps the maindifference is that the Sumero-Akkadian city was a locusof contingent, shifting powers and prerogatives, whereasthe Greek city was built around a more self-conscious,formally constituted civic body.

Surveying the Sumero-Akkadian literary evidence,Hallo has collected references to a proliferation of formalsynonyms and antonyms for Sumerian uru, Akkadian alu.But he also notes that in connected, nonliterary contexts"the concept 'city' is expressed by a single term throughoutvirtually all the long history of cuneiform" (1971, p. 58).Leemans, similarly, speaks of cities and towns as havingbeen "indiscriminately lumped" under uru and alu re-spectively: "If we translate it by 'city,' we use this term inthe wide sense of a big walled town, generally with animportant temple . . . It is more difficult to distinguishbetween town and village, ville and village, etc. In sometexts, especially in the Old Babylonian period, the villageis denoted kaprum, corresponding with Sumerian e-durus,denoting a rural settlement" (1975, p. 135).

Edzard, following North, offers a further qualification:kaprum is not used in connection with nomads or semi-sedentary folk, but only for small sedentary settlements

v on the alluvium (1964, p. 145). Finally, the Chicago"Assyrian Dictionary provides a generalizing comment onthis lesser category: "The semantic range of kapru extendsfrom 'village' in agricultural surroundings, 'farm' for theproducing of barley, 'settlement' of shepherds of a moreor less permanent nature, to suburban agglomerationsaround cities. In the plural (kapratu) the word refers also,in a general way, to out-of-town regions" (8:190, s.v.kapru).

136

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

Thus we can identify a two-stage hierarchy of settle-ment that persisted throughout Sumero-Akkadian nomen-clature, a term for cities of very broad application and amore amorphous category of small rural settlements. Ad-ditional terms are known that refer (at least in somecontexts) to inhabited places of minor importance. Mas'-kanu, literally a threshing floor, appears as early as OldAkkadian times in a context suggesting that it meant afairly important rural place. Curiously, however, therewere seventeen uru-sag or "main towns" in Lagash dur-ing the time of Rimu§ but only eight mas-ga-na-sag, per-haps "main [smaller] settlements" or "encampments," togo with them (Falkenstein 1966, p. 41; Hallo 1971, p. 58,n. 14). This contradicts what is generally known of nestedhierarchies of settlement, in which lesser orders are alwaysseveral times more numerous than higher ones. Hence itis possible that the term referred not to a settlement in theusual sense, but to some functionally specified rural place,perhaps a centrally administered agricultural depot.Mainly in late (Neo-Assyrian) times, a qualification of theword for city comes into use with similar rural qualifica-tions. URU.MES "seems to denote a settlement, probablya manor" (Chicago Assyrian Dictionary 1/1:387-88 s.v.alu). Of course, references to other special-purpose sitesof limited size, such as fortifications, are not uncommonthroughout the historic record.

Reference has already been made to uru-sag, literallyhead-city. In the Akkadian period this must have includedsome farily modest district capitals, since there were seven-teen uru-sag within an area of rather less than 1,600 squarekilometers. By the time of the Third Dynasty of Ur, how-ever, the term has come to refer (in Neo-Sumerian per-sonal names) only to the imperial capital. There are sev-eral other Sumerian terms with similar connotations, allequated in Akkadian lexical texts with alum elum, high orupper city (Hallo 1971, p. 60). References to a hierarchyof urban importance tend to be more consistent and ex-plicit in Neo-Assyrian annals, where Sennacherib, for ex-ample, proclaimed his conquest of "a total of 88 strong,walled cities of Chaldea, with 820 hamlets within theirborders" (Luckenbill 1924, p. 54).

A somewhat untidy scheme of conceptual categoriesemerges from this discussion, although perhaps no moreuntidy than is to be expected when evidence is eclecticallydrawn from royal annals, lexical lists, administrative texts,and personal names over two millennia. Comparableterms in contemporary English are, after all, no less elastic.The title "city," particularly in the western United States,is often only a reflection of the original settlers' aspirations."Town" can extend in meaning from very minor settle-ments to major ones, the latter especially in informal usage(into town, downtown). The city-town-village trichotomyhas little functional significance, in fact, except for formalanalytical purposes. More disquieting as a reflection ondiscrepancies that may have lain behind the Sumerian

and Akkadian terms is contemporary Iraqi administrativeusage. Cities and towns are centers in which governmentalfunctions are exercised, and in censuses and statistical ab-stracts they are frequently recorded as being considerablysmaller than neighboring settlements classified as villages.

Returning once more to Hallo's overview of the cunei-form sources on this subject, the antonyms of urbanismare also instructive:

In literary Sumerian, the contrast "town and country"is commonly expressed by the pair uru and A-adam, liter-ally "town and pasture"; when used in an additive sense,the pair implies the totality of human settlement.... Whileone can only speculate about the etymology of A-adam,other Sumerian antonyms for the city put transparentstress on the hydraulic basis of the cultivated countryside.In contrast to the city, it is "that which is fructified withwater" (a-ri-a, 6-ri-a); it is "the moistened ground"(e-duru5) or the "irrigation district" (a-gar, literally per-haps the "water pocket"). Most of these terms passedwith little change as loan words into Akkadian (edura,ugaru).

With reference to the lexical texts and thus principallyto the high literary language, the contrast clearly definesat least one feature of the conceptual world of scribes andfunctionaries:

on the one side a diffuse, subjective, functional diversity ofdescriptive terms for the countryside, reflecting the urbanpoint of view and a succession of different linguistic strata;on the other a single term for the city, reflecting a basiccommon distinctiveness that apparently outweighed what-ever external differences divided the cities of one age orplace from another. [Hallo 1971, pp. 58-59]

To summarize further, it appears that only two addi-tional generalizations can be made with reasonableconfidence about that cognitive categories of ancient set-tlement. First, recognition is given to a subordinate, func-tionally specialized, nonurban class of settlement, the com-parative paucity of references to it presumably reflectingits lack of salience in the minds of urban scribes and ad-ministrators. Second, slightly redirecting Hallo's com-ment, the term city was applied, often with adjectivalqualifications, as a broadly inclusive category referring tosettlements of very limited size as well as to the very larg-est ones. From that point forward, however, we need toturn to empirical evidence from the survey on what thechanging size distributions actually were.

The basic data on changing settlement size from EarlyDynastic through Middle Babylonian times, drawn fromthe descriptive statements given in the general site catalog,is presented serially for all relevant sites in table 14 in theappendix to this chapter. As I noted above, I have em-ployed fairly broad categories of size rather than specific

137

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

estimates, reflecting the impressionistic quality of muchof the evidence. The smallest category, of 4 hectares orless, presumably applies to villages, small manors, spe-cialized agricultural, processing, or defensive facilities, andperhaps in some cases temporary encampments of non-sedentary peoples as well. The third through sixth cate-gories, covering settlements from just over 10 hectares tomore than 200 hectares, presumably fall in virtually allcases within the uru/dlu "urban" designation of our an-cient informants.

More doubt attaches to category 2, from 4.1 through10.0 hectares of surface area. Quite possibly this conflatessome marginal "cities," some specialized facilities, somemanorial estates, and some larger "villages" according tothe classification of the time, but the limited findings ofa surface reconnaissance provide no means for introducinga finer separation. Ultimately the test will come with moreextensive soundings in sites of the smaller and intermedi-ate categories. Villages, one assumes, probably were theloci of a relatively undifferentiated stratum of peasantcultivators, and hence should yield at most quite circum-scribed kinds and amounts of textual material. Whateverprovided a basis for including a particular settlementwithin the uru/alu category, on the other hand, probablyincluded at least some record-keeping and other forms ofmore or less regular involvement with elites and admin-istrators in the major centers.

It must be conceded that any such simplistic separationas this will probably be confounded by wide variability inancient practice. At least in northern Babylonia, some verysmall settlements were substantially involved, not merelyin routine record-keeping, but in more complex forms ofscribal and administrative activity. Tell al-Dhiba'i, ancientUzarzalullu, for example, covers only about 4.5 hectares,yet excavations there have brought to light not only a tem-ple of considerable size and the elaborate metalworkinginstallation and perhaps shop of a smith, but also a largecollection of tablets including a mathematical text withour earliest known application of the Pythagorean the-orem (Mustafa 1949; Baqir 1962; al-Gailani 1965; Ab-dullah 1967). Nearby, impressively walled Tell Abu Har-mal, ancient Shaduppum, was still more directly, perhapseven primarily, engaged in administrative responsibilitiesusually thought of as "urban." Its site plan includes atemple as well as other public buildings, and there werenumerous administrative, business, lexical, and mathe-matical documents, although it did not exceed 1.8 hectaresin area (Baqir 1947; al-Alusi 1959, pp. 47-48). Directlycomparable cases are not yet known within the intensivelysurveyed area dealt with in this volume, largely becausein the south excavators' attention has been all but ex-clusively directed to the great ancient cities. Tell Sifr, an-cient Kutallu (site 448), is perhaps the smallest site yetto have yielded documents, but it occupies an area ofnearly 30 hectares (Loftus 1857, pp. 263-72). A further

difficulty, of course, is that the use of writing on such sitesmay well have been extremely localized. Hence there isnothing to ensure that evidence for it will be foundwithout a fairly extensive program of soundings.

Table 14 is intended as an aid to the specialist interestedin the history of occupancy at particular sites or groups ofsites. Figure 25 further summarizes and abstracts the samedata, identifying trends in settlement across the two-mil-lennium span. It provides a breakdown by period andsize category of the 441 sites occupied for varying inter-vals during this span. Horizontal connecting lines acrossthe interstices between the vertical columns are intendedto give some measure of the degree of continuity betweenmajor periods, since their height is determined by the ag-gregate numbers of sites occupied in both periods.

There are discrepancies in the quality of data fromperiod to period yet to be dealt with, but they should notintroduce significant distortions into estimates of theproportions of settlements of different size within a par-ticular period. Differences are likely to be small, in otherwords, between the range of ceramic types that can beidentified on large sites as compared with contemporarysmall sites. It seems to follow that shifts over time in theproportions of the total settled area occupied by largeand small sites are a reasonably accurate index of changesin the urban hierarchy. The proportions form a strikinglyregular, even dramatic, progression, as table 12 shows.

TABLE 12 Declining Proportion of Urban Settlement in theThird and Second Millennia B.C.

Percentage LargePercentage Nonurban Urban

Period (10 ha or less) (more than 40 ha)

Early Dynastic II/III 10.0% 78.4%Akkadian 18.4 63.5Ur III-Larsa 25.0 55.1Old Babylonian 29.6 50.2Cassite 56.8 30.4Middle Babylonian 64.2 16.2

% It is evident from these figures that urbanization reachedmaximum proportions in late Early Dynastic times,roughly in the middle of the third millennium B.C. Fromthat time forward, for more than the following millen-nium and a half, trends in settlement ran strongly andcumulatively in the opposite direction. At least until theend of the Old Babylonian period, however, it is alsoevident that the bulk of the southern Mesopotamian pop-ulation not only remained urban but continued to clusterdisproportionately in cities of very large size.

The overwhelming concentration in large cities for aconsiderable part of this sequence must be regarded as ahypertrophic, "unnatural" condition for an agriculturalcivilization with preindustrial transport technology. Fol-

138

oi.uchicago.edu

: 4l 4.0ha 1rn 1114.1-10ha S10.1-20ha e 20.1-40ha =S 40.1-200 ha

Fig. 25. Settlement numbers and hierarchy of size in earlier historic periods.

139

MMi200+ ha

. ... ....... ........... ...

.

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

lowing the standards of population density in settlementsand cultivated land requirements per person that wereapplied earlier (see above, pp. 69, 85), the median dis-tance to a field used as a source of subsistence for a siteof 10 hectares or less would have been less than 1.6 kilo-meters. For an early city covering more than 40 hectares,on the other hand, the median distance would have beenmore than 3.1 kilometers. In the latter case the averagedaily travel time to and from fields thus was an hour orso longer, or even more with teams and agricultural im-pedimenta, to the marked disadvantage of cultivators liv-ing in the larger urban centers.

This should not be taken to imply that a law of "leasteffort" prevails uniformly in matters of settlement. Werethat so, urban concentrations of any kind would be quiteinexplicable. A dispersion of cultivators into smallercommunities often must have been an immediate conse-quence of changes in social stratification and land tenurearrangements, or of an improvement in rural security. Thelatter, in particular, would tend to counteract one of theprincipal forces that probably had produced hyperurbani-zation in the first place. Ethnic factors may also haveplayed a role, with incoming semisedentary folk normallygravitating to small, newly founded communities of theirown. There was a particularly large shift in relative pro-portions of small and large settlements at the outset of theCassite period that may well illustrate this process. But thedebilitating loss in agricultural efficiency that large citiesfaced, because of the wider belts of cultivation theyneeded, must have been a steady and sometime decisiveforce behind the progressive, long-continuing dispersion.

Returning to figure 25, further consideration must begiven to the horizontal lines between the columns thatprovide some measure of continuity between successiveperiods. To be sure, ceramic index fossils confirming oc-cupation of a site in two successive periods do not neces-sarily indicate that it was continuously occupied duringboth of them. If Gibson is right, even a major center likeNippur could have well-documented Old Babylonian andCassite occupations and yet have been abandoned for sev-eral centuries between the two (see above, p. 132). But theproportion of sites of one period on which an occupationduring the following period can also be confirmed is surelyrelated to the degree of continuity between the two pe-riods. A high proportion is a necessary but not sufficientcondition, one might say, for a high degree of residentialcontinuity.

Several general observations about continuity emergedirectly from figure 25. The larger cities, not surprisingly,seem to trace out the smoothest curve over time, with leastevidence of deep interstitial troughs between major pe-riods. However, there is nothing to suggest that the small-est settlements, those in category 1 covering an area of 4hectares or less, were significantly more volatile in theirpatterns of occupancy than the towns and small cities in

the two next larger categories. On the whole, settlementsof all sizes exhibit roughly the same patterns of continuitybetween periods, save that the interruptions are somewhatdamped for the largest ones. On two occasions, however,after the Early Dynastic and Akkadian periods, the small"village" sites appear to have been at least as stable as allof their larger counterparts.

Bearing in mind that considerable periods of abandon-ment may have occurred that would not necessarily bedetected with the reconnaissance methodology followedin this study, figure 25 suggests that at times there wasrelatively high turnover. The tops of the columns traceout a more rapidly fluctuating curve than the troughs be-tween them, since the former include a component ofsingle-period occupations that may have occurred in re-sponse to various temporary state programs of expansionwhile the latter are based only on sites that persistedthrough at least two periods. Differences from one columnto the succeeding one must reflect substantial, continuingprocesses of sedentarization and population growth ifthe later exceeds the earlier in height, or the reverse if itdoes not. But comparing succeeding crests with the troughbetween them is also instructive. In most cases the troughsare substantially lower than both adjacent columns. Thedifference between the site count of the trough and thesite count of the lower of the two adjacent columns is atleast a rough measure of the proportion of total settlementthat was abandoned and then fairly quickly relocated else-where, as distinguished from net increases or decreasesin the settlement pool and from those that were continu-ously occupied. In some cases this applies to almost a thirdof the total. Intervals of especially disruptive relocation,as identified in this manner, will be observed during thetransitional phases leading into the late Early Dynastic,Old Babylonian, and Cassite periods and must suggestparticular volatility in conditions affecting settlement atthose times./ Several factors may be responsible for these oscillations.Certainly one is the physical destruction or forced transferof communities in a period of unification following inter-dynastic rivalries. Improvements in the irrigation systemare another, frequently mentioned concomitant of con-trol by new dynasties. Newly dug or desilted canals andheavy investments in facilities like weirs and regulatoryheadworks would have attracted settlers even without anelement of compulsion. While the construction and main-tenance of major new irrigation works was usuallyclaimed as a royal responsibility, in practice the stateseems generally to have limited its role to providing thenecessary resources to intermediary labor brokers or localcommunity headmen and then inspecting periodically toassure performance (Walters 1970). Without discountingthe presence of a strongly authoritarian element in theorganization of dynastic states, therefore, there is reasonto be skeptical that the primary cause of repeated aban-

140

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

donments and resettlements were levees en masse to openup new areas for irrigation or other, similar applications ofdirect, centralized control. To recall the example of Arabcultivators in the last century, moreover, they were ableto build and maintain fairly extensive local irrigation sys-tems without the assistance of-even in the face of occa-sional military opposition from-the Ottoman state offi-cials in Baghdad (see above, pp. 23-25; Moritz 1888;Fernea 1970; Adams and Nissen 1972, chap. 5). At leastsome of the turnover in settlement evident in figure 25,particularly that among the smaller sites, thus may havebeen outside the realm of state interest or concern. Thatwould apply particularly to nomadic or semisedentaryelements in the population that had not yet been integratedinto-or that had broken away from-such ongoing in-stitutions of state authority as there were in the remotecountryside.

An additional factor reducing continuity of occupancyhas been suggested previously. In connection with thetenuousness of any reconstruction of local geographic andsocioeconomic trends as a zero-sum game, I referred tointerdictions or diversions of some of the major rivers andcanals by upstream dynastic contenders. Downstreamrivals naturally would seek to restore the needed suppliesof irrigation water insofar as they were able; by regularlymaintaining a variety of alternative water sources, byhurriedly constructing substitutes that would serve atleast for short-term, emergency use (e.g., a canal from theTigris), or by military action. The net effect was increas-ingly to convert the basic hydrographic system of theregion from a feared but accepted natural condition intoan object of conscious manipulation. Herein must lie apotent destabilizing influence on settlement-though onewhose actual importance in any given instance may neverbe convincingly established.

The difficulty in determining how far this factor waseffective stems from the always ambiguous intersection ofhuman actions and intentions with the powerful and un-predictable behavior of a natural river system. Understrong kings like Hammurabi (1792-50 B.c.) any outcome,whether the diversion of flow away from the competitorsor the preservation.of the existing system, probably wasby design. Sufficient time, control, and resources were athand to restore even the main body of the Euphrates toits bed if it breached its banks during a flood and brokeaway to form a new, undesired course to the west. Underweak kings, on the other hand, a diversion might be dugand encouraged to grow as a political expedient, but themeans frequently must not have been available subse-quently to restore the status quo. In the declining yearsof a dynasty, moreover, flood breaches surely appeared inweakened dikes without any human intervention and thencould not be repaired with the limited forces at hand. Atleast according to the account of Baladhuri, this was theorigin of the Great Swamp at the end of the Sasanian

period (Le Strange 1905, p. 27). Only if diversions can beassigned to a particular reign, in other words, can theybe attributed with confidence to human as opposed tonatural agencies. But while the survey data document amajor westward movement of the Euphrates during theearly to mid-second millennium, surface collections arecharacteristically inadequate to support a dating withinsuch close chronological specifications.

The question of volatility of settlement has heretoforebeen viewed solely in terms of changing conditions on thecentral Euphrates floodplain. In those terms, it has beenpossible to speak of an at times considerable turnover inoccupation from one period to the next, and to suggestsome of the factors that may have been involved in it. Butwe must conclude this section by considering the questionin a somewhat wider context. In relation to HenryWright's discussion of the region around Ur in the Ap-pendix to this volume, this area was strikingly more stablein almost every respect. Wright finds, for example, thatalmost 90 percent of his Late Larsa-Old Babylonian sites(essentially the same as those called Old Babylonian here)had not been occupied during the preceding period, andthat the same applies to almost 80 percent of his Cassitesites. For central Sumer these figures are much lower, 36and 46 percent respectively. Here we find a remarkabledemonstration of the greater continuity that obtained inthe core of the alluvium than along at least its southernperipheries. In other words, stability was directly corre-lated with centrality. If settlement is any indication, it is tothe core of the heartland that we should look not only forthe firmest retention of traditions but for leadership in allthose institutions and arts whose cumulative growth isencouraged by not being repeatedly and forcibly trans-planted.

UPPER AND LOWER LIMITS OF SETTLEMENT,AND ASSOCIATED PATTERNS OF LAND USE

The urban hierarchy, as elaborated in the previous sec-tion, has few, and only rather general or indirect, impli-cations for an understanding of the patterns of subsistenceon which it was based. It is reasonable to argue, for ex-ample, that the clustering of a large proportion of the pop-ulation in the uppermost tiers of the hierarchy suggests afairly high level of agricultural productivity. Otherwisethe necessary subsistence resources would have been dis-persed over intolerably wide areas. A heavy urban con-centration also suggests a rationalized and efficient systemfor transferring the foodstuffs into the cities. At a BronzeAge level of technology with limited and inefficientwheeled transport, the necessary scale of the operationimplies primary reliance on movement by boat andbarge. ¥

It is difficult to imagine a set of conditions more con-ducive to the precocious development of urbanism, be-

141

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

fore the onset of industrialization, than those that ob-tained on the ancient Mesopotamian plain. A "corona"of irrigation agriculture around each town (Oppenheim1969, p. 6) provided substantial yields of cereals, dates,and garden crops, the heightened irrigation requirementsand productivity of the gardens and orchards dictatingthat they should be concentrated immediately around thetowns, while the fields could lie farther away. Networksof natural waterways, easily modified or extended in theabsence of any natural obstacles, provided the necessaryarteries of large-scale transport. Also provided by theselagoons and waterways (as well as by the waters of theGulf, not far away) was a habitat for fish, probably themost significant source of protein.

Meanwhile the herds, primarily of sheep and goats,occasionally received cut barley but on the whole werealternately pastured on gleanings and stubble in outlyingfields, on young growth of barley, in seasonal depressionsas the water receded, and on the sparser but ubiquitousvegetation of the open steppe (Nissen 1976, p. 33). Cattle,present only in more limited numbers (on the order of 10percent of the number of sheep and goats), must havebeen kept closer at hand, both because their primary useas draft animals required this and because the sparse tex-tual references to milk are almost entirely to cows' milk.Goats' milk, apparently obtained from flocks generallykept too far from the cities to be imported before spoiling,was converted to butter and cheese in minor quantitiesand otherwise must have been consumed locally by pastor-alists and rural cultivators. Sheep and goats were thusthe most distant component of the system, but as sourcesof wool, hides, and (much less important) meat they couldbe driven at little cost to the places of processing or con-sumption. A single herdsman could adequately supervisea flock of about a hundred sheep and goats, includingmoving it to new feeding grounds and even into the citiesfor disposal.

To go beyond this harmoniously meshed but purelyqualitative picture, we must turn from enumerating typesof resources and their procurement to assessing their prob-able scale. Just as the proportions assigned to differenttiers in the settlement hierarchy shifted through time, sowe must assume that the total in all the tiers was notstatic but tended to fluctuate. The reliability of the evi-dence for different periods also fluctuates, underminingany attempt to reach a fully independent estimate fromsurvey data. What is available from the survey, moreover,consists only of estimates of site area. As I have alreadyargued, the relationship of site area to population wassurely not static and in any case was at best a fairly looseone. Thus the question of scale is complex, and the an-swers to it will be necessarily problematic. But onlythrough a quantitative comparison of different periods canwe form an impression of the dynamics of the underly-ing demographic and economic system, of which the pre-

dominance of different categories of settlement are on thewhole only a passive reflection.

Figure 25 provides at least a rough indication of rela-tive population levels during successive periods. Apartfrom other problems, precise estimates are precluded bythe fairly broad brackets for each of the size categories.As a first step toward an approximation, the midpoints ofthe six categories may be said to form the following se-quence of ratios in hectares: 2:7:15:30:100:200. By multi-plying the number of sites in a particular category andperiod by the relevant ratio, we can construct an ad-mittedly rough and provisional index of total settlementareas. This is given in table 13. The area totals therein

TABLE 13 Totals of Assumed Site Areas in Hectares, bySize Category and Historic Period

Size Category

("Village") ("City")1 2 3 4 5 6

Period (±2 ha) (±7) (±15) (±30) (±100) (±200) Total

Late Early Dynastic 52 112 75 120 1,100 200 1,659Akkadian 86 175 135 120 900 - 1,416Ur III-Larsa 286 399 240 300 1,100 400 2,725Old Babylonian 216 315 150 210 700 200 1,791Cassite 330 413 105 60 400 - 1,308Middle Babylonian 200 196 30 90 100 - 616

can be converted to a yet rougher approximation of popu-lation by multiplying once again by our familiar-andprobably conservative-constant of 100 persons per hec-tare (but recall that the constant applies not to net siteareas but to larger rectangular areas enclosing sites attheir longest and widest dimensions). Added to the un-doubted variability that this (or any other) constantmasks, however, are the uncertainties introduced by usingnot actual size estimates but size categories. The cumula-tive discrepancies thus may be large, and a conversion ofareas to population levels admittedly must be veryspeculative.

Even as relative proportions, without attempting tospecify their significance for the determination of abso-lute population levels, there are difficulties with these fig-ures that should not be ignored. The entries depend, to agreater degree than in similar tabulations in the previouschapter, on the identification of ceramic "index fossils"whose spans of use either are poorly known or cannot beunambiguously assigned to a single period. Particularlydoubtful are the groups of types used to identify the Ak-kadian and Middle Babylonian periods, for neither ofwhich could I unequivocally establish clear, one-to-oneassociations of commonly occurring types of surfacematerials with all or part of the chronological span of theperiod.

142

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

A related problem is still more serious. For reasons dis-cussed in the appendix to this chapter, I have not judgedit possible in most cases to distinguish on the basis ofsurface collections between settlements occupied duringthe Third Dynasty of Ur and those of the succeeding Isin-Larsa period. Together, these span some 330 years. Thisis not in itself excessive, at least in terms of this study'semphasis on wide geographic coverage and hence rapidityof application. Included within the span, however, arethree separate, successive political configurations orga-nized around different capital cities. Almost certainly therise and consolidation of each of these dynasties in itsturn was accompanied by substantial shifts of population,in response to numerous changes and improvements inthe canal system and the founding or relocation of towns.This being so, the totals given in table 13 to some extentconflate sequential developments during the Ur III-Isin-Larsa period, thereby considerably adding to the figuresthat are shown. There is no reason to doubt that the maxi-mum extent of pre-Hellenistic settlement (and population)occurred during this period, as the table attests. But themagnitude of the difference between it and the precedingand following periods probably is less than the figures inthe table suggest.

These and similar difficulties sharply restrict and qual-ify any demographic conclusions from the data of thesurvey. Yet it cannot be denied that there is a substantial,internally consistent mass of evidence from which we maydraw some admittedly impressionistic generalizations.Broadly speaking, the population levels attained in thelate Early Dynastic period seem to have continued untilthe last century or so of the third millennium. Internalshifts from district to district occurred during this spanof almost a millennium, but there appears to have beenlittle advance or decline in the regional aggregate. Then,within a relatively short period at the end of the millen-nium, there was a sharp increase in the numbers of sitesin every size category. More attention will be given belowto the meaning of this change for matters of subsistence,but it surely must imply a population maximum wellabove anything seen earlier. Thereafter the trend wasirregularly, sometimes steeply, downward. By about thebeginning of the first millennium B.C., there is a muchsmaller recorded aggregate of settlement area than forany period after the late fourth millennium." This account of demographic change, though internallyconsistent and supported by the available survey data,raises a number of problems when viewed in a wider con-text. Perhaps most important are the implications of thetotal occupied site areas given in table 13 for populationand hence for the corresponding extent of cultivation,particularly as compared with the initial urban climaxdescribed in the preceding chapter. Recall that less than600 hectares of site area was occupied during the Urukperiod, and that even as late as the Early Dynastic I period

the total was only some 1,075 hectares. The latter figurewas somewhat inflated by the great expansion of Uruktoward the end of Early Dynastic I times, moreover, sincesequentially occupied areas could not be distinguishedfrom simultaneously occupied ones. Yet by the end of theEarly Dynastic period it is now suggested that the occu-pied area had climbed a further 54 percent over even theinflated figure. And the occupied area in the late thirdmillennium, admittedly inflated by the same inability todistinguish between sequential and simultaneous occupa-tions, is 153 percent above the Early Dynastic I figure.

,Is it reasonable to conclude that almost a fivefold growthin total population occurred, corresponding to the growthin occupied areas of archaeological sites, between lateUruk times and the Third Dynasty of Ur? Taking intoaccount the deep, almost continuous, kaleidoscopic mili-tary and political shifts of this period of about a millen-nium, could there have been sufficiently high and sus-tained internal growth to account for an increase of thismagnitude? Or do we need to conclude instead that therewas extensive immigration?

Satisfactory answers to such questions are not easy.Much depends on the purportedly constant density of 125 >persons per hectare of occupational debris (cf. chap. 3, n.6), once again conceding that this or any such averagemasks a great deal of hitherto unexplained variance. Butwas the density the same for large as for small sites? If so-and the existing data do not really provide a basis forsupposing otherwise (Kramer 1980)-it is difficult to con-tradict at least the main thrust of the quantitative com-parisons just given.

A direct calculation of cultivated areas (on the basis of1.5 hectares per person) is unreasonable, since for thelate third millennium we are dealing only with assumedaverages of size categories rather than with total mea-sured areas. However, the approximate effect of an in-crease of this magnitude can be seen by referring oncemore to figure 24. The combined area of the two polygo-nal enclaves that are shown for the Early Dynastic I per-iod would, if entirely shaded, be only insignificantly largerthan the cultivated area needed to accommodate the latethird millennium population whose cities and townsmostly occur in the same areas.

Such a reconstruction implies that there was a broad,contiguous zone of cultivation connecting most of thelarger towns and city-states. However, that zone wouldhave constituted only a relatively small fraction of thetotal alluvial land surface. Flanking it along both sideswould have been substantially larger areas in which therewas little if any sedentary population. Admittedly, nocredible, comprehensive statements about population,with which we can check urban densities directly andhence substantiate this picture, have yet been encounteredin ancient textual sources. However, the shift from sep-arate "coronas" of cultivation around each major center

143

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

to a more fully integrated, zonal system is indepen-dently suggested by the evidence of irrigation nomen-clature.

The importance of this evidence has recently beenstressed by H. J. Nissen. He notes that none of the tech-nical terms for irrigation facilities and their managementand normal functioning (e.g., weir, settling basin, gate-house, offtake or inlet) appear in this context before lateEarly Dynastic times (Nissen 1976, p. 23; cf. Sauren 1966,pp. 35-83). Clearly this suggests a rapid technical develop-ment of irrigation works, connected with an equally sub-stantial and rapid growth in their scale, concomitant withthe growth of urban population that has been suggested.Precisely this kind of growth would have been needed ifthere were to be a transition from localized, ad hoc, gen-erally small-scale irrigation concentrated along the back-slopes of particular natural levees to extensive, increas-ingly artificial, intercommunity systems. Probably presentearlier, this tendency can best be seen in the survey datafor the Ur III-Larsa period (fig. 31). Not only are exten-sive, dendritic systems of branching canals then in evi-dence, but there were even subsidiary canals parallelingthe major channels (e.g., west of Adab) to link the up-stream and downstream components of the system morefirmly.

Also to be considered are the implications for urbanpopulation density of average house-plot size as suggestedby real estate conveyances. A distinction from buildingswith public or semipublic functions is not always easyto draw, but Gelb speaks of an "average" Old Baby-lonian private house as having been about 67 gin, or 39.4square meters, in size (1976, p. 197). Following Russell(1958, p. 12), we may assume that a nuclear family of fiveis an ideal, not an average, and set the latter closer to 3.5.Assuming further that at most half of even a denselybuilt-up area could be given over to actual living quar-ters, in order to allow for other kinds of constructionand routes of access, densities of just under 450 personsper urban hectare are quite reasonable.

The difficulty is, of course, that the occurrence of suchdensities in some neighborhoods, and perhaps even inmost districts given over to private housing, does notfurnish a basis for extrapolating the population of entirecities. The Gilgamesh Epic speaks of Uruk, assuredlysomewhat metaphorically, as having contained equalmeasures of orchards, clay pits, and city districts, as wellas the temple precinct devoted to Istar (Chicago AssyrianDictionary 1/1:380 s.v. alu), suggesting that less than athird of the average for the built-up urban area as itexisted at any given time would apply to the whole areaof the city as an archaeologist would calculate it fromsurface data. Moreover, there were surely many special-purpose buildings other than temples within the activelyoccupied part of the city, including storage facilities,workshops, and provisions for temporarily accommodat-

ing herds and perhaps rural folk in times of crisis (e.g.,Smith 1932, p. 297).

None of these countervailing considerations can beproperly quantified, at least at the present stage ofarchaeological and textual study. But we cannot entirelyexclude the troubling possibility that urban populationdensities sometimes were "only a small fraction of con-temporary densities in villages and small towns" (Adamsand Nissen 1972, p. 30). Perhaps significantly, it has beenshown for one of the few adequate bodies of modernMiddle Eastern materials relating density to settlementsize that the same type of reduction occurs. In a groupof fifty-four Khuzestan villages that all occupy less than 4hectares, the average density of the smallest third exceedsthat of the largest third by 68 percent-admittedly withmuch unexplained variance in each category (Wenke1975-76, p. 90). However, one cannot reasonably ex-trapolate from modern landlord-controlled villages of asingle type to ancient towns. Pending much larger ex-posures (or more systematic samples) in excavations, thereis no basis for assuming that average densities oftendropped below 125 persons per hectare in any of thelarger centers, while in some of them it is quite possiblethat the average was a great deal higher.1 Russell's similarfindings for medieval European cities (1972, p. 28) perhapslend further support to this judgment. Let me conclude,then, by reaffirming that something on the order of afivefold increase in population apparently took placeover the millennium or so after late prehistoric times,with consequences for the irrigation regime that havealready been noted.

Two other types of quantitative evidence potentiallyrelevant to a determination of agricultural and populationlevels are to be found in the voluminous administrativerecords of the Third Dynasty of Ur (2111-2003 B.c.).The first concerns workers hired or assigned to assist inthe harvest and other agricultural work in southernMesopotamia; the second, receipts of sheep and wool atstate intake depots like Puzrish-Dagan (or Sellush-Dagan).In both cases very large numbers are sometimes specified,the entire context of the recorded operations making clearthat accurate counts rather than estimates or propa-gandistic statements were regarded as imperative. Neithertype of information can directly furnish data on the totalpopulation of individual towns or even the region atlarge. But together they throw considerable light on over-all patterns of land use and are of some assistance inestablishing more securely the orders of magnitude forurban populations that have already been put forward.w Representative of one type of accounting for harvestworkers is a text from Drehem (Puzrish-Dagan or Sellush-Dagan) that has been extensively discussed by Goetze(1963).2 A total of 21,799 workers are listed, brokendown into contingents under named captains from anumber of towns and cities whose provincial rulers are

144

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

also named. Among the centers dealt with are several that west and Umma about 85 kilometers southeast. If ancan be definitely identified within the intensively surveyed allowance for travel time is made in addition to the timeregion, including Umma with 2,600 eren, Adab with consumed by the harvest itself, the problem arises of how1,800, Shuruppak with 1,200, and Isin with 180. The at least the more remote contingents could have copedgreater part of the group apparently was recruited from with the harvest of their own fields. The spring harvest isthe region northwest of Nippur, extending as far as the period of most intensive labor during the agriculturalSippar at the upper end of the alluvium but not including season (Adams 1965: pp. 14-15). Once the crop is readythe lower Diyala plain across the Tigris. its prompt completion is made urgent by heightening\/In a general way, the quotas assigned to particular cities losses to pests in the increasingly intense heat of earlycorrelate with the sizes attributed to those cities in table summer and by unavoidable waste of grain when the14, on independent, archaeological grounds. The table stalks become too dry and brittle before being cut.places Umma and Adab in category 5, centering in the Closely coinciding with the work in the fields, moreover,neighborhood of 100 hectares, and Shuruppak in cate- were equally urgent needs for large numbers of men togory 3, centering at about 15 hectares. Isin is a special guard and repair the dikes and levees along the majorcase since it is in the largest category only on the basis watercourses lest large areas of standing crops be de-of the change in its fortunes after the Ur III period. At stroyed by the spring floods. For an especially crucialthe time of this text, in the second year of Amar-Sin period of not less than several weeks in April and May,(2045-37 B.c.), it was apparently still of limited im- therefore, competitive demands for labor on state fieldsportance. To be sure, some of the men listed may come around Nippur and in the districts from which thefrom towns and villages under the control of the same workers were recruited would sharply scale down the,ensi rather than from the district capital where he exer- proportion of the local population that could be enlistedcised authority. But provisionally excluding this as a for the purpose. Then our estimate of the density in therelatively minor correction, the two towns to which pop- larger cities might need to rise accordingly.ulations on the rough order of 10,000 have been attributed There is an ordinarily unrecognized factor, however,would have furnished something on the order of one-fifth that reduces or even eliminates the problem of simulta-of this number for agricultural work. Shuruppak, on the neous demands for harvest labor: the date of the harvestother hand, seems to have furnished a much higher pro- is not uniform throughout southern Iraq, but advancesportion, suggesting either that the estimate of its size as one moves northward. Ancient records confirming thisduring this period is defective or that in this case it was phenomenon are unfortunately not available, but thereexpected to recruit heavily from subsidiary settlements is no reason to believe that the sequence as outlined innear it. figure 26 from modern Iraqi agricultural statistics is not

Perhaps all that can be said is that there is nothing a very close equivalent.inherently unreasonable in the idea that one-fifth of the In light of the differences shown here, the work forceurban population was recruited for migratory but tem- probably was initially assembled in the southern part ofporary harvest service. Excluding women and children, the district and then moved northward, closer to theofficials and other exempt categories, herdsmen who districts from which most of the men came, in time tocould not leave their flocks, and surely some additional participate in the harvest there also. Acting to lengthengroups of adult males engaged in other essential services, further the available interval was the greater concentra-an even higher proportion might have been available for tion of barley on the heavy, poorly drained and hencea few weeks or even months without unduly straining more saline soils of southern Iraq than in the northernthe local economy. The total size of the work force is also part of the alluvium (Jacobsen 1958, pp. 12, 26-27), fornot remarkable, at least when it is considered that the even where the crops occurred in adjoining fields theeren were drawn from the entire alluvial plain between traditional practice was to reap the barley before thethe Tigris and Euphrates except for cities like Ur, Uruk, wheat (El-Samarraie 1972, p. 62). That tradition can inLarsa, and Lagash in the extreme south. To phrase these fact be demonstrated in the modern statistical data asobservations differently, even though the initial impres- well, though for simplicity of presentation wheat andsion is that large numbers of men are dealt with, there barley are combined in figure 26.is certainly not a convincing case that average densities y. How large an area could a group of about 22,000 in-in the larger cities needed to be any higher than 125 dividuals have been responsible for harvesting? The ques-persons per hectare to provide them. tion is an important one if we are to reach some under-

Goetze assumed that the entire group referred to in standing of the importance of state-recruited and ofthis text was assembled to assist in the harvest around locally employed agricultural labor. At least an indirectNippur, roughly in the center of the region from which approach to an answer is provided by the Manishtushuthe different contingents were drawn. But the distances obelisk, recording the sale of four large parcels of landare considerable: Sippar was about 135 kilometers north- to that king (ca. 2275-60 B.c.) of the Akkadian dynasty.

145

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

100

80s

60

40-

20

........

/

/........... Thee-Oar and MuthanaS / [ Ur, Uruk]

---- Babil and QadisTya/ / i Babylon, Nippur]

S- BaghdadS // [ Sippar]

/

/ 00#e.000, 1 •l

4-10 11-17 18-24 25-30 1-8APRIL I

9-15 16-22 23-29 30-31 1-5 6-12 13-19 20-26MAY I JUNE

Fig. 26. Percentage of wheat-barley harvest completed at successive time intervals infive Iraqi provinces. From Department of Agricultural Statistics 1971, table 8: "Fre-quency Distribution of Harvesting Dates in the Selected Fields of Wheat and Barley,

Based on the Sample Surveys Conducted in 1970."

The text speaks in aggregate of 964 gurus, "men" whopresumably were the cultivators, who "eat bread" in aceremony formalizing the sale of 9,643 iku, amountingto some 3,402 hectares (Gelb 1976, p. 199). The parcelsthat were sold surely included lands in their alternateyear of fallow as well as the adjoining plots currently incultivation. Insofar as this is a representative sample,therefore, the harvest-labor text records a group capableof dealing with about 3.53 hectares per person and hencewith an area of about 770 square kilometers, althoughnot much more than half of it would have been undercultivation in any given year. Some supporting evidencethat a ratio of this magnitude is reasonably accurateis provided by the Iraqi agricultural and livestock censusof 1952-53. For the four provinces of the time that to-gether constituted the heart of ancient Sumer and Akkad(Baghdad, Hilla, Diwaniya, and Muntafiq) the averagenumber of cultivated hectares (including lands in fallow)per agricultural worker was 3.24 (Principal Bureau ofStatistics 1954, passim).SThe territory for which these harvest-labor contingents

had to assume responsibility thus was a substantial one.Some idea of its economic importance can be gained byconsidering dietary and crop-yield levels that were re-garded as standard during the Ur III period. The averagebarley productivity as reckoned by scribes was 30 gur(-lugal) per bur, or 1,133.7 liters (about 700 kg) per hec-tare (Maekawa 1974, pp. 10-11).3 The minimal yearlysubsistence allowance, reckoned at 2 gur (Jones andSnyder 1961, p. 286), amounted to some 480 liters (orabout 297 kg). Hence the crops harvested by these groups

"%would have been enough to provide for the basic sus-tenance of as many as 90,000 persons, equivalent tothe total population of several of the most substantialcities of the time. Yet we know that Nippur, althoughpresumably nearest at hand, was not among them, sinceit regularly received large deliveries of grain from othercities. The suspicion accordingly arises, although it can-not be confirmed with the available evidence, that thebarley produced in this manner was only partly intendedfor the immediate subsistence needs of the urban popula-tion. Some of the harvest might have been diverted in-stead to export, for example, although only the Elamiteplains around Susa and trading towns along the lowerGulf were within reach of economic shipborne transport.Beyond human consumption, however, lies the possi-bility that at least part of the barley may have been in-tended for animal fodder. This I will return to presently./ In addition to harvest labor, there were direct and sub-stantial bala contributions to the state from individualtowns. In the case of Girsu (modern Telloh), about whichwe know most, the total area cultivated was about a thirdof that collectively harvested on behalf of the state asrecorded in the text previously referred to. Half of theharvest was set aside to meet the costs of production, in-cluding rations for the agricultural workers and mainte-nance of the draft animals. The remainder was dividedequally between the state and the priesthood, with thelatter making provision for seeding, milling, and so on(Gregoire 1970, p. 233; Jones 1976, pp. 57, 60). In theabsence of any reference to the recruitment of harvestlaborers in Girsu for service elsewhere, it appears that in

146

0i

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

the southern cities like Girsu that were nearest to Ur theemphasis was placed instead on direct grain deliveriesof something on the order of one-fourth of the harvest.Perhaps the somewhat different emphasis in the centraland northern part of the alluvium was partly a device bywhich to reduce transport costs, or at least to shift theburden of meeting them onto the shoulders of the pro-Sducers. But it may also reflect the more sedentary, per-manently urbanized character of the inhabitants of thesouthermost districts.

The organization of agricultural work around thesouthern cities thus took a different form, although theheavy hand of state intervention is no less apparent.There are strong if indirect indications of fairly wealthy,powerful entrepreneurs, on the other hand, who assumedsome of the responsibility for fixed deliveries from large-scale cultivation against advances of laborers and supplies(Jones and Snyder 1961, pp. 270-71). Daily allowancesof wages, or less frequently rations, were the normal unitof account, and the figures are impressive even if thenumber of individuals simultaneously employed is usuallydifficult to establish. Texts from Girsu contain totals of43,204 and 27,589 man- and woman-days of labor, forexample, and a text from Umma dealing with womenalone accounts for 93,781 full days' wages (Fish 1953,p. 49; 1956, p. 8). As this indicates, women were widelyemployed in agricultural operations as well as in croptransport associated with the harvest, although apparentlynot in the threshing itself. Their presence in such numbersmay hint at a pervasive social difference from the northernpart of the alluvium, although this might also result onlyfrom the migratory aspect of the harvest labor thathappens to be recorded in the latter./ It is interesting to speculate on where these large num-bers of seasonal laborers could have come from. In asystem of such rigorous state control as is usually picturedfor southern Mesopotamia during the Ur III period, onewould assume that most individuals were locked intoposition with prescribed if seasonally changing duties andwith expectations of regular income from the state or aparticular temple. Perhaps this suggests that there wasa large, fairly fluid lower stratum of Sumerian urbansociety that otherwise receives little mention in templeor royal administrative records. Alternatively, we maycatch a glimpse here of the limited, seasonal interactionof the state system with semisedentary folk who other-wise normally remained outside the perimeters of culti-vation.4

What these texts dealing with cultivators suggest, inshort, is the existence of a complex, geographically dif-ferentiated, and extensive system of agricultural man-agement. Successive intervals of stable state control wereby no means identical, but in all of them there appearsto have been a fairly continuous band of cultivation thatvaried in width but extended down the center of the

alluvium for virtually its whole length, from Sippar tothe head of the Gulf. Marked regional interdependencewas a less constant feature, perhaps confined to the UrIII period. But at least at that time a disproportionatepart of the migrant harvest labor supply was recruitedfrom the generally more rural region of ancient Akkad,the upper part of the plain, and possibly from amongsemisedentary pastoralists not fully integrated within thestate system. By contrast, there was a striking lack ofparticipation by the inhabitants of the religious centerof Nippur, consistent with its privileged status in nu-merous other respects (Cassin et al. 1965, p. 142). Otherlarge southern cities, save perhaps Ur itself, received nosimilar dispensation. However, they seem to have beenallowed to focus their agricultural activities on theirimmediate hinterlands.

This must partly reflect the concentration of powerand wealth around the capital at Ur, in the extreme south.It implies sponsorship of a broad geographic division oflabor in which trade and manufacturing as well as ad-ministrative and religious activities were disproportion-ately concentrated in the southern cities. The extent ofcultivation around the latter thus ceased to be a functionsolely of their own immediate labor supply and foodrequirements, becoming partly also an index of the effec-tiveness of the regime in recruiting labor from remoterural areas. Against this background, cultivated areas andagricultural surpluses, like the population density incities, were increasingly and intimately dependent uponthe degree of political integration and the inward flowof resources that the center demanded from the kingdom'speripheries, rather than being relatively static reflectionsof the prevailing level of technology.>Similar interpretations emerge from a considerationof the state-managed component of the pastoral economyduring the Third Dynasty of Ur. Again the numbers areimpressively large, attesting to what Kraus has appro-priately called a "cortege ininterrompu" (1954, p. 528)of animals, principally sheep but also cattle in lessernumbers and occasionally even nondomesticated species,directed toward the larger temple establishments. Forexample, an aggregate total of almost 350,000 sheep andgoats and somewhat less than a tenth of this number ofcattle is recorded in one text, dating from the forty-eighthregnal year of Shulgi (2093-46 B.C.). Processed throughDrehem, most of them were apparently intended forsacrifice during the preceding forty-nine-month period(Calvot 1969, pp. 103, 108-9, 113). To judge from con-temporary Turkic nomads (Bates 1973, pp. 148-49),herds up to five times as large as the annual total ofalmost 85,000 sheep and goats per year would have beenneeded to sustain the flow of sacrificial animals alone.To be sure, by no means all of these animals necessarilywere maintained within the Mesopotamian alluvium. Itis known, in fact, that during the same period of Sulgi's

147

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

reign, several hundred fat-tailed or kungal sheep wereobtained as "booty" from the land of Martu northeastof the Tigris, and others were shipped in the oppositedirection (Lieberman 1968-69, p. 58). But there can beno question of the presence of very large, state-maintainedherds both within the intensively irrigated belt along thecentral Euphrates branches and on the open steppelandssurrounding it.\., The state's operations in connection with the produc-tion of textiles from the wool of its own herds were evenmore impressive. Jacobsen's fundamental study of theroyal Wool Office speaks of an establishment chargedwith custody over some 6,435 tons of raw wool andemploying as many as 9,000 state-owned male and femaleslaves. Its elaborate organization and record-keeping arenot directly relevant to this study, but the overall scaleis further suggested by the figure of about 2,000 tons ofnew wool coming in largely if not exclusively from theplucking of the royal herds (1953, pp. 172-74, 178).Figures are also available on the average annual yieldper sheep, ranging from almost exactly a kilogram forthe uli-gi variety that predominated around Drehemdownward to about 0.707 kilogram of better-quality woolfrom the fat-tailed variety kept around Lagash (Waetzoldt1972, pp. 5-6). If one assumes an average for all varietiesof 0.85 kilogram, herds totaling more than 2,350,000animals would have been necessary to provide the wool.These herds must have overlapped to some extent withthe herds kept to supply meat and sacrificial animals, butit should be noted that several additional varieties ofsheep are known that are seldom mentioned in connectionwith wool and so must have been kept primarily forthese latter purposes.

As I noted earlier, not all of this immense number ofsheep and goats at the disposal of the crown were de-pendent on pasturage and cultivated fodder from theMesopotamian alluvium. Those designated as "highland"presumably took advantage of the richer grasslands inRowton's "dimorphic zone" east of the Tigris. But suchconsiderations as security, increased transport cost, andthe barriers to efficient, centralized management arisingfrom poor communications would have at least partlycounterbalanced the attractions of underutilized pastur-age there, in the calculus of a state bureaucracy intenselypreoccupied with routinizing operations. Waetzoldt hasshown, on the basis of recorded daily rates of plucking,that more than 200,000 sheep may have been processedeach year at Girsu alone. Still other texts from Ur thatrecord the receipt of new wool from fat-tailed sheep, theLagash (and hence Girsu) variety, may attest the presenceof as many as 500,000 sheep and goats in the Lagash area(Waetzoldt 1972, p. 14). This suggests that most of theroyal herds were kept considerably closer at hand thanin the natural grasslands of the Zagros foothills. Textsrecording the feeding of barley to as many as 52,553

stalled sheep (and 1,522 cattle) at Girsu over a three-month period probably do not fully illustrate the extentof the practice even in that one center, and they certainlyreinforce the conclusion that the distribution of the herdswas heavily clustered within range of easy seasonal move-ment into the main settled areas (Schneider 1927).5

Returning once more to the Iraqi agricultural census of1952-53, it is instructive to consider the number of sheepand goats then held within the four provinces more orless corresponding to the ancient heartland of alluvialsettlement. At that time the total was 1,536,752, includ-ing just under 90 percent sheep and the remainder goats(Principal Bureau of Statistics 1954, passim). Even with-out considering herds kept especially for meat and sacri-fice, the total during the Third Dynasty of Ur as hypoth-esized above was 53 percent greater. Making generousallowance for segments of the royal herds kept across theTigris and elsewhere, does this imply an Ur III magnitudeof land use very little different from that of modern Iraq?

To reach such a conclusion would require us to ignorethe additional, surely very substantial number of animalsthat were privately or communally held by villagers andsemisedentary folk. Their presence is only indirectly at-tested in state and temple records, for example by ex-votoofferings, but the patchwork of stubble and gleaningsand other localized sources of fodder in rural areas af-forded an ecological niche that only small herds in frag-mented ownership could have filled, just as they do today.

o a comparison of ancient with modern conditions mustVbegin with the recognition that the ancient population of

sheep and goats was considerably larger. Nor are thedifferences limited to this contrast alone.

The cultivated area in these four provinces at the timeof the census was 16,800 square kilometers. This sub-stantially coincides with the area blocked out by the linesof ancient settlement in figure 31, but the modern totalmakes no allowance for the many large areas of swampand steppe that can be seen to have lain within the ancientperimeters. Also, the population in 1952-53 was reportedto have included more than 518,000 men actively en-gaged in cultivating within these provinces, surely ag-gregating in the neighborhood of 2,000,000 persons withtheir families. To this total must be added the nonagricul-tural urban population of those engaged in secondaryand tertiary occupations, not to speak of the considerabledrain upon the countryside represented by the absenteelandlords of the time with their retinues. It seems prob-able, in short, that no less than 2.5 million persons weresustained by the agricultural activity carried on withinthis area in the early 1950s.

No accurate estimate of the Ur III counterpart of thisfigure is possible, since only part of the total region hasbeen intensively surveyed. But a reasonable extrapolationfrom the part that is known-even taking the Ur IIIfigure of 2,725 hectares of settlement given in table 13

148

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

at full face value, without regard to considerations men-tioned earlier suggesting that it is probably too high-isalmost certainly less than three times this number, multi-plied by an assumed density of 125 persons per hectare.The maximum population of the alluvium during theThird Dynasty of Ur was probably closer to a half millionthan to a million, in other words, and thus was less thana third of its modern counterpart. Seen in this light, thesubstantially reduced number of sheep and goats kept inmodern times is even more sharply in contrast with theancient pattern.

It seems inescapable that the ratio of sheep and goatsto the human population was about four times greaterduring the Ur III period than it is today. This suprisingconclusion has a number of further ramifications. Tobegin with, the contemporary pattern in which the bulkof the herds depend in the main on pasturage obtainedwithin the perimeters of cultivation would not have beenpossible. Most of the herds instead must have spent mostof the time on outlying steppelands and seasonallywatered depressions, and the numbers are so large in re-lation to the subsistence potentialities of these types ofterrain that many of the pastures were remote from, thebroad band of settlement and cultivation running downthe center of the alluvium. An entirely different class ofsettlements thus was made necessary, as yet unattestedin either the archaeological or the textual record, toprovide temporary to semipermanent shelter for theherdsmen and perhaps their families.

It should not be assumed that these outlying regionswere given over exclusively to herding, though this wassurely the dominant economic activity. Nothing could bemore natural than that some cultivation was added,wherever it was favored by a local source of irrigationwater. Even minor cultivated plots would have improvedthe diet of specialized herdsmen and reduced the fre-quency of their visits to the distant towns along the majorriverine arteries. Moreover, a local barley crop wouldprovide supplementary fodder for the herds, permittingthem to be enlarged without needing to move more often.A kind of semisedentary society thus would have beenencouraged. It was organized around a few centers ofadministration and distribution, but on the whole it wascomposed of communities of very modest size and dura-tion, thinly strung out along watercourses that moreoften were parts of a naturally anastomosing network ofminor Euphrates offshoots than integrally planned com-ponents of an irrigation system. Something of this kindis suggested by an Ur III geographic text dealing witha district well to the west of the major line of Sumeriancities, centering on the towns of Kazallu and Marad(Kraus 1955). Marad, incidentally, was among the townsthat furnished a contingent of harvest laborers in thetext dealt with earlier, sending a substantial group of1,510 that probably was recruited from the entire sur-

rounding district rather than from that fairly modestsettlement alone.

There is a further ramification to these outlying fringezones given over to predominantly pastoral semisedentaryfolk. Their initial formation must have been encouragedby the Ur III rulers in order to give proper care to theimmense royal flocks of sheep and goats. As long asconditions of assured central control and political stabilitycontinued, herd maintenance offered no especially dan-gerous challenges and was merely one of a number ofspecialized activities carried on partly or largely in sup-port of the state. To be sure, the size of the herds de-manded an expansion in the pastoral component of thesociety, with several tens of thousands of individuals andtheir families assigned primarily to this service. As thepowers of the ruling dynasty inevitably began to erodein conformity with the largely cyclical pattern suggestedearlier, however, the attachment of outlying semiseden-tary elements would have been the first to loosen. Theirlargely pastoral basis conferred mobility, and mobilityin turn conferred a greater opportunity either to shiftloyalties to rival powers or to withhold support andtend toward greater degrees of independence and autarky..In other words, the formation of these impressive royal'herds carried within it the seeds of a far-reaching dis-solution of the web of political and economic interrela-tionships, once the initial organizing impulse had run itscourse.

We must also consider the meaning of the high ratioof the animal population to the human population forthe subsistence base and major economic orientation ofthe metropolitan centers of the empire. Similar recordsare not available for other periods, but at least duringthe Third Dynasty of Ur there was clearly an exceptionalroyal emphasis on the production of wool and on thedevelopment of a large-scale textile industry. Woolentextiles, we may assume, were regarded as the basis forfar-flung commercial relations with regions whose naturalresources Mesopotamia altogether lacked. The well-beingof the flocks was a major concern of state policy, there-fore, and is likely to have exercised an influence on otheraspects of agricultural policy, including crop preferences.y A profound change in crop preferences has been sug-gested on paleobotanical as well as textual grounds. Thereis some slight evidence for an almost equal balance ofwheat and barley in the mid-fourth millennium (Jacob-sen 1958, p. 50). By the end of the third millennium,southern Mesopotamia had unquestionably shifted toan overwhelming reliance on barley. Barley is more salt-tolerant, and it has plausibly been argued that the shiftwas primarily a consequence of ongoing processes ofsoil salinization that accompanied widespread irrigationagriculture (Jacobsen 1958, pp. 11-13; Jacobsen andAdams 1958, p. 1252). But barley is also the preeminentfodder crop for sheep, and both the size of the Ur III

149

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

herds and some direct testimony as to stall-feeding sug-gests the possibility that considerable barley may havebeen cultivated expressly for their maintenance. Withoutdenying that salinization may also have played a sig-nificant part, we can thus suppose that the dominanceof barley was partly a reflection of the commercial aspira-tions of the crown. The latter depended, after all, onmaximizing the production of woolen textiles as the onlyvaluable, lightweight, marketable commodity availablein the kingdom's heartland with which to meet the needsof long-distance exchange.

This should not be taken to imply that royal policyfavored the expansion of flocks while in any way opposingthe growth of the human population. Mesopotamia wasunderpopulated in human terms as well, at least relativeto the subsistence potentialities of irrigation agricultureduring periods of stable, centralized authority. Hence thegrowth of the human population, and a concomitant ex-tension of settlement and irrigation, was surely also adesideratum of state policy. It is known that Ur III rulerson occasion founded new towns, including at least one inthe vicinity of Nippur, forcibly drawing in the conqueredpopulations of more distant regions for this purpose.Agricultural pursuits and textile manufacture are specifi-cally mentioned among the activities to which these set-tlements of prisoners were devoted (Gelb 1973, pp. 76,82). The Third Dynasty of Ur foreshadowed the policiesof the later Assyrian rulers in this respect, althoughmarkedly less persistently and on a less grandiose scale.But within the span of only a century or so before Ur'sdynastic control began to crumble, its resettlement poli-cies had not materially altered a balance weighted pre-ponderantly toward flocks.' As briefly outlined above, the Third Dynasty of Urprovides a kind of paradigmatic model of maximizationin Mesopotamian settlement and agriculture. Of course,state policies could proceed in the directions outlinedonly within the technological constraints of the late thirdmillennium. Those policies were maintained with fairlyconsistent force and direction, moreover, only within thecentury or so of the dynasty's floruit. But subject to thesequalifications, the economic achievements of the timesuggest an ideal-typical model toward which other strongdynasties must have repeatedly sought to direct theirenergies.

This does not imply that the institutional patterns ofthe Ur III period were more or less consciously replicatedat other times. In fact, the plane of abstraction main-tained here has largely ignored specific institutional fea-tures. Only the existence of an absolutist state has beenassumed, subject to more or less elastic principles ofdynastic succession and to the inability of any dynastyto stabilize for long either its external frontiers or its in-ternal authority.

Most of the discussion of early Mesopotamian history

has been conducted on a different plane of abstraction,concerned with the administrative and juridical particu-lars of a succession of societal and institutional formsthat appear to have been more or less equally consistentwith a larger framework of shifting dynastic authority.It has been traditionally maintained that the Tempel-

' wirtschaft of Early Dynastic times (Deimel 1931; Falken-stein 1954) gave way to the state economy of the ThirdDynasty of Ur (Kraus 1954), and that the latter afterwardslowly came to terms with increasing aggregations ofprivate wealth (Koschaker 1942). Diakonoff (1954) wasthe first to modify this view, showing that substantialcommunal holdings as well as large private estates existedalongside the lands administered by temples in the EarlyDynastic period. Additional study of early land-sale rec-ords confirms not only the importance of private hold-ings but the diversity of professions represented amongthe sellers and purchasers. In retrospect, as Gelb (1969,pp. 139, 145) has argued, a sweepingly overgeneralizedpicture was reconstructed on the basis of a single archive.What faces the specialist now is the need somehow tostrike a balance between accidents of discovery, differentgenres of material, and the few, generally ambiguousleads as to the relative strength or status of individualsand institutions. And at least equally demanding is thetask of reconstructing a picture that clearly is no longeruniform but instead must take account of much local aswell as temporal variability.

There is an identical need for the later third millennium.Arguments for a state economy were based largely on theabsence of evidence for the private sale of land in archivesthat happened to deal largely or exclusively with stateand temple activities. More recently, contemporary textsfrom Nippur that instead focus on private activities havebegun to yield abundant evidence of a wide range ofoperations consistent only with private ownership of landin title as well as fact. To be sure, documents of sale arestill not unambiguously attested. But Gelb has cogentlyargued that the available data are more consistent withan interpretation that the sale of property had beenformally prohibited than with the absence of the institu-tion of private land ownership (1969, pp. 146-51). Thereare even indications that kin-based or territorially basedcorporate groups continued through at least the OldBabylonian period as part of a mixed pattern of land-holding (Yoffee 1977, p. 145) and in fact extended intothe early first millennium (J. A. Brinkman, pers. comm.).What this suggests, as Maekawa (1973-74, p. 142) hasnoted, is that attempts to trace a line of essentiallyunilineal development (e.g., Diakonoff 1965; Adams1966) generally overlook highly significant reversals,brakings, restorations, internal contradictions, and localdifferences in their search for a sweeping simplicity. Butthe even more important point is that a basic, continu-ously shifting pattern of economic differentiation and

150

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

centralization can be recognized behind the superstruc-tural features and changes that have received greatestattention from the authorities in the field.

Among this underlying pattern's most salient featureswas the encouragement of an increasein population andan extension of settlement, increasing the human re-sources of the state vis-a-vis its competitors and enlargingits income from growing municipal and institutionaltransfers or taxes. Some natural demographic growthprobably would result from the security afforded by astrong dynasty, but forced resettlement policies were alsobrought into play. A hierarchical, centralized structureis apparent in the economic spheres we have considered.This corresponds only in part with the development of ahighly articulated urban hierarchy, for their advantagesas defensive nuclei led to the differentially greater sur-vival of large urban centers even under conditions ofpolitical disintegration. While there was evidently con-siderable expansion in the urban population, therefore,the more significant shift in settlement was toward agreatly expanded number of smaller villages along thedendritic elements of new (or reconstructed) irrigationsystems. Like the canals themselves, many of these com-munities would remain viable only as long as the statewas able to provide an outer envelope of security forthem.

Accompanying the political and administrative cen-tralization was an intensification of the dichotomy be-tween an imperial core and its peripheries. The alluvialMesopotamian plain in a purely spatial sense appears toconstitute the central geographic region, but it was never-theless sharply differentiated in functional and subsistenceterms. Pastoral and semisedentary groups were allowedand even encouraged to occupy great zones of steppe andseasonal swamp outside the perimeters of cultivation, inorder to maximize income from royal herds. Controlledby a painstakingly recorded chain of command so longas dynastic authority remained firm, these groups wereof course especially prone to behave more independentlyunder less rigidly authoritarian conditions., Also under close royal supervision, and dependent ona continuing inward flow of resources from more periph-eral components of the economy, was a manufacturingsector of impressive size and internal complexity. Sec-ondary and tertiary occupations and professions werecorrespondingly numerous, including a substantial bu-reaucracy relying on exhaustively routinized accountingprocedures. Patterns of final consumption are much lessclear in existing documents, which concentrate on pat-terns of production and collection in central depots andattest to only the first links in the chains of dispersal, butclearly the system afforded the capability to sustain asubstantial, even a preponderant, part of the adult, work-ing population in the larger southern urban centers innonsubsistence activities. And the system was also able

to generate a large supply of textiles, other craft products,and agricultural surpluses to employ in commercial ven-tures well beyond its own frontiers.

The human costs with which all this was accomplishedwere doubtless very heavy, but they are difficult to specifysince they are generally ignored in the existing sources.Involuntary labor and forced transfers of agriculturaland other surpluses must have been particularly onerousfor primary agricultural producers, slaves in state fa-cilities, and the lower levels of the social hierarchy moregenerally. Proportionately the largest part of the society,these groups had least to gain from the superimpositionof dynastic authority. They benefited from the militarysecurity, to be sure, but increasing exactions probablyoffset much of this advantage. Hence it may well havebeen the massive human costs that in the end primarilyaccounted for the prevailing brittleness of periods of con-solidation like the Third Dynasty of Ur.

There is a final respect in which policies of maximiza-tion apparently led to the gradual emergence of theirantithesis. I noted earlier that late Early Dynastic cropyields averaged 2,030 liters per hectare, whereas underthe Third Dynasty of Ur that impressively high figure fellsharply to 1,134 liters. In the meantime, seeding rates hadhad to climb just as steeply. "The rate most commonand used over the largest area" in the Ur III period was55.5 liters of seed per hectare (Jacobsen 1958, p. 63),more than twice the average rate at the end of the EarlyDynastic period (see chap. 3, n. 8, and p. 87). As already

\observed in connection with shifts in crop preference,salinization seems to have been a major contributingfactor in this ominous decline-in rate of return on seedeven more than in output per unit of land area. Butsalinization is not an independent variable that is merelytriggered by irrigation agriculture. Its onset and effectsare inextricably intertwined with the intensity of landuse and the irrigation practices that are followed. HenceI must mention once again the prodigious growth ofpopulation between Early Dynastic and Ur III times.Table 13 indicates a 64 percent increase in site area andimplicitly in population, although part of this may stem

\ from the conflation of Ur III and Isin-Larsa settlementpatterns. Much of such an increase must have dependedon extensions in the irrigation system and on an en-largement and stabilization of the supplies it could deliverto the cultivated areas it served.

Quite possibly the availability of water advanced tothe point where in certain districts land rather than waterplaced the critical upper limit on production. In thosecircumstances there would have been a heavy inducementto maximize short-term output by cultivating the samefields every year. Widespread violations of the system ofalternate years in fallow would have further intensifiedthe salinity problem by doubling the rate of applicationof irrigation water, hastening the rise of saline ground-

15i

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

water into the root zone. Probably the cultivators of thetime perceived only imperfectly the relationship betweenoverirrigation and the long-term loss of output to salinity.In any case, uncertainties as to the adequacy of the futurestream discharge to meet the needs of the current harvesthave always supported a practice of overirrigation. Inthe sequel, therefore, the Ur III agricultural regime pro-vided both the inducement and the means to follow irri-gation practices leading to an enormous loss of agricul-tural output.

Note that these involutionary processes tended to becumulative in their effects. Badly salinized lands fre-quently must have gone out of production almost in-definitely, for even in modern times their reclamationdepends on slow, expensive, carefully controlled methodsof deep drainage and flushing. Loss of formerly cultivatedareas intensified the pressure on initially less affected dis-tricts and thus must have extended the problem. Seriousas was the decline in average yields by the Ur III period,it subsequently went on to become almost catastrophic.Jacobsen has shown that by 1700 B.C., shortly before anextensive abandonment of southern Babylonia, yieldsaround ancient Larsa had slipped to a mere 718 litersper hectare. Worse still, more than one-fourth of thearea then in cultivation seemingly was being kept inproduction even though yields were only 370 liters (about228 kg) per hectare (Jacobsen 1958, pp. 39-40). Sincelabor inputs were relatively inelastic, this represents lessthan a fifth of the expected yields eight hundred yearsearlier for a roughly similar magnitude of effort. Theburden on the cultivator had become a crushing one.

Thus long-term agricultural decline was in some waysa direct consequence of its earlier apparent "success." Aswith the ineluctable political processes contributing theearly demise of seemingly highly successful regimes likethe Third Dynasty of Ur (above, pp. 132-33), this high-lights the linkage between expansionist policies and en-suing collapse. Important features of both are to beunderstood only as parts of a single, long-term process.To that end, it may be useful to consider the nadir ofsettled life in the early first millennium B.C. as perhapsthe sharpest imaginable contrast with the Ur III period.No comparable degree of detail is possible, since it isone of the characteristics of the time that textual sourceswere extremely impoverished. Administrative activitymust have been at a very low level, apart from corre-spondence relating to the largely hostile and acquisitiveinterests of the Assyrians, and even archaeological testi-mony is very limited in extent. The contrast is thus par-ticularly sharp with the Ur III period, the most volu-minously documented of all from a textual standpoint.But enough is known to outline at least some of the genericfeatures of the countryside. Together they suggest anopposite extreme, or at any rate a strikingly differentparadigm, toward which the repeated oscillations away

from a condition of economic and political integrationseem to have tended.

Figure 25 and table 13 illustrate the basic conditionsof settlement in Middle Babylonian times, roughly theend of the second millennium B.C. and the first threecenturies or so of the first. There had been more than a40 percent reduction in the number of sites, and a 77percent reduction in the aggregate occupied area, sincethe end of the Ur III period about a millennium earlier.Major urban centers had disappeared almost completelywithin the intensively surveyed region, although Babylon,the capital, was probably still of considerable size. Muchof the population away from the district around Babylonmay not have been sedentary enough to leave substantialarchaeological traces, but almost two-thirds of the nu-cleated site area that presumably accommodated thefully sedentary component is composed of small villagesand towns occupying 10 hectares or less. Accompanyingthis retrenchment, and making an analysis of it vastlymore difficult, was an even more precipitate reductionin textual documentation. Accidents of discovery makethe significance of direct comparisons somewhat ques-tionable, but there are approximately seventy-five timesas many Cassite texts as the 160 or so that are known forthe post-Cassite or Middle Babylonian period (J.A. Brink-man, pers. comm.).

A decline of this magnitude cannot be thought of as asmooth, featureless withering away. Representing an anti-thesis to the unprecedented density and internal articula-tion achieved under the Third Dynasty of Ur, the newconditions must have led to a sharp deterioration in thefabric of urban as well as rural life. They even involve asteep retrogression from the preceding Cassite period,itself an interval of political retrenchment and demo-graphic decline. Only a marginally greater area of totalsettlement appears to have been retained within the in-tensively surveyed region than in the Uruk period, twoand a half millennia earlier, while already then, so soonafter the very outset of settled life, the urbanized propor-tion of the population-not to speak of the vigor ofcultural development-was appreciably greater.

With due allowances for fragmentary data, the socialmilieu of the time has been carefully pieced together byJ. A. Brinkman (1968). He notes that urban life and cul-ture continued, albeit on a declining scale and repeatedlysubject to disastrous interruptions. The indigenous Baby-lonian population was concentrated in and immediatelyaround the miajor towns, perhaps held in place by tena-cious religious traditions and by the growing economicstrength of temple complexes after an earlier nadir ofcorporate activity. Sargon's propagandistic inscriptionsspeak of freeing urban hostages from Chaldean detain-ment, and more generally of the pro-Assyrian loyalties ofthe citizenry of the larger towns. Largely on this basis,some have identified a shared Babylonian-Assyrian re-

152

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

ligious tradition and old cultural ties as a continuing his-torical force (Dietrich 1970, p. 5). Such loyalties mayhave played a part, and certainly the Assyrians lost noopportunity to reaffirm publicly their respect and supportfor the urban and religious institutions of their southernneighbors. But, as Brinkman shows, Babylonian urbanallegiance to Assyria was at best qualified and subject toquick reversal. By the Assyrians' own accounts, the in-habitants of many Babylonian cities at least at timesactively collaborated in armed resistance to the invadingforces. Moreover, "there is no record of 'native Baby-lonians' or their cities revolting against Chaldean leaders,who must have caused considerable disruption by draw-ing down the frequent wrath of Assyria in the late eighthand seventh centuries" (Brinkman 1977, p. 315). At most,therefore, the relationship was one of shifting crosscur-rents and conflicting loyalties. Perhaps the relatively lesscontinuous and active opposition of the cities was largelya reflection of the fact that their citizenry were moreimmediately exposed to Assyrian retribution.

Assyrian kings boast of having reestablished urbanlands and privileges, only to complain later that manyof the same cities they had favored were actively support-ing their former oppressors. Before the direct assumptionof Assyrian imperial control, at least some acts of militaryintervention seem to have been intended primarily to pro-vide assistance to the Babylonian king in his own un-availing attempts to impose order on the countryside.Adding to the complexity of the situation is that moreconsistent support for growing Assyrian suzerainty wasevinced by southern than by northern Babylonian cities.Already for several centuries, the former had been littlemore than island enclaves in a Chaldean sea. Weakerurban adherence to Assyria in the north, by contrast,might have been an outgrowth of the more routine pres-ence of Assyrians in administrative capacities there havingbrought palpably greater costs and fewer benefits. RuralBabylonian dependencies in the northern countrysidegenerally seem to have remained under Assyrian controleven when the cities there were in revolt (Brinkman 1965,p. 243; 1968, p. 229; 1969, pp. 346-47; pers. comm.).All this suggests that pro- and anti-Assyrian attitudeswere probably less independently determinative than theywould have appeared to the Assyrians. To some degree,positions on this issue appear to have been influencedby local cleavages between the towns and the country-side that sometimes extended into the towns them-selves.

Our difficulty in weighing these possibilities, of course,is that so much of the available documentation stemseither from the Assyrian state annals and correspondenceor from the appeals of Babylonian partisans to what thelatter must have hoped were Assyria's decisive interests. Aspokesman for Nippur, for example, urgently requeststhe Assyrian king's assistance in the following terms:

The king knows well that people hate us everywhere onaccount of our allegiance to Assyria. We are not safe any-where; wherever we might go we would be killed. Peoplesay: "Why did you submit to Assyria?" We have nowlocked our gates tight and do not even go out of towninto the . . . We are (still) doing our duty for the king;the envoy and the officials whom the king has sent herehave seen all this and can tell the king about it. But theking must not abandon us to the others! We have no waterand are in danger of dying for lack of water. The king,your father, wanted to give us the water-rights for theBanitu-canal under this condition: "Dig an outlet fromthe Banitu-canal toward Nippur." [The ... ], however, re-fused us the water. The king should now send an order toUbar, the commander of Babylon to grant us an outletfrom the Banitu-canal so that we can drink water withthem from it and not have to desert the king on accountof lack of water. They must not say everywhere: "Theseare the inhabitants of Nippur who submitted to Assyria-and (when) they became sick and tired of the lack of water(they deserted)." [Oppenheim 1967, p. 175 (ABL 327)]

To what extent can we conclude, from self-interestedtestimony of this kind, that these protestations of Assyr-ian loyalty were genuine? Alternatively, the reality mayhave been that Nippur was indeed cut off and beleagueredin the midth of a hostile countryside, unable to mustersufficient forces to arrest the ruination of its own agricul-ture, and so pragmatically sought the assistance of itsonly potential ally. Essentially the same submission wouldhave been made to the Assyrian overlord in eithercircumstance.

But let us turn from the cities, obviously circumscribedand largely powerless enclaves even though they naturallydominated in the loyalties and attention of the local scribalelement. The countryside was largely in tribal hands, be-yond the reach of urban administrators. Aramean-speak-ing groups, their members characterized by a gentilic ad-jective, had been longest in place but were on the wholemost fragmented and least sedentary. There were thirty-six named tribes among them, some under the simul-taneous leadership of as many as six shaykhs (nasiku).The Chaldeans, although more recent arrivals, were morecentralized and hierarchically structured. There were onlyfive relatively larger groupings, together constituting "thereal strength in the land" virtually until the fall of Ninevah(Brinkman 1968, p. 318). Members of each Chaldean"house" claimed tribal affiliation by tracing their descentfrom its eponymous ancestor.

In addition, the Chaldeans in general seem to have beenwealthier, more apt to be settled in their own fortifiedcities, taking an interest in trade, growing date palms, andplaying an active role in the political life of Babylonia;also many Chaldeans, expecially of the ruling families,took Babylonian personal names. The Arameans, on theother hand, seem rarely to have resided in large cities of

153

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

their own, appear more often as members of raiding par-ties, seldom had Babylonian names, and were not activelyinvolved in Babylonian politics. [Brinkman 1977, p. 307]

It is difficult to establish the source or signifiance ofthese differences when so much of our present evidenceconsists of records intended for Assyrian use. Tribalgroupings held a shifting mosaic of lands that interpene-trated with those of the Babylonians. The Chaldeans, inparticular, maintained fairly continuous control of thesouthern Babylonian swamplands that were more remotefrom the Assyrians and more difficult to subdue militar-ily. The Bit Yakin tribe of the Chaldeans, for a time ac-tively in alliance with Elam and the spirit of resistance toAssyria, had its strongholds there. Already by the mid-ninth century its shaykh was termed a "king" by theAssyrians. By the early eighth century the Assyrians spokeof the "kings of Chaldea" collectively, but to Tiglath-Pilesar III a half-century later there were only "headmen."Succeeding variations in usage like these may attest eitherto the fluidity of local patterns of leadership or toAssyrian scribal uncertainties about unfamiliar customsand terms.

The Babylonians, Chaldeans, and Arameans clearlycannot be arranged along a smooth folk-urban continuum.There were sharp disjunctions, different directions ofspecialization, and probably historical reversals of direc-tion as well. In general, however, the widest gulf seems tohave been that between townsmen and countrymen. Whatdistinguished the latter most visibly was that they tendedto be tribally organized, and, as Morton Fried has per-suasively argued (1968), tribal organization is perhaps inalmost all circumstances to be understood as an outcomeof the requirements of interaction with politically moredeveloped neighbors. 6 At least for purposes of this dis-cussion, similarities in the structural positions of theChaldeans and Arameans outweigh their apparently dif-ferent social bases and economic orientations. Both werelargely nominal subjects of the Babylonian kings. Bothwere active in the resistance to Assyrian overlordship,with the Chaldeans in particular suffering massive lossesof exiled population as a consequence. Inspite of this,both were able for long periods to maintain considerablede facto internal autonomy.

The consistent resistance of the tribally organized partof the population to Assyrian pressure is striking. As a re-sult of it, according to Assyrian claims, a total of morethan 450,000 persons were forced into exile over littlemore than a forty-year period in the latter part of theeighth century. Even allowing for some duplication and avery large element of exaggeration, this surely testifiesto the massiveness of an assault that was directed pri-marily against the Chaldean and Aramean countryside.According to Brinkman (1979, p. 235) it was "awesomelyeffective" in destroying the Bit Yakin, initially the spear-

head of resistance, as a military force large enough to bereckoned with. Yet the opposition of other, originallysmaller groupings did not thereby slacken.

This resilence seems in partial contradiction to therelatively low population levels recorded for the inten-sively surveyed area, and it perhaps should serve as awarning against generalizations based exclusively on thelatter. Part of the loss of population that is archaeologi-cally attested even before the major Assyrian onslaughtmay have been more apparent than real, with most of thetribal elements occupying small, shifting settlements thatproduced very shallow accumulations of debris-and thathence easily elude archaeological detection. But even moreimportant, the identification of the Bit Yakin with aswampy refuge and their close association with the Elam-ites suggests that they may have settled primarily in aregion well to the east and southeast of the Babyloniancities along old Euphrates levees. Quite possibly this indi-cates a gradual retreat of the Gulf shoreline, creating anempty niche into which newly arriving tribesmen couldreadily filter. In these circumstances, the formation of akind of no-man's-land within much of the area that here-tofore has been archaeologically surveyed could well bea somewhat misleading indication of population trendsfor the Mesopotamian plain as a whole. A sufficientlylarge proportion of the alluvium has already been studiedwith the absolutely consistent finding of a drastic decline,however, to indicate that remaining regions are mostlikely to provide for more than a fairly modest reductionin the steep and widely prevailing loss.

In its general outlines, the picture outlined above isstrigingly similar to conditions obtaining in southern Iraqduring the last centuries of Ottoman rule (Adams and Nis-sen 1972, chap. 5) and to what little is known of earlierchaotic interludes like the Gutian period. Hence it is thegeneric features of political instability-the decline of irri-gation agriculture, urban-rural polarization, and theheightened influence of tribally organized seminomadicelements-that must be seen as the cyclically opposedcounterpart of periods of strong dynastic consolidation.The interplay of specific ethnic groups and cultural loyal-ties is not thereby made irrelevant. It may help to explain,for example, why tendencies toward integration as well asdisintegration did not produce entirely uniform configura-tions with each rising and falling dynasty (Adams 1978).But the general pattern is most significant for an under-standing of gross changes in settlement and land use.This is particularly true when so much of the evidencefor those gross changes has had to be seen through thescreen of roughly three-century spans into which theceramic indicators were classified during archaeologicalsurface reconnaissance. Conceding the historical impre-cision of any schema that groups distinctive periods intocontrastive categories, the oscillations in agricultural andsettlement patterns for which evidence has been presented

154

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

in this chapter can therefore be viewed as forming parts ofa grand, consistent pattern.

It is also true that during the course of the third, sec-ond, and early first millennia there were significant devel-opments, whether abrupt or slow and cumulative, thatcannot be subsumed within a pattern of oscillation be-tween centralizing and disintegrating extremes. Empha-sizing once more that this is primarily an account of landuse and settlement, not of cultural or political history, Imust now consider the most salient of these changeswithin the more particularistic or narrative framework ofthe succession of historic periods.

THE SHIFTING NETWORK OF WATERCOURSESAND SETTLEMENTS

An overview of the main features of early historic set-tlement in the Mesopotamian plain, from Early Dynasticthrough Middle Babylonian times, is provided by figures27 and 28. The first of these maps includes the Early Dy-nastic I period, thus maintaining a degree of overlap withfigure 9 in which pre- and protohistoric sites concludingwith the Early Dynastic I period were illustrated at thesame scale. Similarly, settlements of the Third Dynasty ofUr and the Isin-Larsa periods (which it has not been possi-ble to separate, as I noted above) form the concluding andbeginning phases respectively in figures 27 and 28. It willbe noted, however, that the second of these maps includesa somewhat larger region at a correspondingly smallerscale. This permits the area around ancient Ur to be in-cluded, as separately described in an appendix to this bookby Henry T. Wright on the basis of his 1966 survey. Alsoshown only in figure 28 are the major centers in the ancientkingdom of Lagash whose locations are known-Girsu,Nina, and Lagash itself. Figure 28 is an essentially com-plete map, in other words, of all known Sumero-Akkadiansites on the alluvial plain between the Tigris and Euphra-tes from Ur III times onward.

A greater density of sites is apparent in the southeasternhalf of both maps, the region of more intensive archaeo-logical survey. The average interval between sites is, cor-respondingly, several times greater in ancient Akkad, theregion to the northwest. However, this is not a genuineregional contrast. It partly stems from differences in recon-naissance methodology, as well as reflecting the greaterdepth of alluvial deposition and the much more extensivedisturbances resulting from modern cultivation that arefound in the upper part of the plain. The relatively lim-ited effect of the latter two factors on the lower plain hasencouraged more detailed treatment of ancient water-courses there, as described more fully in chapter 2.

Relatively straight dashed lines between sites, to benoted primarily in Akkad, are no more than generalizedsuggestions of the paths the major watercourses of thetime may have taken. In Sumer, on the other hand, it has

been possible in many cases to trace actual paths, completewith meanders and meander cutoffs, with the aid of airphotographs. Once again, this contrast must not be re-garded as indicating a genuine regional difference. There isnothing to imply that straight, essentially artificial canalswere characteristic of the upper part of the plain, or thatstate engineers and cultivators in the lower part were con-tent with more "natural" river regimes. If anything, theopposite is likely. Straightening and diking were probablycommoner in the more urbanized, and almost certainlymore densely populated, southeastern region of ancientSumer than in Akkad.

The main modern branches of the Euphrates, shown inboth figures, are far to the west of their ancient counter-parts. Parts of the modern river regime are so linear as toimply that they follow earlier courses laid out for artificialcanals, as for example northwest of ancient Borsippa. Butthe irregularity of most of the rest of the system is strik-ing. There are abrupt changes in direction even apart fromnumerous meanders, and the prevailing pattern (some-what simplified in these maps) is one of repeatedly bifur-cating and rejoining channels. Some areas are inevitablymore suitable than others for irrigation and settlementwithin a prevailingly "natural" riverine system of thiskind. The modern population, therefore, tends to begrouped in irregular clusters, especially where local con-ditions permit dendritic systems of canals to fan out intoadjacent hinterlands, separated by thinly populatedreaches of swamp or by uncultivated steppe.

Essentially the same pattern will be observed in thebetter-known, southeastern part of the ancient series ofEuphrates branches. Continuous, fairly regular distribu-tions of contemporary archaeological sites can be fol-lowed along a few channels there, but it is more commonfor groups of sites to be interspersed with areas apparentlylacking permanent settlement. Hence there is almost al-ways some degree of ambiguity as to which apparent seg-ment of a line of settlement along an ancient watercoursewas connected with which other upstream or downstreamsegment. During the periods dealt with in this chapter, inother words, the uniformity that might be expected of theancient riverine system and adjoining chains of settlementas a result of comprehensive planning and regionwide im-provements is not in evidence. With few and apparentlybrief exceptions, the local district rather than the regionas a whole seems to have been the basis for the design andmaintenance of the canal irrigation and transport system.The main channels themselves accordingly were for themost part closer to the "natural" than to the "artificial"end of the continuum.

The massive westward movement evident in the pres-ent position of the Euphrates can be partly documentedfrom the sequence of the two figures. By the time of thesecond, there had been some westward movement in thecenter of gravity of settlement. This was particularly the

155

oi.uchicago.edu

Dayr

SIPPAR1234

1 Early Dynastic I2 Early Dynastic II/III3 Akkadian4 Ur III-Isn-Larsa( ) Reduced or doubtful occupation* Trace of occupation

) + (4)

(4)34 (0334-(4) * 4 3 4

et:".biWI h 4'03)4 4.I. 34 '3)344·4 ' 4 .(4)

)4'1.234

4 "34

NIPPUR3 \41;ýý4 N,.

Wiloya12 3 *

(34)

133

12 4

34

34'\.! X2)24 134.44 "34 --- ..-

3 " 34- 4

2 44'

234 23

SSHURUPPA DAB4

3 \3 , 34

" *" -. -

)UM M

KISURRA0 234, \<, .,4 , Jidr

4 34 \ , •1234

\SHURUPPAK14 4 -.2,

4! (34 .'4 23

!4 1 31234

43^.f. \ '" ./ ,' ÷

34.)

S 1 , 1.. " '4 4 1: 444 +

1.234I P3v (4)4 BAT1131R† † † † † † † † † † † † † † † † † † † † †4

0 m 525 Km.4 432344l

Fig. 27. Sites and watercourses of the third millennium B.c. on the Mesopotamian plain.

156

+

4 i

4

(3)4.4 (1?)4ýPUZRISH DAGAN 4

4 3

+

123

(

*JTHAa2341234

I34

2.34

3 \

02)34

34

072)

"DI(2

1

4. 234

412344e4./ 4 234

oi.uchicago.edu

!Y4)

SIPPAR ' (2)1 44 .3 1 Ur III-Larsa

S" 2 Old Babylonian3 Cassite

2..4. ' 4 Middle Babylonian(4 - 3A 2 2 23-

2 : ••( ) Reduced or doubtful occupationI\ 3, 2;3 3.

, , 224 123., 3 * Trace of occupation23'3 • ,

4. _ _ 3 _ KvTH -"23 23\

23 V214

A \ Z3 23 2 ,4002) AA0 )00

(00 3Ko 4 '2 (0 -1234 123'- -1 1-. 3 'IZ -,A,3 1233 434 1233~'113 _323 2 123 3-?ý 1234."ri,,34 2 -3i'1 --1234 ·1)-4 121123 3t2340 · 23 p? 13, R ý.3 2323413.0 ý- 23 123'123 23 -233 4 34''ý4, 42j3 -1 ., .,( N34 4?.J'-(1)34 * N2) !ý t3* -1 .1 23-34 4' 4 23 1234 323 3423123 ýt4 134" ,4 Z V23 -1+ 34

\4 23,,?3(3,23 4 .IDAP) 23 r 23 .,222P ISIN 40 jI ,'1234 )? 2 12'n I12ý34 ý -/ 1-1 f. i23' 23 1 123-34 - ~ 7 lJ,,Jd-23 Q -123

123 I tl3 ZRp*'2,3 2 44 2ý3 1' 34---123 .31(' / 3) 41234 -23 g 51-123 V23 ZABALAM-1 > \ h 1'

2.3 , 0 0 0 0012 S-i 1ý3:12 12-

'0!,003'4(00 00 1 (034

\, 3 3, 4....3, 34. 124 23- 34 1 434 .I \ -/-3 ._ký ._ 3 23 -'2 12234-34 34 ~0 . .. - Z4 . 00 .3 .3. (I?.... 3 3, .34 - , . .3i. . .

02)34 00)134 3 ",4 123 --3

'...- -ý _133B 3A D(((1234) _3

00 (3() 00 0

0'0) (000)00I- 121A12r / rs·

; )0141

2 2 .2 0 (

3 233 V12

3 N_1. -2ý '42 1234.\12

2 , \ý .3 3

o~oo-34oo .13 Oo 00023123Laar

Fig. 28. Sites and watercourses of the late third, second, and early first millennia B.C.on the Mesopotamian plain.

157

LAGASH

NINAI120

a~ii.i

~ ~~

-y··

(131--

g

\

· Ct 1234 '12r·

iii

0 25 Km

'·;1i··-··

6 \ ··I····

·sir

"·.)

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

case to the west and south of ancient Isin. The westwardmovement of power and population from Kish to Baby-lon as the major northern city additionally must imply ashift of the same kind in Akkad. Babylon lies alongside theHilla branch of the present Euphrates, the more easterly ofthe river's two main beds, but among other ancient townsof any importance only Borsippa, Dilbat, Marad, andperhaps Kazallu can be located near or between the Hillaand Hindiya branches of today. To judge from limitedand problematical surface collections, none except Baby-lon were of major size or more than local influence withinthe long span of time covered in this chapter. Tentatively,therefore, the shift appears to be of later date. Consideredin the abstract, site distributions as reflected in these fig-ures seem to indicate that most of the Euphrates, or atany rate the main flow that was utilizable for irrigation,continued down from the center of the alluvium and pastall the major Sumero-Akkadian cities.

On closer inspection, however, the sequence of figures27 and 28 supports a somewhat different conclusion. Thegeneral direction of flow of virtually all the major water-courses in the third millennium was in roughly parallellines from the north-northwest, conforming to the slopeof the alluvium itself. The pattern was more complicatedin central and southern Sumer, to be sure, but this south-eastern region was one of minimal slope. As I have noted,moreover, it was the most heavily urbanized part of thecountry. Many of the canal lines shown in figure 27 thatfail to take the prevailing direction can be thought of asradiating from or converging on individual cities.

In figure 28, on the other hand, we can see a number ofelements of a radically different pattern. Beginning in thesecond millennium, the older channels were supplementedby new ones crossing the plain from the west-northwest.These serve many of the same cities that had been locatedoriginally along the north-northwest channels, but theydo so by following the plain's contours perpendicular toits slope or by cutting diagonally across those contours.Here we are clearly dealing with canals that must havebeen largely artificial in their construction, even thoughsome are apparently more than 100 kilometers long andmust have been correspondingly large in capacity. Theobvious explanation is that they were designed to tap agreatly increased proportion of the Euphrates flow thatnow was finding its way down the more westerly of itschannels rather than down the center of the alluvium.

Associated with that shift in flow must have been anunprecedented degree of waterlogging of the westernpart of the alluvium. The formation of large, relativelypermanent swamps and lagoons would have increasedthe isolation; and hence the autonomy of the region, aswell as increased the difficulties of constructing canals totransfer water eastward across it. Some insight into thecontemporary character of the countryside is perhaps of-fered by a shallow lake bottom covering more than 100

square kilometers, now only periodically submerged andgenerally dry but thickly strewn with shells, whose originsmay well go back to the mid-second millennium (see site1572 in the general site catalog).

Striking as they undoubtedly are, these changes neednot be thought of as a full and irrevocable replacement ofone system by another. The lesser slope of the increasinglyartificial canals would have increased their tendency toaccumulate silt, making the older levees more attractiveconduits for irrigation water at least in terms of re-quiring a smaller labor input for maintenance on the partof downstream users. If there were accidental or politi-cally motivated diversions upstream, however, the newcanals would still permit life to continue in the ancientcities in the center of the plain. The actual amount ofwater that shifted to the more westerly Euphrates chan-nels is thus difficult to gauge. An irregular series of move-ments followed by partial corrections is likely, such thatmassive new irrigation works had to be constructed if thesouthern cities were to guard against the catastrophic lossof irrigation agriculture, their primary basis of subsistence.But the proportion of water flowing in the various chan-nels at any one time must have varied in accordance witha host of local factors, and the westward movement of theEuphrates probably should be seen as a long-term "proc-ess" rather than a single "event."

The position of the Tigris is considerably more obscurethan that of the Euphrates. Even Akshak, probably themost substantial of the ancient towns along or in thevicinity of the Tigris, has not yet been positively located.In fact, not a single settlement on the alluvium identifiedwith the Tigris in pre-Hellenistic times can be identifiedthat would permit the location of any part of the Tigrisbed (or beds) to be specified. The hydrological back-ground for this unsatisfactory state of affairs has been de-scribed earlier (see above, pp. 6-7), and here I needonly note once more that in any case there was no band ofdense cultivation and urban settlement along the Tigriscomparable to what existed along several Euphratesbranches. As I also noted earlier, however, the apparentprehistoric confluence of the Tigris and Euphrates nearthe upper end of the alluvium was not necessarily fol-lowed by an early, complete, abrupt separation of the tworivers (see above, pp. 16-18). In addition to a branchsomewhere in the vicinity of the modern Tigris, someTigris water may at least periodically have joined withwhat have been described as "Euphrates" branches flow-ing nearer the center of the alluvium. The final separationmight well have occurred in connection with the westwardmovement of the Euphrates just referred to, in the secondmillennium.

The important ancient mound now known as Tell al-Wilaya has some bearing on the probable course of theancient Tigris. As I noted earlier, Postgate has plausibly,if still not conclusively, identified the site with the city of

158

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

Kesh. If the identification is correct, Postgate observes,"the 'Kesh-river' would almost certainly have been abranch of the Tigris, and so Kesh would have lain on aline of communication by which the cities of southernSumer were connected with the Tigris, Aksak and othertowns of the Diyala group" (1976, p. 80). Wilaya's posi-tion, somewhat detached from all the centrally locatedwatercourses and markedly closer to the Tigris, clearlysupports this speculation. A site of its size could not haveexisted in isolation from some substantial source of water,and the Tigris is by all odds the most likely. But it is to beregretted that the admittedly limited survey coverage inits vicinity has failed to reveal a suggestive line of neigh-boring settlements from which either the Tigris course orthat of the "Kesh-river" leading from it might be inferred.

The Sumerian antecedent for the name Tigris is Idigina(Akkadian Idiglat), but its geographical application issomewhat obscure during the periods we are concernedwith here. Edzard and Farber identify the name as usedin Ur III economic documents with a canal flowing alongthe approximate line of the present Shatt al-Gharraf,which runs southward from its offtake above a weir nearKut, while literary references seem to apply to a water-course farther east (1974, p. 269). Numerous references towork done along the Idigina banks or at the Idigina Riveralso are found in the economic texts from Umma (J. W.Turner, pers. comm.). This strongly suggests that territor-ies administered from Umma extended northward to theTigris, and it at least implies that during the Third Dy-nasty of Ur the northern hinterlands of the city werewatered in part by feeder canals from the Tigris.' In earlier times references to the Idigina converge on acanal situated along the eastern frontier of the kingdomof Lagash, hence well to the east of the modern Shatt al-Gharraf and beyond the range of survey coverage (Edzard,Farber, and Sollberger 1977, p. 217). A suggestive begin-ning has been made at locating the position of one oranother of the watercourses to which the name referswith the aid of archaeological survey techniques (Jacob-sen 1969, p. 105), but full resolution of the problem willrequire much further work in a still essentially unchartedarea. It is apparent once again that a watercourse existedsomewhere to the north, whence irrigation supplies couldbe brought by canal into the region of Lagash, Larsa, andother southern cities, but as yet there is no basis for de-ciding whether to place it in the latitude of Wilaya, in thevicinity of the present Tigris bed, or perhaps still farthernorth (Adams 1965, p. 41).

In Cassite times there is a tantalizingly unspecific ref-erence to Tigris water being introduced by canal into theNippur region. A letter found at Nippur, probably a copyof one sent to the king by an official stationed there inthe thirteenth to fourteenth centuries B.C., suggests thatthe latter "would even dig out the namgdru-canal fromthe Tigris" to provide water for certain date palms (Biggs

1965, p. 97). A glance at the distribution of Cassite settle-ments in the district for which Nippur is likely to havebeen administratively responsible (fig. 34) indicates sev-eral possibilities for this canal to the east and east-north-east of the city. But the wording of the message implies awatercourse that was in at best intermittent use. More-over, the suggested course of the canal emanates from thegeneral vicinity of Tell al-Wilaya, by then long aban-doned. If anything approaching contemporary conditionsalong the nearest portion of the Tigris bed obtained at thetime, waters withdrawn along the suggested course wouldhave needed to be brought up from the deep Tigris bedthrough the use of lifting devices (Biggs 1965, p. 102).

Similar ambiguities surround the naming of the "Eu-phrates" branches in the center of the alluvium, eventhough in this case we can follow at least portions of theircourses in considerable detail. The essential difficulty isone that Nissen (1976, p. 13) has adumbrated, that theadjacent position of a particular watercourse and a par-ticular town was self-evident to a scribe of the time andhence was almost always omitted from the texts as redun-dant information. It is evident that there was a westerlybranch of initially limited importance, even though itsidentification with the present Shatt al-Hindiya is perhapssomewhat too facile. This is the Abgal of the Akkadianperiod and perhaps earlier, becoming in time the Paluk-katu and ultimately the Pallakottas that was familiar tothe Romans. The name probably survives today in thetown of'Falluja. Farther east lay the Arahtum, along acourse southward from Sippar through Babylon, Dilbat,and Marad and thence perhaps toward either Isin or Uruk.Beyond it lay the most important early bed, in historicalif not necessarily in hydrological terms. Named Buranuna(Akkadian Purattu), from which the word for Euphratesis derived, it seems to have flowed at one time throughSippar, Kish, Nippur, and Shuruppak to Uruk and thenceto Ur. Much less certain is the identification of a still moreeasterly branch as the Zubi. It is also a matter of disputewhether a major channel connecting Adab, Umma, andBad Tibira with Larsa, below which it apparently bifur-cated into branches flowing toward Larsa and Lagash, wasever generally referred to as the Iturungal. Whatever itsname, however, it must have derived at least in part fromone or more easterly branches of the Euphrates fartherupstream-possibly the Buranuna, possibly the Zubi.Adding further confusion to this highly uncertain pictureis the fact that many parts of the entire system were fre-quently referred to merely as the "Sippar River" (Edzardand Farber 1974; Edzard, Farber, and Sollberger 1977,passim; Adams and Nissen 1972, pp. 42-47; Jacobsen1960).

Little is to be gained by attempting to specify the loca-tion of smaller named components of the system, of whichthere are many, when even the major channels presentsuch difficulty. The reconstructions of watercourses

159

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

shown in figures 27 and 28 have not been derived fromtextual considerations at all. Instead, they are based onsuggestive lines of contemporaneous archaeological sites,sometimes adjoining ancient stream or levee traces thatcan be seen in the air photographs or in the LANDSATimagery. It should be kept in mind that the watercourseroutes thus reconstructed in no sense exhaust the possi-bilities even for the major channels. This is especially sosince the two maps cover a span of some two and a halfmillennia of repeated modifications, reconstructions, andnaturally induced changes.

If we assume that the canal running through Maradwas the Arahtum, for example, the name may have beenapplied to a series of widely divergent alternative routesbelow the town as particular channels silted up, wereabandoned, and ultimately were replaced by new ones.The only levee that can be detected in the satellite imageryin this case seems to be directed toward Uruk and Larsa,but it seems very likely that a branch, and perhaps eventhe entire flow of this channel, was diverted instead to

\Isin during periods of that city's hegemony as a capital. Ifso, it would be fairly natural for the name to be trans-ferred along with the water. Still later, Edzard and Farbersuggest that with the drying up of the former channelfarther east known as the Buranuna this name was trans-ferred to the Arahtum (1974, p. 272).

In other words, there is every reason to expect that no-menclature was as flexible in its application to the entirenetwork of streams and canals as the latter was continu-ally changing. This sharply limits the utility of the textsfor reconstructing not only the canal system but the ad-joining settlement patterns, not to speak of the subtle andpervasive changes both underwent. Written records pro-vide a picture that is tantalizingly detailed for limitedperiods and areas, to be sure, as is especially well illus-trated in the kingdom of Lagash (Falkenstein 1966, pp.17-41). But until excavations have made it possible toidentify firmly far larger numbers of small as well as largesites, the picture provided by the texts is also on the wholeexceedingly fragmentary. Most significantly, there are vir-tually no secure geographic reference points except for ahandful of the major cities. It is in the light of these de-ficiencies that we should now turn to the far more limitedand superficial-but also far more systematic and "com-plete"-data of archaeological surface reconnaissancewithin the intensively surveyed region.

Figure 29 illustrates the late Early Dynastic distributionof sites, to be directly compared with figure 21, in whichthe identical region is mapped for the Early Dynastic Iperiod. A series of striking changes apparently had oc-curred during the intervening interval, most notably theextensive abandonment of scores of smaller, outlying set-tlements away from the major watercourses. Perhaps onecan speak of "linearization" of the pattern, with ribbonsof close-spaced towns (and presumably adjacent strips of

cultivation) replacing a looser and much less dense ar-rangement in which there had been many minor effluentsand small, disarticulated clusterings. The Early Dynastic Imap suggests a mosaic of small, adjacent patches of steppe,swamp, and differentiated cultivation. By the Early Dy-nastic III period the patches had coalesced into muchlarger, probably more contrastive, bands and zones. Justas most of the villages in Uruk's protohistoric hinterlandsdisappeared, their inhabitants drawn into the city, so therewas an emptying of the outlying terrain around morenortherly cities like Adab and Nippur. Perhaps the largestsingle movement involved the abandonment of a series ofsites south of Umma, a number of them of urban propor-tions. Some of the population of this district was drawnoff southeast, into Bad Tibira, but the greater part is morelikely to have contributed to the rapid growth in the po-litical importance of Umma that took place in the lateEarly Dynastic period.

There is a perceptible increase in the density and extentof settlement along the more easterly of the two mainwatercourses shown in these maps. This finds direct ex-pression in the size and distribution of sites along the linerunning from above Adab to below Umma, and perhapsalso in the leading political roles played by Umma and itsdownstream competitor Lagash, in late Early Dynastictimes. But it is also noteworthy that there is little if anyfurther growth at the same time along the Shuruppak-Uruk watercourse, in spite of the 54 percent expansion insettlement for the region as a whole that was reportedearlier (table 13). If suggestions of a possible steep declinein Uruk's population are taken seriously (see above, p.132), it even appears that the Adab and Umma districtsmay have grown directly at the expense of the districtsbetween Shuruppak and Uruk. And it is noteworthy thata canal leading from just above Umma (site 175) in thedirection of Uruk apparently came into greater promi-nence as the Early Dynastic period continued. This invitesthe speculation that water shortages were being felt in theUruk area, while there must have been supplies to spareon the Umma channel, since otherwise the forces of Ummawere easily in position to prevent such a diversion.C The Akkadian period, shown in figure 30, further accen-tuates the shift of settlement away from the Shuruppak-Uruk channel. Occupation at both of those ancient citieswas severely limited, if indeed they were not largely (ifonly temporarily) abandoned. Similarly, the reoccupationof an abandoned prehistoric town (site 1237) southeast ofNippur hints at the possibility that water in the Nippurchannel was now being diverted toward Adab, rather thanbeing allowed to maintain its former direction towardShuruppak. The Adab region thus became particularlydensely occupied, and it is arguably the largest urban con-centration yet known within the Old Akkadian realm.Attention should be called to site 1188, among several oth-ers of apparently urban dimensions. With the provision-

160

oi.uchicago.edu

(

'p

k

111

tp

C,lp

*1237

I -\ ^A~

* 0.1-4. 0 o L* 4.1-10 HA. "kJ0 10.1-20 0 oCCr* 20.1-40 @ P* 40. OCCu

@ 200., POLITICAL CAP

Fig. 29. Late Early Dynastic period settlement patterns.

161

ISINE

c.

c.-

-p

A

I\

I! +\\\\\\

\

BAD TIBtRA 0 '

\

0 LARSA

do

i \kpC,

oi.uchicago.edu

A%00

Li %%

t- \'

r(

* .1032

\

I//..,/

0-xCýID

*

0

S 1383

*I

/

\\

/II

/ -

S0.1- 4.0 / -"* 4.1-10 HA.0 10.1-20 O OCCUPATION DOUBTFUL0 20.1-40 @ POSSIBLY LARGE BUT* 40+ OCCUPATION UNCONFIRMED

@200+. POLITICAL CAPITAL 0

SSHURUPPAK

0200

I/

//

131d

0242

NI /

- ~URUKWC~C

A. ~ \15 NM _

BAD

) LARSA

Fig. 30. Akkadian period settlement patterns.

162

,/

/

0

C,

c

· 1P

'p,"

<>

S*Wilayaj.

*A

0

0

0

/

/I!

0

S\

!

t

i,

oi.uchicago.edu

SO639

\- ' K \RIS(

,v \ 0 , C. . . ,S.1237 + ' ON

I - N PUR \ ^ / /

.DABA

R 0SI

. / * l- -. --

4 \ \ /\ \\l o0 -

S\ \ \

,' \-"' to-'- 'p

S\ \S \ \ . * 1 T 1

S..* 7-/ K\LSO ,P' \ \ D O F

S,, ,' -- "- ,"\YLRE U \ T ,

. / • / • \

\ \ o\\ * O

\ 0DAGAN 1231 1 ,1 --

: @ * 1237 *^ ~ / ^

4 OCCUPATION U I \'

L C PI T K ISURRA. \0 IA _ * . UMMA

\ *\ \

* 40. OCCUPATION UNCONFIRMED 0

200.. POLITICAL CAPITAL o 15 KM. ^. 4

Fig. 31. - Ur III-Isin-Larsa period settlement patterns.

163

--

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

ally assigned name of Tell al-Hafriyat (descriptive of theextensive illicit digging there), it has become the focus of aprogram of scientific excavations by the Oriental Instituteunder the directorship of McGuire Gibson.

Of special significance as an indicator of growth is theappearance of a dendritic fan of canals or river effluentsnorth of Adab and east of site 1188. This suggests the in-troduction of a zonal irrigation system on at least a localscale, as opposed to the confinement of irrigation to thebackslopes of the main levee. In one sense this representsthe modest beginnings of a reversal of the trend towardhypertrophic urbanization that had culminated in the lateEarly Dynastic period. But the reoccupation of formerlysettled districts probably was now on the basis of arti-ficial canalization under the aegis of a victorious dynastywith new resources of wealth and labor to deploy, ratherthan along small, insequent streams of a prevailingly nat-ural character.

With the advent of the Third Dynasty of Ur, large-scalecanal systems of this new kind quickly reached a peak ofdevelopment. The increasing linearity of most of the mainchannels is clear in figure 31 from the regular position-ing of adjacent sites, as well as from numerous survivingtraces of levees (many of them repeatedly reutilized inlater periods). Around Adab the relationship of settle-ments to the watercourse on which they depended is es-pecially well highlighted, largely as a result of the ex-tremely rapid surface erosion going on in the area owingto the presence of a heavy belt of dunes (fig. 32). Adaband its dependencies as they may be seen here probablywere already flourishing before the Ur III period, but atleast the smaller towns are likely to have reached theirmaximum size at this time. Adab itself, on a substantiallyelevated mound, would have been visible from a greatdistance. From each of the fairly substantial towns strungout along the main watercourse below it, however, severalother towns and villages would have been within easyvisibility of an observer of the time. And the watercourse,whether it is thought of as a "river" or a "canal," wascertainly of impressive proportions. While there are somedeviations from linearity in its course, they are extremelymodest in relation to its apparent width. Its propensity todevelop meanders was natural and could not be entirelyprevented, but they have been kept limited by what musthave been an unremitting program of maintenance. Theentire layout is one that would have been optimal both fortowed riverine commerce and for irrigation agriculture,with minimal loss of fields and gardens owing to uncon-trolled meander-cutting.

One must bear in mind, however, that improvementslike these incurred substantial hydrological costs. Canali-zation of a lengthy reach above as well as below Adabartificially increased the stream's gradient over what itwould have been had meander paterns been allowed todevelop naturally. As Schumm observes, "a river in most

Fig. 32. The third millennium Euphrates at ancient Adab,also showing modern (ca. 1962) dune formations.

164

L

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

cases can be straightened or made more sinuous, but thereis a limit beyond which the channel becomes unstableand aggradation and blockage of the channel or severescour and bank erosion result" (1977, p. 149). In fact, theapparent width of the stream below Adab may be a re-flection of these processes at work, with a wide, shallow,braided pattern having been substituted for what wouldhave occurred under less disturbed conditions. One cansee in figure 31, moreover, that the predominantly linear,canalized system is interspersed in at least two parts ofthe region shown by areas in which meanders were al-lowed to develop vigorously and perhaps with very littleeffort at control. This can hardly be understood as a fea-ture of intentional design, since the areas in question,including one around Isin and one in the vicinity ofZabalam and Umma, would have represented a consid-erable disruption of local land use and communicationspaterns. Probably their existence must be understood as anunavoidable compromise between the watercourse controland modification objectives of state planners of the timeand the powers of the natural hydrological system to re-assert itself (cf. above, p. 21).

As figure 31 shows, zonal irrigation systems were intro-duced in many parts of the region. Perhaps the largest, atleast within the intensively surveyed area, lay north andeast of Nippur. There we can identify a roughly rectangu-lar latticework of new canals adjoined by scores of newsettlements. Most of the latter are fairly small and musthave been primarily occupied by cultivators, but at leastone (site 639) was unquestionably of urban proportions.Conceivably this could be the one in which Shu-Sin (2036-28 B.C.) resettled the war captives he brought from Shima-num and surrounding districts (Gelb 1973, p. 76).

There were other latticeworks of new, integrated canalsystems, seemingly smaller than the one northeast of Nip-pur. Most were later in their time of construction and pri-mary use, since they adjoin cities that emerged (or re-emerged) into political prominence only after the collapseof the Third Dynasty of Ur. One lay south of Isn andpresumably postdates Ishbi-Erra's (2017-1985 B.C.) usurp-ing of Ur's powers and administrative system. Its westernand southern peripheries abut the frontiers of modern cul-tivation and hence the limits of survey, so that perhaps itwas originally more extensive than is shown in the figure.But the absence of reported remains of this time on admin-istratively recorded sites within the cultivated zone (seeabove, pp. 15, 43) argues against any substantial under-estimation in the area of the system owing to this factor.

An additional latticework of intermediate size ransoutheast from Uruk and its northern environs to and be-yond Larsa. It too must date to the "Zweite Zwischen-zeit" of contending successor states after the fall of Ur,possibly owing something to the Sinkashid dynasty inUruk (ca. 1865-1810 B.C.), but more probably reflectingthe vigorous economic policies of Larsa kings during the

long span of that dynasty's prominence between Gungu-num (1932-1906 B.C.) and Rim-Sin (1822-1763 B.C.) (Ed-zard 1957; Falkenstein 1963). The scale of state-sponsoredirrigation activity of that time is illustrated by a series oftablets from the reign of Sumuel, in about 1880 B.C.Records were kept of the fabricating of more than1,300,000 bricks, about a third of them not merely sun-dried but also fired, and the whole constituting a volumeof some 30,000 cubic meters, for the installation of a singlereservoir at the mouth of the Isin canal (Walters 1970, p.137). Much was also made by other Larsa rulers, includingRim-Sin, of the prosperity induced by royally sponsoredcanal projects, some of them even "leading to the sea,"while in his time as well as that of Sin-iddinam (1849-43B.C.) undertakings as ambitious as the restoration of (aportion of) the bed of the Tigris were proclaimed in royalannals (Sollberger and Kupper 1971, pp. 191, 205-6).

Finally, textual sources from Umma indicate that a sim-ilar latticework extended eastward from that city in thedirection of Lagash, and the number of canals and fieldswhose names are mentioned in Umma texts suggests thatUmma's canal system must have been of considerable size(Sauren 1966). In that direction again, unfortunately,modern frontiers of cultivation have heretofore foreclosedthe possibilities of intensive archaeological survey.

There is less to be said of the Old Babylonian period,primarily because it was a time of sharp economic anddemographic retrenchment. Figure 33 provides no evi-dence whatever of newly resettled areas or other new irri-gation initiatives, and table 13 records a 40 percent declinein the aggregate area occupied by the major urban cen-ters (those over 40 hectares in size).

By the last decades of the eighteenth century B.C.,dated tablets had virtually disappeared from southerncities (Stone 1977, fig. 2), probably not signifying theiruniform and complete abandonment but surely suggest-ing a deep disruption in the routine fabric of civil admin-istration and ritual. Old Babylonian sites directly on thebed of the Adab-Umma channel (e.g., sites 1175, 1460),the main artery of settlement in the Ur III period, suggestthat its flow had ceased entirely. Quite possibly it hadbeen replaced by much smaller canals running along itsbanks or levee, but at the very least the volume of wateravailable for irrigation must have been greatly reduced.Moreover, the possibilities of shipborne commerce to andthrough this particular region must have essentially dis-appeared. Citing the repeated references to the extensionof irrigation systems and the founding of new settlementsin northern Babylonia by the predecessors of Hammurabi,Nissen rightly observes that these forms of greatly in-creased water utilization upstream were surely an essen-tial source-one might even guess the essential source--of the ensuing decline throughout the heart of Sumer(1976, p. 24, n. 79).

The pattern of Cassite occupation, shown in figure 34,

165

oi.uchicago.edu

% %

00I 0-A

- cc I

S(14<4,

/

',<>

'5.A-

r

N

$

~0 \

IL N

·

/IN

r -

" \

( \)

/-·//

* \

\ · ,.K4

Ok

KISURRA

- 0

S10.1-202'1 40 0 OCCUPATION DS201-40 POSSIBLY LARGS40. OCCUPATION U

@200*. POLITICAL CAPITAL

I ~/i

URUKO

A- 0\i 5. n

OUBTFUL 0E BUTINCONFIRMED

0 15 KM

/

0 i

* \ 0

S \ I .; " .

Woo ^

Fig. 33. Old Babylonian period settlement patterns.

166

4-

-320N

I0

ffr

/

% % ..

"x

~1395

·,\O

I

oi.uchicago.edu

A

\~1

0

0

k

/

*ISIN

*. *0

01389

+

* 0

So.1-4. '0 -* 4.1-10 HA.* 10.1-20 0 OCCUPATION DOUBTFUL

S20.1-40 POSSIBLY LARGE BUT* 40, OCCUPATION UNCONFIRMED

O200÷. POLITICAL CAPITAL 0 15 KM.

0

Fig. 34. Cassite period settlement patterns.

167

tp

10

.0

^ "~ -s

· ·

\

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

initially appears little different from Old Babyloniantimes. On closer inspection, however, the dependence ofthe southern part of the region on lengthy canals from the

kvest-northwest had become much more pronounced. Itis not at all certain that significant supplies of irrigationwater were continuing to find their way into the regionaround Uruk, not to speak of other ancient cities to thesouth and east of it, along any of the old levees that leddown from the north and north-northwest. There was akind of geographical fragmentation, in addition, in which

-districts began to detach themselves from one another andto constitute increasingly distinct enclaves. And the declin-ing total population suggests that even within these en-claves an at most fairly extensive and dispersed form ofagriculture was carried on. The cumulative effect of thesetrends becomes clearer if the map of Cassite settlementand irrigation is compared not merely with that of thepreceding Old Babylonian period but with the markedlythinner occupation of the following Middle Babylonianperiod as well. The latter is shown in figure 35.

Three principal enclaves can be distinguished withinthe intensively surveyed area. The largest seems to havehad Nippur as its "capital," and Nippur's rich archivalmaterials of Cassite date indeed clearly identify it asone of the most important administrative centers in Baby-lonia as a whole. Isin was the major city in the second,supplemented by the impressive and previously recordedcenter at Umm al-Khezi (site 1389) and perhaps for a timealso by Jidr (site 004). Uruk exercised the same dominanceover the third.

The formation of these local realms seems to have coin-cided with an abrupt, unprecedented change in the hier-archy of settlement sizes. Dispersed, small villages attaineda higher frequency than ever before, as did towns of lessthan 10 hectares of occupational area. But at the sametime there were sharp drops in both the numbers and theaggregate area of urban sites larger than this. Cities of thelargest size category disappeared altogether (cf. table 13,fig. 25). A typical unit of settlement in both the Cassite andMiddle Babylonian periods consisted of a string of ex-tremely small sites, relatively close together along a canaland often separated from other such strings by considera-ble distances. The pattern for each of the enclaves tendedto become one in which one or two large, probably well-fortified centers dominated their hinterlands, while mostof the population lived in small agricultural villages. Akind of "feudalism" is evident, perhaps not in the senseof institutions and relationships specifically recallingmedieval European vassalage, but in the more generalizedsense of a decentralized system of landed, patriarchalauthority with a wide bifurcation between the mass of de-pressed agricultural population and a strong military elite.

One of the important Cassite-Middle Babylonian ca-nals supplying the district north of ancient Uruk has beenfortuitously exposed by surface erosion, permitting a

more detailed glimpse of the rural irrigation regime. Thecompact, resistant clay of the canal's bed, some 8 meterswide, could be followed fairly continuously for more than5 kilometers, even though the associated spoil banks hadbeen almost completely scoured away by the wind. Withinthis distance six roughly contemporaneous archaeologicalsites were identified (1570, 1584, 1589, 1590, 1592, and1594), including three of 0.1 hectare or less, one of0.6 hectare, one composed of two adjacent mounds aggre-gating 3.6 hectares, and one so small that it may have con-sisted of only a branch canal head gate and closely associ-ated structures.

Clearly, only a modest population depended for its liv-ing on at least this particular channel, even along thewell-exposed segment where it can be followed with highexpectation of virtually complete recovery of adjoiningsites. The absence of contemporaneous sites for some dis-tance upstream or downstream along the same general linereinforces this impression of an extremely dispersed aswell as limited rural population. Yet the uniformity andlinearity of the canal and the surviving traces of its spoilbanks leave no doubt of the labor-intensive character ofits construction and maintenance. At site 1590, in fact,there is a kind of metaphor of the succession of irriga-tion and settlement systems. Surface riffles and vegetationpatterns trace out the shifting bed of a meandering, essen-tially "natural" stream from the north or northeast, flow-ing past the nucleated ruins of a town (site 1591) thatmay have survived into the Old Babylonian period. Likethe town (and perhaps for the same reason), the streamexpired at about this time. Running across and superim-posed upon the traces of its bed was the straight ribbon ofthe bed of the Cassite-Middle Babylonian canal from thewest-northwest, serving few if any towns and accompa-nied instead by a string of small, amorphous hamlets.

This local vignette, or at any rate the larger pattern itexemplifies, suggests a different perspective on the cyclicalpattern of centralization and fragmentation to whichmuch of this chapter has been devoted. Middle Babylon-ian population densities seem to have drifted downwardto levels that had not been obtained since the Uruk period,three millennia earlier. But this emphatically did not implythe full restoration of earlier conditions, as if one hundredand fifty intervening generations had had little appreciableeffect. The Cassite and Middle Babylonian cities, after all,were the repositories of a rich literary and cultural tradi-tion that had taken shape over this long span. The BronzeAge had come and gone, but the Iron age now beginningwould transform not only local technologies and econo-mies but also the mechanisms and motivations for interre-gional contact and conflict (Childe 1942, chap. 9). Newdomesticates had been introduced or borrowed, amongthem the camel, that enabled nomadic peoples of the aridSyro-Arabian steppelands to begin to encroach upon thecivilizations of the Fertile Crescent as a historical force

168

oi.uchicago.edu

A

O0

K

.. \1U N*1N '4

NN N 1 o~1-,

Kd·N

NO I\.' o~

*~ 0

NIPP~h@~

S,\

1231 N

-~ V.

C I

C

SISIN

\\

//

01389N.

K

~ -

N /- Ž ;*--- .

NN,

I- 0\

- '4

1 V.1- O OCCUPATION D* 2-0 POSSIBLY LARG

* 40. OCCUPATION U

O 200*, POLITICAL CAPITAL

)OUBTFUL5E BUTJNCONFIRMED .

0 15KM.

K

\

*

0

, N

Fig. 35. Middle Babylonian period settlement patterns.

169

4+

\

S\

t" t

'N

J,C·

450

k

//

//

- ./

/

•-32-ON\\

---

K

N

N

N-

0~.,

\N'-,~ N,,

%. 0_

\,

± |32"N

/. ^-

o\ \\

\%%\

\\

\

\\\

\\\

N\\

\O

N

- -N'

NNNN

N /

+

\\

\\\\-B

0

N

1 \ -

N

i

I

S'

\

I

t

%

a

/

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

(Bulliett 1975). Above all, the frontiers of the knownworld had spread immeasurably.

Even at the strictly local level, however, the pattern wasappreciably different from that in the fourth millennium.Dispersed or not, the regime was highly articulated po-litically, authoritarian, and hierachically structured intodistinct social classes. With the perfection of a writingsystem and the growth of a tradition of literacy had comesubtle, complex, tenaciously retained bureaucratic formsof high administrative "density." Extensive canal irriga-tion had been introduced that involved planned, massiveinterventions in natural systems of drainage. Rather thanwithering away under conditions of population decline

and ensuing economic and political stagnation that be-gan in the late Old Babylonian period, these public workswere even intensified for a time before the growing urbanisolation and paralysis accompanying Aramean invasionsin the tenth and eleventh centuries and the later Neo-Assyrian conquests. No comparable feature was presentthree thousand years earlier to even a remotely similardegree. There were profound continuing oscillations inSumero-Akkadian society, settlement, and patterns ofland use, in other words, but we cannot fail to recognizethat here were cumulative, underlying forces for changeand development as well.

APPENDIX

THE SURVEY DATA BASE: LATE EARLY DYNASTIC-MIDDLEBABYLONIAN SITES AND CHRONOLOGICAL INDICATORS

As I indicated earlier (pp. 130-31), the data on which theforegoing discussion is based are more limited and lessreliable than those employed in chapter 3. A high propor-tion of historic sites were found to have had multipleoccupations or reoccupations, making estimates of set-tlement size for successive periods particularly difficult.Hense it is impossible to summarize the descriptive data onsite size more concisely than previously. Occupations forvarious periods are given in table 14, in terms of fairlybroad size categories rather than estimated areas, buteven within this looser framework it has not been possibleto make all assignments with equal confidence. Readersare referred to the general site catalog (chap. 7) for fullerdescriptive material on individual ancient settlements.

No attempt was made to record systematically theceramic and other dating indicators observed on sites thatwere occupied during this range of time. It is not at allclear that systematic, randomized investigations of surfacecollections provide a useful approach to the analysis ofcomplex, multiperiod occupations, at least without de-voting extremely long periods of analysis to each of theindividual sites so considered. But the presence of fairlywell-defined dating indicators was noted, as was theirmore or less extensive areal distribution over the sitesurface. Continuing the narrative of Appendix A to theprevious chapter, the following were the principal "indexfossils" that were distinctive, common, and securelyenough dated to be useful.

Late Early Dynastic Period (Early Dynastic II-III)

The criteria employed during the Warka survey (Adamsand Nissen 1972) were followed with little change. Conicalcups (AD) continued at least into Early Dynastic III times,

but the later forms tend to be distinguishable from theearlier in that they became progressively wider and shal-lower. Supplementing them were small cups with in-ward-beveled or inverted rims (AE) and larger cylindricalbeakers (AF). "Fruit stands" (GC) continued, probablybecoming considerably more common. Finally, bakedplanoconvex bricks (LG) are not infrequent on sites ofthis period. This form of brick made its appearance inEarly Dynastic I times, but all but one of the observedexamples stem from sites on which the surface materialreflects an occupation lasting into the later part of theEarly Dynastic period or even later.

Akkadian Period

A somewhat different approach to periodization istaken in this study than in the Warka survey, in that herethe Akkadian and Ur III periods are kept separate. Thebasis for the change does not lie in improvements in strati-graphically secure, excavated findings but instead stemsmainly from observed homogeneous grouping of surfacecollections. According to the severity of the reservationsindividuals will wish to attach to any chronological con-clusions drawn from surface materials, the differentiationbetween the two periods that is now proposed must beregarded with a corresponding degree of provisionality ifnot skepticism. Some independent support for differenti-ating between the periods is perhaps afforded by the sig-nificantly different settlement patterns that emerge whenthis is done, a finding that is the basis for much of the dis-cussion in chapter 4.

The apparently coherent conspectus of types that wasused to denominate the Akkadian period included thefollowing: horizontal ribs, triangular in section and usu-

170

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

ally forming a concentric series, applied to the shouldersof large ovoid jars (Delougaz 1952, pl. 101; McCown andHaines 1967, pl. 81:9; Hrouda 1977, Taf. 26); large, deepbowls with a prominent, large-diameter spout with beadedlip, usually with down-sloping flange rims (Delougaz1952, pl. 169; C.053.312; Hrouda 1977, Taf. 27; McCown,Haines, and Biggs 1978, pl. 47:7); ovoid jars with roundedto bluntly pointed bottoms, slightly constricted necks, andsharply profiled rims (Delougaz 1952, pl. 160: B556.540;McCown and Haines 1967, pl. 80:18); and incised-combmeander patterns between horizontal bands of similar in-cisions, typically on shoulders of large jars. It must beconceded that a number of these types may well havecontinued for some time under the Third Dynasty of Ur.

Ur III-Isin-Larsa Period

While the Akkadian and Ur III periods have been dif-ferentiated in this account, an opposite approach is takenwith regard to the division that was recognized betweenthe Third Dynasty of Ur and the Isin-Larsa period at thetime of the Warka survey. The problem to which this isa perhaps debatable response emerged with full forceonly upon the publication of the long ceramic sequencespanning much of the first, second, and third millenniaB.C. at Nippur, not yet available at the time of the Warkasurvey fieldwork. As Donald Hansen succinctly statedbefore publication of that large and stratigraphicallyfairly secure body of material, some forty pottery typescould be recognized in levels succeeding the onset of theThird Dynasty of Ur. These pottery types, he argued,"present a continually evolving series; there are no sharpbreaks until the end of the First Dynasty of Babylon"Hansen 1965, p. 210). Leaving the Old Babylonian aspectof this generalization for further discussion below, hisobservation suggests the difficulty of trying to impose abreak after the Ur III period on the basis of surface col-lections alone. To be sure, McCown's final analysis ofthe "Range and Frequency of Pottery Types" (McCownand Haines 1967, table 2) lists six that began only withthe Larsa period. Upon closer inspection, however, theseall prove to be either infrequently occurring or else basedon features of generalized vessel form that are difficult todistinguish in the typically fragmentary material to befound on site surfaces. Hence a grouping of Ur III andIsin-Larsa diagnostic criteria appears to be the least ob-jectionable course of action, even though it has the severedisadvantage of conflating historically very distinctiveperiods.

The following ceramic features were considered tobest reflect occupations within this span of time: cylindri-cal or ovate jars with concave or inset necks, beveled orflaring rims (the latter frequently with profiled undersur-faces), and low disk bases (McCown and Haines 1967, pl.84); flaring, carinated bowls with grooved vertical rims(McCown and Haines 1967, pl. 82; Hrouda 1977, Taf.

27-28); "collander" sherds with evenly spaced, circularholes (McCown and Haines 1967, pl. 82); large flaring-sided jars with a characteristic decorative pattern ontheir slight shoulders and vertical to slightly flaring necksthat consists of horizontal ridges and grooves combinedwith a comb-incised meander (McCown and Haines1967, pl. 84:21); narrow stump bases of small oval orcylindrical jars (Delougaz 1952, pls. 152, 184); and insetbases of thin-walled, well-made cups with vertical orconcave sides (Delougaz 1952, pl. 153; McCown andHaines 1967, pl. 89). Stump bases and thin-walled cups,it should perhaps be added, are apparently indicative ofoccupations during the latter part of this interval. At afew sites an effort was made to discriminate between UrIII and Larsa occupations on this basis, but this generallyproved more difficult than in the Diyala region, for ex-ample, where the cups in particular were notably moreabundant (cf. Delougaz 1952, pl. 115).

Old Babylonian Period

The absence of a clear succession of types coincidingwith the rise of the First Dynasty of Babylon has alreadybeen indicated. McCown's table of distributions indi-cates that only three types went out of use, and six othersappear approximately contemporaneous with that his-toric northward shift in the fulcrum of political power.None of these individual replacements prove to be es-pecially useful when one is confined to working withsurface collections. If one views the distribution of typesat a slightly higher level of aggregation, however, apicture of substantial, fairly rapid change in the con-spectus of pottery types emerges that is more helpful.Six types went out of use during or immediately afterthe late Larsa period, and no less than seventeen wereintroduced either during the late Larsa period or at theend of it. Regarding the late Larsa period as a kind oftransitional interval, therefore, it seems both possibleand worthwhile to distinguish broadly between Ur III-Larsa surface collections on the one hand and (lateLarsa-) Old Babylonian collections on the other.

Two features in particular are sufficiently common andunambiguous to permit identification of Old Babyloniansurface collections. The more important is a globular jarof moderate size with a low, slightly flaring collar, oftena pair of horizontal grooves incised on the upper body,and a low ring base creating the impression of an attachedbutton (Delougaz 1952, pl. 163:B.656.720; McCown andHaines 1967, pl. 90). The second, also widely occurringand useful, is a large, rounded jar or urn with a pro-nounced, generally horizontal flange rim over a series ofhorizontal grooves or ridges, generally with a ring base(McCown and Haines, pl. 89).

Cassite Period

Hansen notes the "tremendous change and break in

171

oi.uchicago.edu

TABLE 14 Periods of Occupation and Size Categories of Sites from the Third, Second, and Early First Millennia B.C.

No. |c • • ,c ~ p S •.

001002004005011016019020021025027032039040041049050052061070078095099100129130131139140142145148149150151152156157161162164165168169172175176190197198200204213235242247248253254255258259261263267278279280283287294300301311315

Ut MtSite c S |No. L. : - c S C

I----------1

I------------!

II I I rTT T T I I II In m~mrrnirn rr nrrrrnrrrra _ _ _ _ ,

r- - - - -- - - -1

I 1

r--------

I- -- -- I

nrlr rnrr r rrlrrr r rrTT~ ITIIC

r I

r

I-- - - - - - - - -- I

I-- - - - - - - - --1

~'////7/~1'~'////////I//fl

I 1

_____I 1

I-- - - - - - - - - - - - --- - -r

I-- - --1

r-- - --1

I-- - --r

r I

___ I I r rT T T T 1 lTll.n

111 1 I I rR r Irrn- I

r---_=--~--~-~-~-~-~3

r I

I I

I I

I

r

L

r

r

I- - -1

r 1

r 1

c-------

r I r 1

I 1 I- - - - - -I

r

Irmmmrmrrmmn

~TIIIIlnlllllTUmllnTTlllrrlmm

I_ 1

I I

_ ~n77 1 177 rrrT 177 T 1711- - -- I

_ nlnnlllll~IllllnllTIII~lIlllIlTl(llrm

aTlll~m)

~nT rr IlI I T TrII i i I lr

I I

nIIllnTll~TTTlllllnrrlrllll~I 1

1 lIlI

~--------------------------

rrrrrr~_~I~I~~

I 1

______ I

I :

I- I___

~--------1

I-- - -- - -1

I 1

I 1____~

I 1_____

I- - -1

I 1

I---------1

I

r 1

I

_____________ rl

529535573576584586627639667671672673674692706713714715720721722 __734738740745

320 i333 L336 i341 ..351 c352 _353 -355358 i359 c360 1 ..361365 L _385 I I392 c400 -----403 c405 .412 ----413 (424 i425 --427 E428429 _-_430431 - - - - - -432 Va, a,433 E437 -438 -439 r1---443 r--444 1zm7446 -448 E-mz7a'451 _ _ rm r453 r---455 I ---457 E---- -...458 c459 -62 r

463466- -

Uruk -ffl _ _ _ //////////77Larsa ._____ f m - - ffflfim

Shuruppak mI - -m -_ _ Ul_ 0ll3u

... 2.-- -C

r 1

I 1

I 1 I 1

I 1

_______I I

r I

r J---.

r I

r-- - - - - - - - -- I

r 1

r 1

I

L 1

r- - - - - - I

r I

Sir 5a i c 9 i

No. s

-- -,

n r r nln r Illr

---1

1

'-------

-------

c746747748749770771772781783794795798802807815818820822824826827828831832834835838844845854858861865869871875898932933944947951954959960970971972976983989993994996997999

100010011003100510061007100810091019102110241025102710311032 ..zz..1034103610371049

172

r I

1

___I 1

r I

I 1_____

r I

r I

r---------1

r

I 1

r 1

r

I---------I

r- - -1

1---7 1

r- - - - - - - - -1_ _ _ C

I I

I

L 1

~----------------~

~-------

~---------------

~----------------------~

I 1___1

I i L-

I 1

I

I

I 1-- - - - - --1

I-- - - - - --1

[ 1

___I 1

___I

___I

I-- - - - -- - - - --1

I I

I 1

I 1

I

I I

I

r 1

I 1

__,I 1

I I

I

I - - - - - - - - --1

rrrrrm

~--------

I

I 1

r I

1

___ r 1

r I

I 1

r I

___ r-----

_______1 I

r I

1

,I I

m i i I i i IIa

_____A

oi.uchicago.edu

Site -45

No. IcCK Q

105410551056105910671069107110751077108010821083108610881089109010911093109410951098109911001102110311091112111311141115112911371143115211571161116611691170117111741175117711781179118011811183118811961199120512061208121012171218121912211226123012311233123412351236123712401244124712491250125612571259

E 1

1

r 1

,,, IIII1T11I1IIIB

I I

I I

I 1

111 I I 1 17 T T I rrl

r

I

~--------

I

I

I 1

I-- - --1

1

I-- - - - - --1_

I-- ---- --1

I

L

I

I 1

I 1

I --I

I

r- - -- --------- -

r- _1

I 1

r 1

r-- - --1

n nrrrr r r ri In

L I

I 1

I·-· .---~-lj

F I

r- ~------77-

r77I---- - -

C7"

cLzj

-I I

Site ý 4 4 Cs

No.

126712711278127912841290129312941299130113031304130713081309131113171318132013211322132313241325132613311333133713391340134113421345134613471348134913511352135413591365137213731376137913811383138613891390139213941397139914001401140314041405140914101411141214131414141814201423142514261428142914301431

r I

I I ~----------------------~

I

I _

I . 1

(-- - --rr7n171-- - - IIIT111)-· - ---

I 1

1 7

I

I I

Lr- - - - -1

r-- - - - ---- 1

I -· t

I-- - - - - -- 1

I 1___1

I~JI I

I------------------I

I . 1

I

I 1r_ _ _ _ _~ _ _ _

I 1

I I-- -- ;I 1

I-- - - - - -1 I

I 1

I

I 1

I I

1 1

I I

I 1

I I

~-----1

I 1

I 1

I-- - - - - --1

I 1

I I

I 1

I 1

I 1

I

I I

nnnn I 1__I 1

~rlll~rTT71TTmllTT1LIIn17/////1

I I

I-- - ------------- I

r 1

I - - - -- --1

I-- - -- - -- I

I-- - - -· - --J

I I

b I-- - - --

I - - - - - --I

I---------·--I

I 1

I I

I- - - - - -1

r

I 1

r 1

I-- - - - - --1 1

(----------------;1

I---- - - - -- I

I-- -- ;Y71/1~-- - -- I

I I

I

I I

I

-r a

Site a -

No. ' T O4 Z CIO

1434 j

144014421445144614491450 -14521454' --14571458 c1459 c1460 c1461146214631464147314761401 T1502150315081509151015131515152615281540155515631570157115741575157915831584158915901591159215941599160016011602160316051610161216131618162816301639

A 221A 259A 261A 264A 273A 275 TS

Nippur ffAdab ff

Isin E-KisurraWilaya arn

--_ 0.1-4 ha-_i 4.1-10 ha

rn 10.1-20 haTrace

r--------

I 1

~--------~

r- 1

I I

I-------1

--------------- I

I I

1 '-----1

r 1

I-- - -- I I

II II 1 r II 111111

I 1

11171T1IrI17n---------

I 1

I 1

~ILLLLIlILIlILIlIII

__ r-i

i ]

I IE--1

e 1

i i

-- -i i

2 20.1-40 ha1 40.1- 200 haiTmn 200+ ha- Max.

173

I I

I 1

I I

t-- - -1 71I1 1 I 111rl rn

I I

I-- - -- I

I I I 1

I I

I 1

I I

I 1

I I I I

I I

1/7/n 1

lTIllIlllln llITll~mIllV~1'/7//AIIIIIllll~'////////

l~TTTlITTllTnrlnTl~I7T1TTTTllnllI7Trl-T~

-- -V/////~rmlmlllmIlIrrrrrrrlrrlrlrrmnr

l~m

l~lllllll~TTrrmTlZnT7n

oi.uchicago.edu

Integration and Fragmentation under Successive, Contending Dynasties

Mesopotamian culture" at the end of the Old Babylonianperiod (1965, p. 213), and this is confirmed by the veryextensive, nearly total replacement of the inventory ofceramic types at Nippur. Many of the new types wereof course repeatedly recognized during the survey, buta single one of them was so nearly ubiquitous that inmost cases it was allowed to stand individually forCassite occupations. This is a tall chalice with a concaveneck and almost cylindrical body, tapering toward a mas-sive, solid base with a large, disk-shaped foot (McCownand Haines 1967, pl. 98:14-16; Hrouda 1977, Taf. 28;McCown, Haines, and Biggs 1978, pl. 50:6).

A second useful shape for dating purposes is apparentlya derivative of the globular jar with "button" foot ofOld Babylonian times. Two changes generally occur inthis form. Perhaps more important, the proportions ofjar body and collar change, with the latter becomingvertical or concave and much taller, in fact frequentlycontributing as much or more to total vessel height asthe body itself. In addition, the button or low disk basecontinues but often is attached to a short cylindricalpedestal rather than directly to the lower body (McCownand Haines 1967, pl. 98:1-5, 11-13; Hrouda 1977, Taf.28; McCown, Haines, and Biggs 1978, pl. 49:6, 8-9). Athird common vessel form is a low, thick-walled, care-lessly made flaring bowl with a rounded and sometimesthickened rim and frequently a string-cut base (McCown,Haines, and Biggs, pl. 49:10). Some examples have agroove or slight carination on the upper exterior, andassociated rim forms may be widely everted, beveled, orincurving (McCown and Haines 1967, pls. 97, 100;Hrouda 1977, Taf. 28). Note that while this series offorms does not antedate the Cassite period it may havecontinued for a half-millennium or more after that.

Middle Babylonian Period

The long interval from the decline of Cassite hegemonyuntil the rise of Babylonian resistance against the Assyr-ians on the eve of the formation of the Neo-Babyloniankingdom is poorly attested archaeologically. Ultimatelywe may expect that continuing excavations at Nippurand Isin not only will shed light on the vicissitudes ofurban life in a sharply declining number of political cen-

ters but will clarify the ceramic and architectural se-quences as well. In the meantime, however, this remainsa little-known intercalary period, with attitudes towardits limited cultural attainments aptly summarized bycasual reference to it among fieldworkers as the "V.D."(Various Dynasties) period. Within the framework ofsharply curtailed, perhaps often impermanent settlementand a general lack of secure ceramic stratigraphic criteriawith which to make firm attributions of what sites didcontinue in at least periodic use, the survey was forced torely largely on imprecise, inductively established criteriathat are at best highly tentative and insecure.

The problem of identifying Middle Babylonian diag-nostic criteria, at least until stratigraphically secure typo-logical series spanning the period are made available fromexcavations, is rather like the previously abumbratedproblem with the Jemdet Nasr period. Distinctive traitslimited to this period are not known, so that for a siteoccupied at this time alone one is forced to rely primarilyon the presence of certain continuing traits, togetherwith the absence of others that are directly associated withthe periods immediately preceding and following. Sitesare described herein as exclusively Middle Babylonianwhen they lack Cassite chalice fragments or any of thesomewhat less distinctive Neo-Babylonian traits, butwhen their surface materials in other respects are quitesimilar to those of Cassite sites. A subjectively perceiveddifference is that pottery manufacturing techniques be-came even more careless after the rather low standardsof the Cassite period-jars and particularly bowls weremade with thicker walls, and the rims of the latter wereprogressively thickened and rounded until they might bedescribed as having a ropelike appearance.

Commonly occurring on Middle Babylonian sites arelow tubular pot stands, with simple rounded or rope rimsand more or less constricted midsections (McCown andHaines 1967, pl. 102:20). Excavated examples are un-fortunately still too rare for us to speak with confidenceof the longevity of this type. Quite possibly it covers theentire span from Cassite through Neo-Babylonian times.The same applies to small cup bases that may be char-acterized as irregular, rodlike stumps (McCown andHaines 1967, pl. 100:21-23).

174

oi.uchicago.edu

5Culmination and Collapse

of an Agrarian Baseand Urban Superstructure(Neo-Babylonian-Late Islamic Periods)

The culminating historical epoch to which we now turnsaw a repetition and intensification of many of the cen-tral themes of the preceding one, attaining both a newlevel of economic accomplishment and its seeming antith-esis. For the longer, initial part of it there was a cumu-latively impressive, if irregularly sustained, growth inthe scale and integration of the agricultural regime. Atleast partly as a result, by the Sasanian and Early Islamicperiods, city-building, population density, and most othermanifestations of complex, differentiated social life hadattained levels transcending anything known earlier. In-strumental in these developments were cadres of admin-istrators under increasingly centralized direction. Pro-viding the improved coordination needed for widelyextended empires, they also made possible unprece-dentedly destructive and despotic exercises of state power.Linked in another, probably complementary way to thegrowing socioeconomic core were networks of commer-cial intercourse that penetrated more deeply than everbefore into the day-to-day interpersonal relations withinthe society at large.

But then, corresponding to the somewhat similar re-versal during the second millennium B.C. that was dealtwith in chapter 4, seemingly well integrated systems frag-mented before a growing array of political, economic,and ecological disorders. The basis for an urban modeof life subsequently was eliminated for centuries withinthe immediate region of the archaeological portion ofthis study, and until almost the advent of the modernera even the nonurban, sedentary population shrank backbelow any recorded point after the fifth millennium B.C.However, the curve of growth and decline-in bothepochs-was characteristically less smooth than sweep-ing summaries like this may seem to imply. At all timessuperimposed on the slow, underlying pattern of advancefollowed by deterioration was a harshly sudden, uncon-

trollable convergence of external and internal, stabilizingand destabilizing forces that produced destructive short-term oscillations of their own.

Alongside this similarity between the two epochs withregard to the broad outlines of a developmental pattern,admittedly there will be found many vital differences.Technology and organizational consciousness, tending togrow accumulatively, must engender some such differ-ences. Others involve the greater dependence upon ex-ternal forces of what took place during the later of thetwo epochs. Partly, this is a matter of a significant changein the scale of the key units of study; city-states becamesubmerged in the large regions around them. But it isalso a matter of more extensive as well as intensive pat-terns of interaction that pay less respect even to regionalboundaries. To be sure, the Mesopotamian city-states ofthe third and second millennia B.C. likewise cannot beunderstood as truly isolated in their processes of growthand change; this was the meaning of the rejection ofthe zero-sum game metaphor in chapter 4. But the fabricof interactions across local, regional, and all other tradi-tional barriers now became perceptibly denser, and thelessons that can be drawn by concentrating exclusivelyon a sharply defined geographic focus of study becomecorrespondingly fewer and more qualified.

For much of this new epoch the peoples of the lowerMesopotamian plain were more or less firmly incorpo-rated in larger, longer-lived, more heterogeneous empiresthan had existed previously. They usually provided adisproportionately large part of state revenues becauseof the scale and productiveness of Mesopotamian irriga-tion agriculture. But for the most part they did not furnishthe consolidating impulse or determine the main insti-tutional forms that the empires took. Lying within rangeof increasingly powerful and sophisticated military forcesfrom Anatolia and the Mediterranean littoral as well as

175

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

from Iran, the occupants of the alluvium were repeatedlysubject to destructive attack. Great power rivalries beyondtheir control, in other words, more and more frequentlydictated the oppressive tax, service, and other burdensthat the indigenous Babylonian population was forcedto bear.

We must assume that there were two interacting eco-nomic elements involved in the subordination of Baby-lonian autonomy within larger, imperial frameworks.One was the relative productivity of irrigation agricul-ture, disregarding local and temporal fluctuations andconsidering it only in the aggregate. When means werefound to suppress internal urban rivalies while alsoholding off external predators, this permitted rulers toamass reserves larger than in any neighboring region.With its networks of rivers and canals permitting rapid,economical concentration of these reserves by waterbornetransport, Babylonia could be made to serve a new strate-gic role as the granary of much more extensive realms.Its permanent pacification was vital for this purpose,even though the more peripheral provinces coulq be muchmore loosely knit together by more or less centrallysupported armies. Insofar as we think of Babylonia'sagricultural surpluses as the foundation for precariouslymaintained but often long-lived polities of this new kind,our attention is directed to what must have been in-tensifying networks of local or regional production orexchange. The difficulty is that such networks can as yetbe discerned only very dimly and at long intervals, sincethe major interests of historians and archaeologists havebeen directed elsewhere.

The second subordinating element has tended to re-ceive much more attention. Its real importance is some-what difficult to judge, since most of the evidence thatis adduced derives from the circum-Mediterranean worldrather than from Babylonia itself. But unquestionablythere was a massive new preoccupation with long-distancetrade, ultimately binding the western Mediterraneanworld to India and even China. The heavy drain of Medi-terranean gold to the farther Orient for spices, silk, andother luxury products is thought to have influenced Ro-man expansionist policy, of which Mesopotamia was re-peatedly a victim (Nodelman 1960, p. 109). Commerce-dependent states like Palmyra, Petra, and Gerrha soughtto monopolize control of the transit traffic for their ownadvantage. The resulting moves and countermoves formtoo far-flung and complex a sequence to be gone intohere, but their net effect was certainly not to enhanceeither the independence or the well-being of the Mesopo-tamian countryman.

The growth of the transit traffic and the emergence ofthe new, desert-based kingdoms are interdependent. Un-derlying both, surely, was the spread of the domesticcamel, for without a reliable, economical mode of long-distance transport under arid conditions the erratic and

costly conditions of supply probably would have divertedthe growth of demand in the Mediterranean world inother directions. But beyond the often considerable in-fluence that the new polities had even within Mesopo-tamia proper was a yet more profound shift. The camelfor the first time made highly mobile but dispersed groupsof desert tribesmen militarily formidable. Politically andethnically stable and insignificant through the secondmillennia B.C., the southwestern border of the Tigris-Euphrates alluvium accordingly had already begun tobe a source of incursions and disruptions in Neo-Assyriantimes. Much later, the profound transformation effectedby the Muslim conquest would have been unthinkablewithout the camel as the primary means of the Arabarmies' supply, rapid mobilization, and successful attackon larger Sasanian armies.

Yet the massive conversion to Islam that followedonly after a delay of several centuries, within the alluviumas elsewhere, was not in any sense the first substantialabandonment of indigenous Mesopotamian traditions inmatters of religion. Syncretistic elements drawn fromother areas of course were present much earlier. But itis significant for the more general decline in Mesopo-tamia's cultural distinctiveness or salience-a processwhose onset long antedated Islam and the growth of de-cisive military pressures based outside the fertile, denselypopulated river valleys-that by the Sasanian period in-digenous beliefs had largely lost their institutional ex-pression. What the Arabs encountered instead was asociety compartmentalized into self-enclosed Christian,Jewish, and Zoroastrian "minorities," all of whose im-mediate religious antecedents lay primarily in otherregions.

The sources for a study of settlement, land use, andagriculture during the periods of concern here are asheterogeneous as this sketch of a gradually submergedautonomy implies. Archaeological surface collectionsserve as the main point of departure. Details of the surveydata derived from them are provided in this chapter'sappendix, together with an assessment of special prob-lems affecting the data's consistency and reliability. Par-ticularly for the earlier part of the epoch, such problemsassume great importance. The Neo-Babylonian, Achae-menian, and Seleucid periods are very poorly defined interms of archaeological criteria that can assist in ac-curately dating surface collections. Within that long spanof time, even the order of occurrence of settlement pat-terns that can be differentiated on the basis of typologi-cally distinguishable surface collections is debatable, andabsolute dates can be assigned only with wide marginsof uncertainty. Conditions are improved in every respectfor the Parthian, Sasanian, and Islamic periods, althoughcrucial ambiguities remain in the absolute dating of typo-logical sequences.

The problems and potentialities of the documentary

176

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

record are of a different order. Knowledge of cuneiformwriting on clay tablets flickered on into the first centuryA.D., but for the last few centuries was an increasinglyarcane practice that had less and less to do with day-to-day administrative and economic requirements. Impres-sive archival activities had continued through the Seleucidperiod, in considerable part carried on by temples thathad assumed complex, widely ramifying economic func-tions. Much remains unpublished, however, and institu-tionally oriented studies that would make the materialintelligible to the nonspecialist are just beginning. Atleast with respect to agricultural management, the out-standing exceptions are a group of Neo-Babylonian-Achaemenian texts from Uruk and the so-called Murashuarchive from Nippur, dating about a century later in theAchaemenian period, to which a fuller discussion will bedevoted below.

The records in Aramaic that complemented and re-placed late Babylonian unfortunately were written mainlyon perishable materials that have not survived. Hencethere is a deep lacuna in our knowledge of the Parthianperiod, compensated for only to a very modest degreeby numismatic studies and by limited, frequently ten-dentious references to conditions and campaigns in Parthiathat are to be found in Roman sources. There is someimprovement in the Sasanian period, particularly as aresult of historical compilations made only after theIslamic conquest. References to conditions of town andrural life are rare in such materials, although some con-temporary gleanings along those lines have been ex-tracted from the Babylonian Talmud. Even for the firstcentury or so of Islam, strictly contemporary records fromIraq, not subject to anachronistic emendations by laterpolemicists and copyists, are virtually absent. Later ma-terials appear in a veritable flood, although in large partwith religious, political, and narrowly urban preoccupa-tions that limit their usefulness for reconstructions of theagricultural economy and conditions facing the mass ofthe rural population who sustained it.

Thus there is a discontinuous, disjunctive quality tothe kinds of data on which this study can draw. Thegreater degree of internal continuity characteristic ofthe region in the third and second millennia B.C. unques-tionably permits a more systematic, consistent approachthan is possible for this later epoch. Were a fuller archae-ological record available, it might go some distance inrepairing these deficits. But once again most of the archae-ological fraternity still tends to link archaeological priori-ties to the acquisition of texts and the limited-primarilystate and temple-interests that literacy served, ratherthan to the most effective service that can be performedin reconstructing broader cultural patterns; hence it hasturned its attention to other periods.

Conceding difficulties, it is yet useful to begin assem-bling what is known and to suggest a few of the plausible

possibilities to which that knowledge points. As previ-ously, within the area of intensive coverage the archaeo-logical survey provides information on changes in demog-raphy and settlement distribution that can appropriatelybe summarized before we turn to the less uniform butmuch more richly detailed textual materials. We cannotassume, of course, that this information adequately repre-sents the whole of the alluvial land surface. None of themajor power centers of the time lay within the areafor which intensive coverage is available, although theenvirons of Babylon, Seleucia, and Ctesiphon have beenreconnoitered more selectively. The presence of thosecenters suggests that the northern part of the plain mayhave been more densely occupied than the region forwhich at least an approach to quantitative treatment isnow possible. The northern part was also, at least inParthian times, more exposed to the destruction accom-panying periodic Roman invasions. But I will postponefurther discussion of the representativeness of the avail-able data, and of geographic contrasts and historical de-velopments that bear on the question, in order to beginby providing a clearer picture of the scale of settlementin at least one region. Parenthetically, a separation be-tween settlements and the irrigation systems that sus-tained them is obviously somewhat artificial. Neverthe-less, I will also withhold comment on the latter until alater section of this chapter, when textual as well asarchaeological data on agriculture can be brought to-gether to provide a more integrated treatment of thesubject.

AN OVERVIEW OF LOCAL DEMOGRAPHICAND SETTLEMENT TRENDS

In spite of the uncertainties I outlined earlier, there isno doubt about the rapid, continued growth that gotunder way during-or perhaps even slightly before-theNeo-Babylonian period. This is most simply shown bythe rising numbers of sites as furnished in table 20 inthe appendix to this chapter. The total increases from134 in the Middle Babylonian period to 182 (excludingquestionable or "trace" occupations) in the Neo-Baby-lonian period, to 221 of Achaemenian date, and to 415that are apparently Seleucid-Parthian. If we combine Neo-Babylonian and Achaemenian on the grounds that col-lections assigned to the two may not be readily distin-guishable (below, pp. 228-31), the joint interval wouldbe credited with 257 sites.

We may be justified in assuming that one componentof this growth was the natural increase of local popula-tion, under conditions of relative prosperity and tran-quillity imposed by strong dynasties. It is also likely thata large proportion of the Chaldeans and Arameans whohad been forced into exile during the later years of Neo-Assyrian rule afterward drifted back to their former

177

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

homes in southern Babylonia, perhaps accompanied byother elements left rootless in the Assyrian homeland afterthe fall of Nineveh. But there were in addition numerousJews and other groups brought into Babylonia after thewestern conquests of Nebuchadnezzar II (604-562 B.c.).The available documentary evidence for the latter clearly(although indirectly) suggests that large masses of peoplewere involuntarily transferred as part of intensive Neo-Babylonian efforts to rehabilitate the central region of adomain that previously had suffered severely (Eph'al1978, pp. 81-82).

The extent of growth is even more striking if we con-sider areas of occupation, and thus implicitly population,rather than numbers alone. The gross Middle Babylonianoccupied site area (table 13) was 616 hectares, giving anaverage site size of only 4.6 hectares. Calculating on thesame basis for the combined Neo-Babylonian and Achae-menian periods, it was 1,769 hectares, or an average of6.88 hectares, and for the Seleucid-Parthian period 3,201hectares, or an average of 7.71 hectares. Thus the incre-ment in average site size (most rapid in the Neo-Baby-lonian-Achaemenian times), when combined with theincreasing numbers of sites, suggests a growth in thepopulation of the region by more than five times duringa span of five to seven hundred years. As the appendix tothis chapter specifies, to be sure, population density mayhave been somewhat lower within the sprawling, partlydiscontinuous sites of the Parthian period than withinearlier, more nucleated towns, but correction for thisfactor would lead to only a very modest lowering of whatis still an impressive rate of increase.

To speak of average site size may be somewhat mis-leading since it tends to imply distribution around a mode.The reality, as we have seen for earlier periods, is thatsettlements formed a hierarchy of different sizes. Usingthe categories already applied in chapter 4, table 15provides the distribution for the periods indicated.

TABLE 15 Numbers of Sites by Size Category and Period ofOccupation, ca. 700 B.C.-A.D. 200

Neo-Middle Babylonian Seleucid-

Size Category Babylonian Achaemenian Parthian

6 (200+ hectares) - - 25 (40.1-200 ha) 1 5 44 (20.1-40 ha) 3 6 153 (10.1-20 ha) 2 19 342 (4.1-10 ha) 28 70 951 (0.1-4 ha) 100 157 265

What emerges most clearly from this tabulation isthat while the numbers of sites in all categories increased

dramatically, the rate of increase was significantly greaterfor large than for small sites. Average site size increased,in other words, because the proportion of large sitesincreased over time rather than because the number ofsmall sites declined or became static. The point is perhapsclearer if we distinguish a broadly "urban" from a "non-urban" category. Applying the-admittedly arbitrary-criterion developed in chapter 3, occupations larger than10 hectares may be tentatively designated as "urban."On this basis, only 36 percent of the occupied site areain Middle Babylonian times was found in settlementslarge enough to be characterized as urban. Combiningthe Neo-Babylonian and Achaemenian collections, thisfigure had by then increased to 51 percent. By the Seleucid-Parthian period it had risen slightly further still, to 55percent. This was almost exactly the same proportion asduring the Third Dynasty of Ur, completing the reversalof a devolutionary trend that had continued for almosta millennium and a half.

Whereas there was about a threefold overall increasein the number of sites between the onset of the Neo-Babylonian expansion and the Seleucid-Parthian period,the above tabulation indicates more than a ninefold in-crease in the number of urban sites as here defined. More-over, the majority of the urban settlements do not appearto have been already existing smaller communities thatwere merely expanded. Of the 29 recorded jointly forthe Neo-Babylonian and Achaemenian periods, 21 pro-vided no trace of a Middle Babylonian occupation. Evenfor the slower expansion in the number of urban sitesbetween that time and the Seleucid-Parthian period, aslight majority (28 out of 55) seem to have been newlyfounded. It appears, therefore, that we are dealing withfairly abrupt, probably state-directed, policies of settle-ment formation in the case of the urban communities.

Some comparative perspective on this process of re-settlement can be gained from a consideration of theconcurrent sequence of development on the lower Diyalaplains. The latter region had reached an extremely lowlevel of population density during the unsettled condi-tions of the late second and early first millennia B.C. Butthe resurgence that followed was even more substantialand rapid than on the central Euphrates floodplain (seetable 16).

As might be expected in view of its original positionof greater exposure to Assyrian and Elamite incursions,the Diyala region appears to have undergone a wider andmore rapid oscillation between abandonment and re-settlement. The sixfold growth in the number of sites ofall sizes between Middle Babylonian and Parthian timeswas approximately twice as large as that occurring con-temporaneously in the more central area between theTigris and the Euphrates. By the Parthian period theproportion of Diyala settlement that may be defined as

178

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

TABLE 16 Comparative Data on Lower Diyala SettlementDistributions, Middle Babylonian-Parthian Pe-riods

Neo-Middle Babylonian- Seleucid-

Size Categories Babylonian Achaemenian Parthian

6 (200+ hectares) - - -

5 (40.1-200 ha) - - 84 (20.1-40 ha) - - 33 (10.1-20 ha) - - 152 (4.1-10 ha) 1 7 381 (0.1-4 ha) 33 74 141

Source: Adams 1965, tables 15-18, slightly corrected to conformwith site register entries in Appendix C. The originally separate Neo-Babylonian and Achaemenian tabulations have been combined. Al-though its position at the time cannot be plotted in detail (see below,p. 225), the Tigris is assumed to have followed approximately itsknown medieval course through the 'Ukbara or Ishaqi district northof Baghdad (Adams 1965, p. 91), rather than the line of its modernbed farther to the east. Hence this table also includes a few sitesrecorded in the Ishaqi area (Adams 1972) that appear likely to havebeen situated on the Diyala side of the Tigris.

urban was only slightly 'smaller than its midalluviumcounterpart, even though as late as Achaemenian timesit is not possible to identify a single urban site in theDiyala survey data.' For the whole alluvium, the im-pression of demographic trends that emerges-applicable,incidentally, to periods other than this one-thus is oneof relatively greater continuity and a somewhat dampenedcycle of growth in the core area, while the opposite heldtrue on the peripheries. Yet we have seen that even inthe core the extent of the transformation by the Seleucid-Parthian period was such as to introduce the distinctpossibility of state-initiated schemes of urbanization andagricultural expansion.

There is one curious feature of the changing numbersof sites in different size categories for which no explan-ation is readily apparent. It emerges not from the above,very condensed tabulation but from the site-by-site rec-ords summarized in table 16. If we consider the Parthiansites, a higher proportion of those that were "ui"ban" (20out of 55, or 36 percent) than "nonurban" (68 out of360, or 19 percent) continued to be occupied in Sasaniantimes on the same or an enlarged scale. That differenceis not surprising; one would expect larger communitiesto exhibit a significantly greater continuity of occupancy.But this turns out not to be the case if we examine Neo-Babylonian or Achaemenian sites. The proportions arereversed: Only 6 out of 28 (21 percent) of the "urban"sites continued into the Seleucid-Parthian period, but 90out of 266 (34 percent) of the "nonurban" ones appar-ently did so. Why should there have been a significantlyhigher mortality among the larger settlements than among

the smaller ones after the Achaemenian period? Onecould perhaps speculate about the purported associationof Hellenism with the founding of new cities, but theGreek cities that can be identified were almost certainlyof far greater size than the "urban" category as it is con-sidered here. For the present, this unexpected reversal ofpattern can only be set aside as an unexplained-butsurely historically significant-anomaly.

The earlier trends in population and urban growththat have been described continued through most of theSasanian period (A.D. 226-637). This was the apogee ofthe premodern development of the region in every eco-nomic as well as demographic respect, although we shallconcentrate here only on population and urban variables.Once again, the basic data on the Sasanian settlementsystem within the central Euphrates floodplain is pro-vided by archaeological survey. Because the ceramicchronological markers of the period are comparativelywell understood (even if the absolute chronology still con-tains many ambiguities), a somewhat fuller quantitativespecification of the extent of the system is possible thanfor most preceding periods. Aiding in this objective isthe fact that Sasanian remains are both relatively recentand more extensive than those for any other period, forthis means that their outer limits are rarely concealed byan overburden of later occupational debris. Table 17accordingly provides estimates of individual site areasnot merely in fairly broad size categories but in hectares.

For readier comparison with the conditions attainedalready by the Parthian period, however, it may be usefulinitially to summarize the Sasanian data in terms of thesame categories used earlier. Recall that both in chapter4 and heretofore in chapter 5 rough population estimateshave been derived merely by taking the approximate mid-points of the size categories and regarding these as aver-ages that could be applied to the categories as aggregates.This is also done in table 17 for purposes of comparison,but with the totals based on individual site computationsprovided in the adjoining column.

The averaged areas in most cases are tolerably close

TABLE 17 Numbers and Occupied Areas of Sasanian Sitesby Size Category

Averaged MeasuredNumber of Occupational Occupational

Size Category Sites Area Area

6 (200+ hectares) 1 200 2305 (40.1-200 ha) 11 1,100 710.94 (20.1-40 ha) 22 660 586.93 (10.1-20 ha) 46 690 655.62 (4.1-10 ha) 141 987 898.81 (0.1-4 ha) 376 752 709.5

Totals 597 4,389 ha 3,791.7 ha

179

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

to the measured ones. Category 5 is the significant ex-ception; major urban centers at least in this region andtime period apparently were all grouped in the lowerportion of the range. Even with this substantial dis-crepancy the sum of the differences is only about a sixthof the measured total. To compare aggregate settlementand population with that of the Parthian period, sinceno Parthian counterpart of the individually measuredSasanian site areas is available, the averaged total mustbe used for both periods. As I noted earlier, the figureof 3,201 hectares in the Parthian case is the one to becompared with 4,389 hectares in the Sasanian period.Thus there had been an increase of approximately 37percent after the Parthian period. When we bear in mindthat there had been extensive local abandonment of thesouthern part of the alluvium by late Parthian or earlySasanian times in connection with the apparent spreadof swampy conditions there, including a major declinein one of the region's two largest urban centers (Uruk),the proportion of this increase is even more impressive.

A comparison with the maximum extent of settlementunder the much earlier conditions outlined in chapter 4is also instructive. Recall that the area of built-up settle-ment during the Ur III-Isin-Larsa period was 2,725 hec-tares, calculated on the same averaged basis as was em-ployed in the center column above. Sites of the third andsecond millennia B.C. give the impression of being morenucleated than Parthian and Sasanian sites, possibly be-cause their growth was often constrained by fortificationwalls, whereas in later periods defensive works werepractical only for the larger cities. If this is so, whateverthe reason, a direct comparison between the demographi-cally and economically most advanced of the early periodsand the apogee of settlement in the early Middle Ages isvery difficult. On the one hand, equivalent sites may havehoused somewhat greater populations during the earlierperiod. On the other hand, the earlier period is on thewhole more likely to conflate successive developmentsunder different political auspices, so that it may sig-nificantly exaggerate the population at any one time.Moreover, the surveyed region was in the very heart ofSumer in the earlier period, so that findings on populationdensity there are likely to be near the maxima for thetime. For the Sasanian period, however, the surveyedregion was remote from any centers of even regionalpolitical importance.

Taking all these considerations into account, signifi-cantly greater population densities were to be found inthe Sasanian period than during the Third Dynasty ofUr, more or less corresponding to the contrast betweenthe suggested maximum settled areas of 2,725 hectaresand 4,389 hectares-61 percent greater. At the most, onemight want to accept the more accurately measured figureof about 3,800 hectares for the later period while leavingthat for the period around 2000 B.C. unchanged, still in-

dicating a very substantial growth of population withina surveyed area of identical size. The effect of this in-crease in density was surely to require something closelyapproaching continuous cultivation over wide areas, andthe computer simulation program originally worked outfor prehistoric data in Appendix B to chapter 3 providesa graphic illustration of this is figure 36. Based on averagesite densities of 100 persons per hectare (as calculatedfrom the rectangular coordinates of site measurement)and average requirements of 1.5 hectares of land perperson, it converts recorded Sasanian site distributionsand areas into a representation of the cultivated zoneoptimally located for the sites in question.

There is an exceptionally close correspondence betweenthe limits of this simulated zone and the limits of thesurveyed area itself, indicating that virtually continuouscultivation within large parts of the latter would havebeen necessary. The exceptions include a west centraldistrict (centering on the long-abandoned city of Isin)and one in the south (centering on Larsa, also long aban-doned by the Sasanian period). The latter had been givenover to swamps by this time, but in the former weshall find presently that there was a massive (if probablyshort-lived) Sasanian attempt at reclamation (below,p. 210).

The implication of figure 36 is that the uniform darkarea would be extended in most directions, if conditionspermitted surveys of intensity comparable to the presentone to proceed in those directions. This underlines thereal difference between the Sasanian period and earlierantiquity. As we shall see presently, it lay not so muchin an early medieval increase in density as in an immenseexpansion of the cultivated area so that the total popu-lation of the alluvium was significantly larger.

Characteristic of the central Euphrates floodplain atthis time was its declining access to the increasinglycentralized levers of imperial power. This suggests thata different kind of comparison also may be instructive,with the Diyala plains immediately adjacent to the seatof Sasanian authority at Ctesiphon. Sasanian surfaceceramics were not so readily identified at the time of theDiyala study as in the present one, and the need to con-duct much of the earlier study in cultivated areas furtheraffected the completeness of its findings. These differencesprobably cause a disproportionate reduction in the re-corded number of smaller sites, slightly skewing the per-ceived settlement pattern in the direction of a greaterapparent degree of urbanization. But, even taking thisinto account, the contrast between the urban hierarchiesin the two regions is so substantial that it is surely sig-nificant. Considering only the number of settlements ineach size category specified in table 17, the correspondingdistribution for the Diyala region is as follows (FromAdams 1965, table 18, with slight corrections from siteregister):

180

oi.uchicago.edu

0

0*

ZibliyatZibliyat

Jidr/

S.

1

0

Ruqba Medain

Fig. 36. Simulation of Sasanian cultivated area within the intensively surveyed region.

181

c

o 4 qo

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

6 (200 + hectares)5 (40.1-200 ha)4 (20.1-40 ha)3 (10.1-20 ha)2 (4.1-10 ha)1 (0.1-4 ha)

2126

3169

174

The two topmost categories are slightly better repre-sented on the Diyala plains than in the midalluvial re-gion, while settlements falling among the lower fourare more than twice as numerous as on the centralEuphrates floodplain. Proximity to the political centerthus was accompanied not merely by a higher degree ofsettlement nucleation but by marked population agglo-meration in the larger urban categories at the expense ofsmaller urban centers, towns, and villages. It is quiteuncertain whether average regional population densityalso was appreciably higher in the immediate hinterlandsof Ctesiphon. Differences in survey procedures and effec-tiveness obscure this question, but in addition the land-use simulation given above appears to argue negatively.With so much of the land surface already under cultiva-

tion, appreciably larger populations could have beensupported only if more intensive cropping systems werein use around the capital-or if there were a large netinflow of food from fairly distant regions. But a compari-son of the lower Diyala and the midalluvium regionsdoes support the existence, at least under the conditionsobtaining in the Sasanian period, of a strong positiverelationship between the degree of population concentra-tion in large cities and political centrality.

The data in table 21 are graphically summarized infigure 37, a conventional histogram indicating site dis-tribution by size for the intensively surveyed area duringthe Sasanian period. In spite of the great numbers ofsmaller settlements, only 42 percent of the total occupiedarea lay in sites of 10 hectares or less that are here desig-nated "nonurban." The arbitrariness of that definition ismade apparent by the absence of clear breaks in the long"tail" extending to the right on the histogram until amuch higher figure is reached. The first broad gapstrongly suggesting a meaningful discontinuity, in fact,is that between 35 and 45 hectares. Size alone is a quiteinadequate basis for distinguishing functionally differ-

380

360

140

120

100

80

60

40-

20-

0-

50 60 70 80 160

Upper Limit of Size Categories in Hectares

Fig. 37. Size distribution of Sasanian sites.

182

120-

100

so

80"

60

E

40-

20

0

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

entiated tiers in an urban hierarchy, but at present thereis no other information available with which to supple-ment (or replace) this criterion.

The very broad size distribution of the larger sitescomposing this "tail," surely most or all of them urbanby any reasonable definition, is perhaps the most impres-sive feature of the Sasanian site hierarchy. It suggests thevigorous, strongly differentiated development of townlife throughout the region, rather than a pattern of con-centration around one or a few major centers of power.That description recalls our earlier discussion of urban"primacy," a condition in which Uruk, then the largestcity, substantially exceeded its normal place in the rank-size logarithmic succession of cities constituting a singlesystem (above, pp. 72-75).

A rank-size distribution of the Sasanian data is plottedin figure 38. The top of the graph may roughly approxi-mate a normal distribution, but the bulk of it, applyingto all but a few of the largest settlements, is convex andthus is the opposite of a condition of urban primacy. Torepeat the characterization of the "flattop" morphologyof this curve that was given earlier, it is "indicative ofvery imperfect integration at the level in question" (Skin-ner 1977, p. 241). The proliferation of smaller urbancenters in the histogram thus should be understood asbeing only very loosely connected with the major cen-ters, as if they were largely unaffected by centralized im-

perial policy. I will show presently that this is entirelyconsistent with what is known of the regional economicand administrative history of the Sasanian period.

The maximal development of these patterns has beendescribed merely as Sasanian, but it would be misleadingto characterize the whole period of the dynasty as one inwhich these maximum limits were consistently ap-proached. Insofar as the surface collections allow us todifferentiate early and late subperiods, the latter areclearly preponderant throughout the intensively surveyedregion. On historical grounds, too, there is reason toplace the fullest growth of Sasanian urban settlement-and the agricultural system that sustained it-in the earlyto mid-sixth century A.D. Then in the late sixth and earlyseventh centuries there probably was a sharp reversal,although this cannot yet be detected in the findings ofsurface reconnaissance. The full Sasanian achievement,in other words, is likely to have been of strictly limitedduration. Moreover, many elements in the ensuing longdecline almost certainly antedated the Islamic conquests.Although obviously not independent of political, reli-gious, and ethnic factors, the cycles that are discerniblein the waxing and waning of settlements do not conformclosely with the traditional chronological divisions as-signed to successive empires.

Although the onset of the reaction probably precededthe Islamic period, it is in the succession of Isla ic set-tlement patterns that we find the most compelling evi-dence for an exceptionally broad and deep reversal. Theceramic surface collections permit a succession of phasesof settlement to be identified fairly unambiguously. Asis discussed more fully in the appendix to this chapter,there is much less certainty about the absolute dating ofthese phases. Perhaps most difficult to identify are re-mains of the first century or so after the conquest. By themid- or later eighth century A.D., a variety of new glazedwares were making their appearance, attesting occupa-tions during the heyday of the 'Abbasid Caliphate. Per-haps to be considered also under the latest part of the"Early Islamic" rubric are splash-glazed "classic" sgraf-fiato styles. Usually associated with the brief floruit ofSamarra as the 'Abbasid capital (A.D. 836-92), smallnumbers of sherds of this type distinguish a subset ofEarly Islamic sites that may have continued for sometime into, and probably even through, the tenth century.Subsequent to this, corresponding roughly with the late'Abbasid period, is a phase here designated as MiddleIslamic. Finally, Late Islamic is defined for present pur-poses as beginning with glazed styles associated with theIlkhanid period (A.D. 1258-1410).

Table 22 in the appendix to this chapter details theoccupations that may be assigned to these successivephases site-by-site. The summary statistics in table 18,drawn from this longer table, make the astonishing ex-tent of the decline more apparent. As is explained in the

183

nan

Fig. 38. Rank-size distribution of Sasanian settlements.

cc

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

TABLE 18 Numbers and Periods of Occupation of IslamicSites by Size Category

EarlyEarly Islamic + Middle Late

Size Category Sasanian Islamic "Samarran" Islamic Islamic

6 (200+ hectares) 1 - - - -5 (40.1-200 ha) 11 4 - - -4 (20.1-40 ha) 22 7 1 1 (1)3 (10.1- 20 ha) 46 22 2 4 52 (4.1-10 ha) 141 76 + (3) 13 9 + (3) 191 (0.1-4 ha) 376 173 + (73) 21 26 + (31) 52 + (19)

appendix, there is a good case for adding the fairly largegroup of provisionally attributed sites (shown in paren-theses) to the totals given in the Early Islamic column. Onthe other hand, it is doubtful if more than a handful of theprovisional sites shown in the Middle and Late Islamiccolumns may be added to the totals indicated there.

Even if one takes fully into account all the sites that areonly provisionally attributed to the Middle Islamic period,these figures show that by the eleventh or twelfth centurythe total occupational area had shrunk to only about 6percent of what it had been a half-millennium earlier.Urban life on a substantial scale of course still continuedin other parts of the Mesopotamian alluvium, but in thiscentral region only four sites, none exceeding 20 hectaresin size, were left that fall within the urban category as heredefined. The Late Islamic column suggests a modest re-surgence of settled life after the thirteenth century, butearlier trends can only be described as a prolonged, pre-cipitous decline.

Some clarification of the nature of the change may begained by comparing the Sasanian and Early Islamic col-umns, the only ones with enough entries so that a regionaltrend can be described with considerable confidence.There was an aggregate loss of about 55 percent of theSasanian settled area, but the decline was not uniform. Thereductions for each settlement size category betweenSasanian and Early Islamic times are as follows:

6 -100%5 -64%4 -68%

3 -52%2 -44%1 -35%

According to these figures there was some reductionin every size category, but losses were in general muchgreater among the larger settlements. The decline in thesmallest two categories was in fact fairly modest. It seemsto reflect little more than the loss of settled area owing tothe progressive growth of swamps in the southern part ofthe region (cf. figs. 45 and 48), while the abandonment ofthe major towns was a geographically less localized proc-ess. This may suggest that the Islamic conquest was fol-lowed by large-scale population movements away from

the region, possibly in the direction of newly foundedurban centers like Basra, Kufa, and Wasit. Generalizedattrition, whether politically or environmentally induced,seems to be ruled out as a cause by the differentiated na-ture of the abandonment. Much of the rural populationseems to have remained in place, and it appears that thepropensity of the urbanites to drift toward the new Islamicfoci of prosperity and power was roughly proportional tothe size of the urban settlements with which they werealready familiar. However, this plausible but quite specu-lative reconstruction cannot be supported by specifictextual references to the settling of large numbers of non-Arabs in the new garrison cities. Except for the extensivecolonization of the Basra region about the end of theseventh century, the sources speak of non-Arab urbanimmigrants solely in terms of defecting military units andcaptives (Morony, pers. comm.). And, unfortunately, thelimitations of archaeological surface collections as a meansfor dating abandonments makes it impossible to specifythe processes involved more exactly from the other direc-tion. A shift occurring within a few decades would havebeen so abrupt that it would hardly escape the notice ofthe chroniclers, but the shift to be observed in this in-stance could equally well have been a less perceptible onethat continued for more than a century.

In any case, as a result of the selective abandonmentthere was a shift of the balance in the direction of ruralagriculturalists. The latter, for the first time since theMiddle Babylonian period a millennium and a half earlier,emerged once more as dominant population component.After a period of no more than a few centuries duringwhich the population approximated that of the Neo-Baby-lonian and Achaemenian periods in the early stages of theupward cycle of growth, it plunged to a level of less thanhalf what it had been even during its Middle Babyloniannadir.

We must bear in mind, however, that the data thus farmarshaled in support of an irregular but progressive with-drawal apply only to the old urban heartland in the centralEuphrates floodplain. The question is, How representativeis this of trends elsewhere? There is no doubt of a generalfall in population throughout the entire region if we ex-tend the temporal scope of the discussion to include thelater years of 'Abbasid misrule and the destructive Ilkha-nid onslaught that followed. But this already implies thatin some districts the decline proceeded much more slowlythan in others, and that it may have been completely ar-rested or even reversed for considerable periods. Such isparticularly likely to have happened around the newlyfounded Arab cities, and especially in the sustaining areaon which after the mid-eighth century the growth ofBaghdad must have largely depended.

Yet even around Baghdad the immediate impact of theIslamic conquest seems to have been a considerable dis-ruption of settlement patterns and an associated decline in

184

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

population. At least for the lower Diyala plains to thenorth and east, the transition between Sasanian and EarlyIslamic times was accompanied by the following changesin number of sites in each settlement size category (FromAdams 1965, tables 19-20, including only pre-Samarransites from the latter, slightly corrected to conform withsite register entries):

6 -50%5 -58%4 +12%

3 -26%2 -41%1 -35%

There are unexplained anomalies if this tabulation iscompared with the foregoing one. Smaller urban centers,those in categories 3 and 4, either fell behind the generaltrend or reversed it completely.2 But it is fairly clear thatthe large Sasanian centers were once again in this casemore vulnerable than the small settlements presumablydevoted largely to agriculture, although the differenceswere less pronounced than they were farther south. Com-parison of the table and the map for the period (fig. 51)suggests, furthermore, that much of the reduction in thenumber of small settlements derives not from a generalprocess of widespread attrition but from localized pocketsof abandonment. Granting the incompleteness of our dataand its probable unrepresentativeness as a detailed ac-count applicable to the whole alluvium, the rough simi-larity of trends in two large, differently situated regionsdoes argue strongly for a more or less massive demo-graphic decline during the half-millennium or so after thebeginning of the Islamic period.

To recapitulate briefly, the settlement data to whichthis section has been devoted trace out a curve thatascends to a climax in the late Sasanian period and thenfails more abruptly to the lowest level in almost five mil-lennia. To be sure, there is ample reason to doubt thesmoothness of the cycle. Doubtless the broad periodiza-tion we are forced to use obscures many short-term oscil-lations related to conjunctures of ecological, political, andeconomic events and policies. These may well have beenthe only reality of which the population at large had anyconscious awareness. But the longer cycle was no lessreal for lack of popular recognition. It encompassed notonly a burgeoning and then catastrophically decliningpopulation but a cycle of urban growth and abandonmentthat sensitively responded to, and even amplified, theswings in population.

The problem of the unrepresentativeness of our datamust be mentioned once again in concluding this overviewof population trends over a span of a millennium and ahalf. A comparable sequence of change for the lower Di-yala region greatly reduces the possibility of serious mis-interpretation and has been included above partly for thisreason. But a series of at least modest surprises is all butcertain until the entire alluvium has been systematically

surveyed. Note has repeatedly been taken of a tendencyfor the intensively surveyed region to become politicallymore marginal than it had been in earlier antiquity, per-haps as a consequence of the westward movement of theEuphrates away from the older cities on the central flood-plain that has been documented already in chapter 4. Italso remains to evaluate, insofar as possible, the quanti-tative implications of the survey findings with the aid ofrelevant textual sources. But the archaeological evidencefor an underlying cycle involving a sevenfold rise and evenmore complete collapse transcends what can be even po-tentially confirmed at every point by largely accidentalcongeries of texts. It also transcends the many uncertain-ties over the timing of this cycle that are associated withdependence primarily on archaeological surface collec-tions for dating. In that sense the archaeological evidenceis not ancillary to traditional epigraphic and even histori-cal studies but precedes them, providing an orientationfor interpretive efforts rather than merely assisting inthem. This, at any rate, is the conscious spirit in whichit is only candid to admit that the present synthesis isundertaken.

CHANGING AGRICULTURAL ANDFISCAL MANAGEMENT

A frustrating but unavoidable feature of employingarchaeological and textual evidence in a common frame-work is that they can seldom be made fully congruent.In this case, generalizations have emerged from an archae-ological survey that can only very hesitantly be appliedoutside its stated boundaries of intensive coverage. Archi-val sources, in the meantime, rarely if ever even touch onthe same set of demographic and settlement generaliza-tions. All too frequently, moreover, they are so laconic aswell as incomplete that it is tempting not to move very farbeyond the mass of particulars to which they are devoted.Where generalizations were ambiguously articulated in theancient sources themselves, as in some of the early Arabaccounts of fiscal and agricultural policies, their signifi-cance for the archaeologist is reduced by our inability toreproduce the geographic boundaries within which theywere thought to apply. If we are not to sink to the mereenumeration of uncoordinated bodies of data, in otherwords, we are compelled to concentrate upon a few, fairlyabstract but crucial areas of potential linkage-and thenoften to approach even those areas tentatively andindirectly.

The Neo-Babylonian and Achaemenian Periods(ca. 700-330 B.c.)

There is no ready means of avoiding these limitationsfor the earlier part of the long historical epoch dealt within this chapter. Neo-Babylonian, Achaemenian, and Se-leucid cuneiform sources, to be sure, do exist in profusion.

185

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

Many still await publication, but even those that have re-ceived comparatively detailed study are of little assistanceon rudimentary matters like the scale of the agriculturaleconomy.

The importance given to Babylonia in Achaemenianroyal titulary, and the choice of Babylon as the dynasty'swinter residence down to the time of Xerxes, surely indi-cate that great political or symbolic importance was at-tached to it during at least the initial decades of Achae-menian rule. Only a major center of production wouldhave been in a position to provide one-third of all theAsian tribute the great king received, or the unparalleledtribute of one thousand talents that Herodotus (1.192;3.93) claims was levied on it. But the currency of Baby-lonia's prominent place in the titulary outside Babyloniaitself is open to some doubt, so that this may have beenlittle more than a calculated bid for local support. Someuncertainty is also justified over whether the deliveries oftribute were regular or merely exemplary. Moreover,translating aggregates said to derive from an undemar-cated large region, only part of it archaeologically sur-veyed, into specific densities of land use from which theywere extracted by taxation or various other forms offorcible transfer is a challenge not easy to meet. For allof the generalized praise that Xenophon (Oeconomicus4.8-15) heaps on the quality of agricultural managementin what must be Babylonia, making clear that there wassome level of direct royal involvement in making agri-cultural improvements, the realities of economic life re-main elusive.

The so-called Murashu archive from Nippur allows usto approach the problem from the other end. Not im-mediately relevant are the wide-ranging economic activi-ties and institutional interrelationships of this powerfulfirm, both on its own behalf and as a subsidiary crownagent that leased and then undertook the subleasing ofland, water, seed, draft animals, and other factors of pro-duction. But much information is conveyed in passing onagricultural if not demographic questions. The difficultylies in determining the representativeness of this informa-tion. Oppenheim (n.d.) has suggested that the Achaemen-ian rulers "created ... what might be called a project ofinternal colonization in the region of Nippur, which wasto become the economic capital of the satrapy." His inter-pretation is supported by the existence in the vicinity oflarge estates belonging to members of the royal family andother high Achaemenian officials (e.g., Dandamayev 1969,p. 306). If correct, it implies that the local economy mayhave been more substantially altered as a result of royalpolicies than any other part of the Mesopotamian allu-vium except perhaps the immediate vicinity of Babylon.But so large a proportion of the available texts are fromNippur that little can be said with any confidence aboutits similarities to or differences from other regions. Stress-ing the resulting uncertainty, Oelsner (pers. comm.) finds

it inherently more probable that the position of Nippurwas essentially typical for important towns of the periodthan that it was exceptionally favored.

The quantities and varieties of produce testify to thevigor of the agricultural regime in which the Murashufirm played an important part. Barley and wheat contin-ued in their traditional major roles, but there are alsonumerous references to spelt, millet, sesame, onions, andpulses. Flax, too, may have been grown on an extensivescale, a new development suggested by references to it inMurashu tenancy contracts. As Oppenheim notes, "thisinformation must be combined with the remark in Stra-bo's Geography XVI 1, 7: 'Borsippa .. . manufactureslinen in great quantities,' all of which is in keeping withthe sometimes astonishing quantities of linen (21,600 unitscalled qatu) mentioned in the linen weaver's account Nbn.164" (1967, p. 251; note that a line is supplied here thatwas inadvertently dropped in the original publication).Figs are not attested, and there is only one reference tothe cultivation of grapes, but dates were cultivated on avery large scale. At their maximum for a single year theobligations to the firm included one for 20,000 kur (30,000hectoliters) of dates, equivalent to 350 kilograms of silver.Another title to 11,270 kur implies that the debtors hadthe usufruct of a similar number of palms, since one kurwas about a palm's average yield (Cardascia 1951, pp. 2,18). To judge from the average density of about 300palms/kur (227/hectare) within orchards near Uruk atabout the same period, their holdings would have beennot less than 50 hectares. Date orchards more than sixtimes as large were common in the Uruk region, trulymassive units of intensive land management (Cocquerillat1968, pp. 22, 32, table A).

Reflecting the new emphasis on specialized intensifica-tion and productivity were steeply increased rates of seed-ing. The most common rate in the Ur III period, a previousera in which many policies tended toward maximization,had been about 55.5 liters of seed per hectare of barley(above, p. 151). Now it moved to 112.5 liters in the Urukarea (Cocquerillat 1968, p. 28), and to 133.3 in the vicinityof Nippur (Jacobsen 1958, p. 44). Yields seem to have in-creased accordingly, from an average of less than 1,200liters per hectare during the Third Dynasty of Ur to about1,575 around Uruk and to an "expected minimum" of2,000 liters around Nippur. For these figures to be main-tained-if indeed they were maintained-close attentionto irrigation schedules would have been needed in orderto avoid excessive irrigation and the rapid onset ofsalinization.

And there were developments that may be related toimproved irrigation controls as well as to the increasinglyintensive character of the entire agricultural regime. Al-though the details are obscure, according to Oppenheim,"the fact that two harvests are reported indicates the sizeand the reliability of the water supply provided by the new

186

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

irrigation installations and, still more important, by newtechniques. The use of novel terms such as GIS.APIN andSilihtu 'sluice gate' may indicate technical innovations thatmade better utilization of the available water possible"(Oppenheim, n.d.; cf. Cardascia 1966).

Water was one of the more expensive factors of pro-duction in the Murashu firm's operations. In one case agroup of landowners are reported to have had to give athird of their crop as well as moderate payments in silverfor the use of water from a royal canal for only threedays a month (Dandamayev 1969, p. 307). Land wasacquired from a variety of property holders, but waterrights came exclusively from a distinct branch of the royaladministration. Based upon a comparison of leases withand without water rights, it has been suggested that an-nual rental fees generally were as much as 70 percenthigher in the former case. Canal frontage may, in fact,have been the significant entry in royally maintained landregistries, in which military service was a condition of thetenure of fiefs, and royal involvement in the irrigationsystem is also suggested by the naming of judges andfiscal officials who were responsible for canal districts(Stolper 1974, pp. 44-45, 49, 61). Clearly, the crownrecognized not only its potential monopoly in this areabut a valuable source of administrative leverage, and ithad not been slow in moving to take advantage of both.What is less clear, probably because the records of the irri-gation administration have never been recovered, is theextent to which the crown itself was responsible for theimprovements in the system rather than merely benefit-ing from them by right of conquest.

This uncertainty extends beyond the leasing of waterrights, for there is something perplexing about the rela-tive costs of all the factors of production. Land was ap-parently the least of them, although generalization is madedifficult because leasing contracts fail to deal with grada-tions of quality-and frequently also to specify area aswell as price. Its annual costs appear to range between v2kur and 1 kur (150 liters) of barley per kur (1.324 hectare)of field devoted to that crop. At the most, this is consid-erably less than 10 percent of the amount that wouldnormally be harvested. Date orchards, similarly, appar-ently were priced at very low levels. In an example fromNippur the annual rent in silver on a kur of orchard wasonly fifteen shekels, the average value of the produce ofonly fifteen trees when three hundred trees were to beexpected in that area. Even at the upper limit thus farknown, an example from Uruk, the average produce thatmight be expected from only forty palms covered therental of an orchard with three hundred trees (Stolper1974, pp. 191-93; Cardascia 1951, p. 2; Cocquerillat1968, pp. 32, 38).

By contrast, the rental of the draft animals and plowsand harness with which to work the land was exceedinglyexpensive. The annual cost of an unequipped ox ranged

from 10 to 25 kur of barley, that of a fully equipped ox upto 37.5 kur. Since a sixth century Uruk text shows that 25plows and 100 oxen were needed to work 625 kur of land,the cost to the tenant of plows and draft animals usuallyranged upward from four times the rental of the landitself. Relative to these seemingly "supplementary"charges for water, oxen, and equipment, Stolper is cer-tainly right in terming the basic cost of land "so low as tobe negligible" (1974, pp. 194-95; Cardascia 1951, p. 136;Cocquerillat 1968, p. 42).

What again remains unexplained is the basis for rela-tive restriction of access, to draft animals as well as towater, that maintained this pricing structure for a consid-erable period. The feudatories of the crown as well asother landowners were placed in the position of derivingincome only from the least valuable, almost "negligible"factor of production, while the crown as well as the Mu-rashu firm was able to benefit from the whole repertoireof them (Stolper 1974, p. 200). Cardascia's and Oppen-heim's speculation that GIS.APIN refers to irrigationmachinery would offer an attractive solution to the prob-lem. With the spread of some new lifting device, the de-mand for draft animals might well outrun the supply atleast for a time. But whether the term has this meaning orinstead merely refers to some type of plow is still a matterof debate.

Stolper's argument is compelling, that "the evidentcheapness of land in the Murashu texts can be taken toreflect its availability in increasing supply; while the highcost of canals, livestock, and equipment may be held toreflect competition for scarce factors needed to exploitand develop new lands." The Neo-Babylonian and Achae-menian periods were, in other words, "times of slow butdistinct expansion of population and resettlement of long-abandoned territories" (Stolper 1974, p. 199). But someaccount must also be taken of a factor of coercion-thatis, that relative costs were administratively imposed ratherthan arising from the free play of supply and demand insomething approximating a "market." Without someother, as yet unrecognized set of constraints, for example,it is not clear why low land rentals should have beenaccompanied by the striking increase in seeding rates thathas already been mentioned. If the impetus lay with thestate, was the maintenance of this price structure some-how related to the crown's well-known propensity toamass a hoard of silver and other treasures for larger im-perial purposes (cf. Altheim and Stiehl 1963, 1:120)? Ifinstead (or also-the alternatives are not mutually ex-clusive) the impetus lay with local landowners, was theirobjective possibly to economize on the use of expensivedraft animals? Maintenance of existing, extensive ratherthan intensive, cultivation practices by renting additionallands at very low cost would have been the obvious al-ternative. Even water costs would not necessarily have in-creased if the same canal frontage was maintained and the

187

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

depth of cultivation was merely extended farther down thelevee backslopes.

Water costs and management are perhaps the key tothe riddle. If they were linked by the state to canal front-age, then expansion into new areas along long-disusedlevees incurred steeply rising costs for water. Deepeningthe belt of existing cultivation was also possible, but thecloser the farmer approached the backslope depressionsthe greater were his losses from waterlogging and salinity.By hugging the crests of the existing levees already servedby canals, increasing his seeding rates, and rationalizinghis methods and schedules for the application of irriga-tion water, he maximized his production under the exist-ing institutional arrangements.

What I am groping toward, but admittedly do not havethe evidence to reach at this juncture, is an applicationof the concept of marginal utility. This is not at all todeny that the exceptionally low price of land must be tiedin some way to a "surplus" of it and hence to relativelylow levels of population. But is it not possible that theMurashu firm generally dealt with lands that were onlymarginally suitable for new agricultural investment, eitherbecause of the high cost of development and water or be-cause of increasing salinity? If so, it would be a mistaketo generalize from the Murashu records, valuable as theyare, to the conditions facing the society as a whole.

The tantalizing lack of correspondence between even awell-studied, highly illuminating archive like this one andthe broad demographic and settlement concerns of anarchaeological reconnaissance is thus emphasized onceagain. There is a broad but vague implication in the docu-ments of rising population levels that were still well belowthe carrying capacity of the land, and of an orientation ofstate policies as well as of agriculturalists' decisionstoward an expansion of production not only through in-tensification of the agricultural regime but through its ter-ritorial expansion. Numerous canals are mentioned byname, together with their adjoining settlements (Cardas-cia 1951, pp. 2-3), some of them obviously those that havebeen merely identified by number in the archaeologicalsurvey. But names and numbers are exceedingly difficultto align with one another on the basis of the availableevidence, and they are likely to remain so for as long asexcavators avoid the smaller, less "promising" moundsand confine their attention to a handful of traditional cen-ters like Nippur and Uruk. Many activities and relation-ships that would be essential to a real understanding ofthe system, above all those concerning the coercive pow-ers and prerogatives of the state, are naturally beyond thereach of the archaeologist alone but are even left unspeci-fied and hence elusive in the texts.

This is not to deny the value of certain general cor-respondences between the textual and the archaeologicalpictures. Figure 39 sets forward the latter for the inten-sively surveyed area, and a comparison of Neo-Babylo-

nian-Achaemenian with Middle Babylonian (fig. 35) set-tlement confirms that the earlier enclaves had undergonelateral extensions in virtually every direction. Similarly,the converging line of evidence for a rising populationduring these periods that derives from surface reconnais-sance has already been presented in the preceding section,although with serious reservations about the absolutechronology of the archaeological observations that consid-erably reduce its confirmatory relevance.

Another suggestive element of correspondence con-cerns the general pattern of the watercourses that is shownin figure 39. It must be admitted that the position of indi-vidual components of the system generally has had to bereconstructed from overlying levees of Parthian or Sasa-nian date. But given the greater extent and density of sub-sequent settlement, it is to be expected that very fewcanal traces will have survived that are unambiguously(i.e., exclusively) datable to the earlier periods. In anycase, the figure appears to indicate a marked shift awayfrom earlier patterns.

As exemplified by both the Cassite and the MiddleBabylonian maps, earlier layouts tended to involve lineararrangements of sites along parallel, rather widely sep-arated canal or river branches that were connected only atlong intervals. What the Neo-Babylonian and Achaemen-ian systems exhibit, by contrast, is an interlocking, muchmore "artificial" grid of watercourses that broke large,contiguous areas of cultivation into polygons of fairlyuniform size and shape. In the older pattern one mustassume that a fairly broad uncultivated band was leftbetween the parallel branches, while the new, more tightlyarrayed polygons suggest that fairly broad expanses ofcontinuous cultivation had begun to be introduced. Pre-cursors of these zonal lattices may have been introducedas early as the Third Dynasty of Ur (cf. fig. 31) but hadnot survived the political vicissitudes of the following cen-turies. At the least, therefore, they were now reintroducedafter a long absence, and on a considerably expandedscale.

The significance of the new pattern is to be found bylooking not backward in time but forward to later periods.For the Sasanian and Islamic periods the largely completenetwork of surviving canal levees makes it clear that theentire system had been closely and purposefully interdigi-tated. As a result water supplies could be shifted fromany major branch to any area in need virtually at will,surely providing for more adequate, secure irrigation in avariety of contingencies. The new pattern is to be seen,in other words, as an essential part of the expansion andintensification of agricultural production that has alreadybeen associated with the Neo-Babylonian and Achaemen-ian periods by scholars concentrating upon the cuneiformtexts.

Other questions on which there is at least some po-tential complementarity between the archaeological and

188

oi.uchicago.edu

K\ i

(4

<0

'-

NEOBABYLONIAN (A),ACHAEMENIAN (M) &NEOBAB.-ACHAEM.( )

A

0

+-

Fig. 39. Neo-Babylonian-Achaeienian period settlement patterns.

189

r

I

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

the textual evidence are of a more "contextual," wide-ranging geopolitical character than those heretoforediscussed. Systematic surface reconnaissance, with its rela-tively limited zone of coverage, is not well suited to con-tribute to such an approach. Even though less completeand probably less precise in its attribution of periods ofoccupation, the smaller-scale map in figure 40 that in-cludes most of what is known in all of Babylonia providesa better basis. For the Neo-Babylonian and Achaemenianperiods consideration should be limited to only those sitesaccompanied by the numbers 2 and 3.

There is an obvious source of uncertainty in the greatlacuna south and southeast of Babylon. Some reconnais-sance coverage has been obtained in that region, but itseffectiveness was sharply reduced by recent heavy alluvi-ation and by a mosaic of difficult topographic conditionsincluding swamps, cultivation, and interspersed patchesof dunes. Surely Babylon's enormous size and politicalpreeminence during these periods would have led to at-tempts to develop this region immediately downstreamof it, even though the consequences are not apparent.Whatever developments may have taken place there, how-ever, appear to have had little if anything to do with theeconomic and demographic resurgence of Nippur and alarge surrounding region much farther to the east andsoutheast.

The watercourses serving Nippur and its hinterlandsmay have been regarded either as full-fledged branches ofthe Euphrates or merely as canals following the levees offormer Euphrates branches, but at least the major onespreserved the general northwest-southeast orientation ofearlier times. To be sure, there is reason to believe that aNeo-Babylonian canal of some importance was con-structed running east from immediately north of Babylonto serve the environs of Kish, and it may well have beenextended farther to intersect with one or more of the oldparallel channels that approached the Nippur region fromthe northwest (Gibson 1972, fig. 13; Adams 1972, maps1F, 6). We are handicapped in estimating how far thisnew watercourse may have reached during the periods inquestion, since the crucial area for determining that liesin a kind of interstice west of the zone of this intensivesurvey and beyond the eastern perimeters of coverage ofthe Kish and Akkad surveys. But at least there is nothingto suggest that the new channel supplanted the older onesand led to their abandonment.

The continuing economic independence and viability ofthe more easterly region is also suggested by the consider-able efforts that were made to restore the former connec-tion between Nippur and Uruk, which had been inter-rupted in Middle Babylonian times and perhaps during theCassite period as well. With all due caution about leapingto geopolitical conclusions from mere canal alignments,the picture is not one of Babylon's complete hegemonyand the corresponding subservience of regions farther

east. In that sense it may lend some support to Oppen-heim's suggestion that Nippur served as a kind of eco-nomic capital of the Babylonian satrapy, at any rate in

'the later Achaemenian period after extensive destructionin Babylon in the time of Xerxes (486-65 B.c.).

Uruk, on the other hand, seems to have had a more di-rect connection with Babylon. Survey is lacking in muchof the intervening terrain, but the LANDSAT imagery(fig. 6) suggests the existence of a broad old levee fromBorsippa to Uruk, and even continuing southeast acrossthe present lower course of the Euphrates into the vicinityof Ur, that is very likely to go back to Neo-Babylonianor Achaemenian times. Perhaps this more direct connec-tion helps explain why there were Uruk temple holdingsin the territories of upstream towns like Marad and Sip-par, and holdings around Uruk in the case of Sippar (Coc-querillat 1968, p. 35). Nippur lay closer at hand, but itsinfluence over southern Babylonia may have been sup-planted by Babylon and other towns to the west duringthe long period when the canal connection between themseems not to have been functional.

An additional factor involved in Uruk's geographicallyextensive influence must have been the great herds of ani-mals that were administered by the temples there. In thecase of the Eanna Temple, surely one of the largest landedeconomic establishments of the time, it has been estimatedthat during the reign of Cambyses (530-22 B.c.) its hold-ings included as many as 150,000 animals-principallysheep and goats, but also cattle, donkeys, and smallernumbers of other species. Often dependent on stall-fedfodder for part of the year, these herds at other timesranged into northern Babylonia and even across the Tigris(San Nicolo 1948, pp. 277, 285). Primary attention in theforegoing pages has been given to field and garden culti-vation, largely because this was the principal concern ofthe archival sources dealt with earlier. The Eanna textsmake clear, however, that the Babylonian economy con-tinued to comprise an essential balance of cultivation withhusbandry. And, equally important, they suggest that theforaging requirements of the herds must be seen as one ofthe most durable, widely ramifying, and effective forms ofinterregional articulation.

One additional observation can be made on the basisof figure 40. There is nothing to suggest that a networkof canals ran southeast from the district immediatelysurrounding the cities of Seleucia and Ctesiphon, more orless parallel with the right bank of the main Tigris me-ander-belt levee, before the Seleucid-Parthian period. Onthe other hand, it will be observed that there is a groupof Neo-Babylonian and Achaemenian sites beginningabout 40 kilometers to the southeast. Again we encountera lacuna in archaeological coverage beyond the limits ofthe intensive reconnaissance, but then there are suggestivegroupings of sites (including, e.g., site 552) that may wellhave been connected with the others by a canal. This must

190

oi.uchicago.edu

.234 "23 3

23 2 (. 33. 34 34'323. .34) 2.3 &4 4

2 44 34.

,34 , 30342 32

'34. -../-_\ - ./ ,

... \. -. --- . * /\ -.

.23(4) " 34 • -3 3 34

, , \ 34 4 ) 'C4343( .23 4)

23 3 (3,2 "("'.. ".-'2" / .. ( 23" \'.,

234j. i3 4 23 ./(234) (234)

2i3) 4 -.- 23 ) %12)34.12 34 .4 -123 1-4H, \ . .. ., .- _, , \

. . , -4 1-

!(1) 4 .2. 2234 2(234)- 34'.343.4 (3 234:

2 .1 ( 234 2342 3 / 2233 ' 4 1 _ 2 12((3(2) '23 4 23

1((2))3 4

0 25 Km.N=W

23. )" 1"23(4) "\2 \

, 23. -- / 2, 23?34

. 23 * 23 223.(23)4 23423. 232 32

S 3 , (23 2

(3 3 \ \. 2 \2 3 (12 I

23 23)

1 23 2 )* 34

-4 (3)_ 34 % .. -..23 3 4- 3 ( 32

4. )12324) 12( )4

" 22•4 3 23 2

234 \C - 3 . ) . -% 22 3 24)

S 23 "23(4) 23423 )" .• . "34)) \ .3 .• ';. 2S-4." ."-23 ",

---- 23 "• 3\-, - --34.2)24 2

-l'b· 222.2.2 1222223 2 2L IIII 2222 2..v, 2II 222 2.,.III . 2. 2 L 11111 IU 211 2 2 (32L2-V3 '2 II I LVIJ.t3b

potamian plain.

191

23 SIP 2 .12312 --

2323 2 2 23

S 12(3)

12 31'

2*.(*234

.n only

2,1

3 (234)

234 ,

\i

-34

234

342 234

'23

-234 234 0)4,.234(2(3 4 23 23234.

' N23)4(.32_

123(4)* ·

3--/)

23\

/2 23

23 \ 23-23(4i .j 23

2"(234)23 ,

\•L'ARSA

It

.*34

(M)23

i

1-

-2

3.4234 12

.4-~ 23

\i

,23(4) 3.4

34

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

indicate a major new initiative in either the Neo-Baby-lonian or the Achaemenian period directed toward em-ploying the Tigris as an irrigation source. Its headworksprobably lay in the vicinity of a little nineteenth centuryArab fort and an extensive series of nearby ruins recordedby Collingwood as Kalat el-Buddah and the Gorgotmounds (Barnett 1963, fig. 4). One would expect that asubstantial settlement, perhaps even on the scale of theruins of Opis, might be associated with the main offtakeof a major new canal system of this kind, but that is ofcourse unsupported speculation." At least for the present,what is more interesting is that the fundamental shifttoward the Tigris as the main axis of trade, settlement, andcommunications-characteristic of all later periods-mayhave been foreshadowed even before the arrival of theconquering Greek army of Alexander.

North across the Tigris lay the lower Diyala plains, andI will conclude this consideration of Neo-Babylonian-Achaemenian irrigation and settlement patterns with abrief discussion of the more modest developments there.The relevant data on site size and location have been pre-viously published (Adams 1965), but only with very gen-eralized suggestions on what the accompanying water-course patterns were. Rather than being derived from theconventional map series that served as a base for theoriginal publication, with its exceedingly incomplete andoccasionally inaccurate renditions of former canals andancient canal levees, the watercourses shown in figure 41conform with the ancient irrigation patterns that havebeen independently traced from the aerial photographs(fig. 8).

It is clear from this new mapping that the beginnings ofa trend toward settling the right bank of the Tigris duringthe Neo-Babylonian and Achaemenian periods had noreal counterpart on the left bank. A renewed growth ofpopulation and spread of settlement after Middle Baby-lonian times is just as evident in the Diyala region as far-ther south, but it does not appear to include new enclavesdepending primarily on the Tigris for their irrigationwater. Instead, the overall pattern is essentially the sameas that found throughout the third and second millenniaB.C. (Adams 1965, figs. 2-3), with settlements accompany-ing the numerous branching mouths-predominantly nat-ural-through which the ancestral Diyala joined theTigris.

In most other respects as well, the pattern shown inthe accompanying map (fig. 41) provides no suggestionof changes to come. The two main upper branches ofthe Diyala lay in positions paralleling the present streamto the east and west, not greatly different from wherethey had been a millennium or more earlier. Even thoughthere had been a substantial recovery of population levelssettlement was still confined to villages and a small butmore rapidly growing number of small towns. Nothingis to be seen like the polygonal arrangements of sites on

the ,central Euphrates floodplain at this time, with theirlarge, probably continuously cultivated areas served bylattices of canals implying some degree of state initiationand management. As late as Achaemenian times theDiyala plains were still a provincial backwater, but theywould not remain so much longer.

The Seleucid-Parthian Period (ca. 330 B.C.-A.D. 226)

An understanding of the strategic significance of theshift from reliance on one primary riverine artery toanother is crucial before we take up the Seleucid-Parthianperiod and succeeding periods in detail. It is as a readilynavigable connecting link between a series of adminis-trative capitals and trade emporiums-Seleucia, Ctesi-phon, Charax Spasinou, Forat, Vologasias, and Baghdad,to name only the most notable-that the Tigris gainedascendancy over the Euphrates.4 The advancement ofnot merely riverine but oceangoing trade is implied asthe dominant consideration behind the move, an en-couragement of a rising tide of luxuries from the fartherOrient partly for local elite consumption but especiallyfor large-scale overland transshipment to the Mediter-ranean world. Only in the context of greatly expandedhorizons of intercourse, in other words, is the shift com-prehensible at all.

But with the development of a new main axis ofcommunications primarily for that reason came a chainof consequences that could not have been anticipated.The topography of the upper part of the alluvial plainmade easterly extensions of Euphrates left bank canalscomparatively simple to construct, and the relative Tigrisand Euphrates levels there permitted properly gradedflow only in the direction of the Tigris. When Seleuciawas first sited on the right bank of the Tigris, such anextension not only would have provided a fairly man-ageable urban water supply but also would have ex-tended that city's command of routes of communicationswestward to the Euphrates. The major drawback of theTigris had been that the ultimate destination of most ofthe imported commodities made transshipment necessaryin order to have them follow the line of the Euphratesvalley westward. When navigable waterways united theEuphrates with the Tigris near the upper end of thealluvium, that problem disappeared.

The positive inducements for powerful new groupsof merchant venturers like the Greeks, Palmyrenes, andNabateans (and their allies and suppliers) are obvious.So also are the advantages that would be seen by Persian,Greek, and other administrators, conscious of the ex-panding imperial frontiers of their responsibilities. TheTigris was navigable for much larger vessels of deeperdraft. Its development as the major artery assured theavailability of international prestige goods to local elites,while at the same time enhancing state income throughtaxes on the transit traffic. But we must not overlook the

192

oi.uchicago.edu

10 1 25km.

t 0.1-4 ha.0 4.1-10 ha.

o//

\

S

--- 0•

Fig. 41. Neo-Babylonian-Achaemenian watercourses and settlement patterns on thelower Diyala plain.

193

*

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

accompanying negative consequences. The effect of theshift was slowly but permanently to reduce the statusof the lower alluvial plain of the Euphrates from whathad once been the most advanced locus of powerful statesand cities to a subordinate appendage.

This was so in three respects, First, the introductionof transverse canals from west to east, cutting obliquelyacross the natural orientation of drainage (and the olderirrigation system) from the northwest, greatly increasedthe agricultural and settlement potential of the upperplain. In time, therefore, a whole series of similar chan-nels were developed. These were (from north to south)the 'Isa, the Sarsar, the Malik, the Kutha, and the GreatSarat-Nil of the classical Islamic period (Le Strange1905, p. 24), which in modified form still exist today.The impetus behind new canal construction graduallyshifted northward, in other words, with the advantageof opening new lands where slightly higher naturalgradients made salinity a less chronic (or at any rate,less immediate) problem.

Second, the consequences of the new urbanizationalong the Tigris and the development of a new irriga-tion grid in the northern part of the Tigris-Euphratesalluvium were by no means limited to the lands betweenthe two rivers. Some of the new cities like Ctesiphonlay on the opposite bank of the Tigris, and it was onlynatural that the lower alluvial fan of the Diyala Rivershould receive the same stimulus. Previously the lowerDiyala plain had been politically marginal and thinlysettled, relative to the central Euphrates floodplain, butnow a basis was laid for that area to become even moredensely urbanized. The results of my archaeological sur-vey of the Diyala region can be recalculated in terms ofthe same size categories used in this study, and a com-parison illustrates strikingly the dramatic rise in its im-portance (see table 19).

The entry in table 19 for the period around 2000 B.c.is intended to provide a baseline from which to assessthe subsequent changes. At that time, and in fact through-out the more remote antiquity of the region, the aggre-gate occupied area on the lower Diyala plains amountedto less than a quarter of that in the intensively surveyedpart of the central Euphrates floodplain. Moreover, thepercentage of urban sites was substantially smaller. Bothproportions dropped lower still and remained at thoselow levels through the Achaemenian period. But by theSeleucid-Parthian period the percentage of urban settle-ment was already appreciably higher than it was to thesouth of the Tigris, and by the Sasanian period the totaloccupied area within the two regions was nearly equal.

The immense expansion of settlement that went onin the Diyala region should not be thought of as a smoothand uniform infilling of lightly occupied districts. Hadthat type of growth occurred, we would more likelyfind a proliferation of small settlements such as might

TABLE 19 A Euphrates-Diyala Comparison of Trends inSettlement

Lower Diyala Central Euphrates

Estimated EstimatedEstimated % Urban Estimated % UrbanHectares of (10+ Hectares of (10+

Period Settlement Hectares) Settlement Hectares)

Ur III-Isin-Larsa 602 35% 2,725 75%

Neo-Babylonian-Achaemenian 134 7 1,769 51

Seleucid-Parthian 1,857 69 3,201 55

Sasanian 3,489 75 3,792 58

Note: The two left-hand columns are derived from Adams 1965,tables 12, 16, 17, 18, and 19. The Neo-Babylonian and Achaemeniantotals have been aggregated, and the first three listings have beenrecalculated by size category rather than estimated area, to provideequivalences for the figures used in this study. The right-hand twocolumns are taken from figure 25 and tables 15 and 17.

have been established by groups of agriculturalists mi-grating in search of less saline soils or less oppressiveconditions of tenure. But instead the pattern to be seenin figure 42 is a markedly discontinuous one with, asjust noted, a higher proportion of urban sites than inregions to the south that formerly were more thicklysettled. Large areas whose agricultural potential is con-firmed by their later utilization still were virtually unoc-cupied, while elsewhere the grid or lattice pattern of siteand canal distribution suggests that other large areaswere completely cultivated. The center of gravity of thepattern shifted considerably to the east and south, al-though a few isolated but quite large urban centers alsoappeared west of the modern course of the Diyala andin the northern part of the region.

Some development of the Tigris left bank belowCtesiphon is indicated, corresponding to what had al-ready begun on the right bank below Seleucia a fewcenturies earlier. But the cluster of substantial towns andcities that were founded at this time 50 kilometers or soto the east of Ctesiphon is not so easily explained as aderivative of the founding of the new Parthian capital.Most of the larger sites define a southeasterly axis fora major part of the Diyala's flow, well to the north ofCtesiphon and in fact competing with it for Diyala water.Moreover, the aforementioned cluster traces a fanlikepattern of distribution, very similar to later patterns inthe same area that were made possible only by a greatweir across the Nahrawan canal north of Uskaf baniJunayd (Jacobsen 1958; Adams 1965, figs. 18-19). It isthus more than likely that this impressive barrage hada smaller Parthian precursor on a Diyala channel thathad undergone some enlargement-and that may be

194

oi.uchicago.edu

0 10 25 km.W- NC31 ---

Fig. 42. Seleucid-Parthian watercourses and settlement patterns on the lower Diyalaplain.

195

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

regarded in a sense as the Nahrawan's antecedent. What-ever the administrative and fiscal responsibility that theParthian state authorities in Ctesiphon may have exer-cised over this enterprise, it is obviously not to be re-garded as a product of a mere coming together of scat-tered groups of peasant immigrants seeking new lands.

No details on irrigation and settlement trends com-parable to these for the Diyala alluvial fan can yet beprovided for the northern part of the Tigris-Euphratesalluvium, but it would not be surprising if they were toindicate a generally similar trend of rapid growth. Northof the latitude of Kish and Babylon there had also pre-viously been comparatively sparse settlement and fewcenters of urban proportions. By Parthian times, on theother hand, Gibson is able to identify four major citiesto the north and northeast of Kish alone and to speakof the "incredible density of population implied by milesof virtually continuous settlement" along the Kutha canalin the Sasanian period (Gibson 1972, pp. 51, 57). Butwhat must have held back the northern end of theTigris-Euphrates alluvium, at least to some extent, wasthe periodic destruction visited upon it by invadingarmies. Babylon had suffered extensive destruction underthe later Achaemenids, and a substantial part of its popu-lation must have been drained off to Seleucia soon there-after. Seleucia in its turn was destroyed by Roman attackin A.D. 165, founded anew in early Sasanian times, thensacked once more and permanently deserted in 283(Streck 1917, p. 24). Somewhat better protected behindthe Tigris River, cities and towns on the Diyala plainsfared better.

In short, the shift to the Tigris and the introductionof a new irrigation grid triggered secondary changes inan adjacent area that were perhaps even more extensivethan the primary ones. After the intense development ofthe lands along the lower Diyala the preeminence ofsouthern Babylonia was irretrievably lost.

These first two factors involve the comparatively morerapid growth of settlement and agriculture in the northbecause of advantageous new programs that were un-dertaken there. Their negative influence on the southwas only indirect, in that part of the population musthave been drawn away to participate in them. But thereis a third factor that involved directly deleterious changes.The new transverse canal pattern permanently disruptedits older counterpart. For millennia the latter had pre-served the natural condition of an alluvial fan, in whicha number of channels separated from one another nearthe head of the plain at Sippar and followed parallel orradiating courses down its length. Under that originalpattern the sources of water for the south were less im-mediately subject to diversion and control by northerncompetitors for the same scarce resources. Now, how-ever, there first were an increasing series of diversions tomeet the demands of settlers on the upper plains, with

only the excess being passed along to the region of theancient cities of southern Babylonia. Moreover, the re-quirements of the upper plains would have to be reckonedto include navigation as well as irrigation; it served needsof the state other than agriculture to maintain a relativelylarge diversion of Euphrates water into the Tigris nearSeleucia through the tails of the Nahr al-Malik. 5 Thusthe south gradually lost its access to a dependable watersupply, while the volume of water carried by the Tigriscorrespondingly rose.

This does not imply that water failed to reach thesouth, for the great swamps that prevailed there in classi-cal Islamic times conclusively show otherwise. But itwould have tended to arrive more irregularly, with short-ages manifest particularly during the early, low-waterpart of the growing season, in unmanageable flood crests,and via the tails of lower Tigris diversions at levels fre-quently too low to command the available land for large-scale and canal irrigation. Not immediately, but over thelong run-it must be stressed that this was not a constantbut an irregular process, and that it continued over cen-turies-the inhabitants of the lower plains found it moreexpedient to drift northward themselves than to resistthe conditions that were closing in upon them. Hereinlies the underlying cause for the inexorable northwardmovement of the perimeters of settlement whose begin-nings were already evident by the late Parthian period,and which by Middle Islamic times had seen an almostcomplete abandonment of the entire southern portionof the Euphrates floodplain.

This discussion of the implications of the shift hasbeen in a sense premature, in advance of an account ofgeneral conditions obtaining on the lower plain duringthe Seleucid-Parthian period. While the reorientation ofthe canal network at the head of the alluvium had begunby Seleucid times, the dire long-term consequences thathave been sketched were by no means felt immediately.In the region around Uruk, on the contrary, the Parthianperiod has been interpreted as "the culminating epochin the entire settlement record" (Adams and Nissen 1972,p. 58). This was because by Sasanian times the abandon-ment of the southern part of the region was well underway, but in any case the existence of an impressive finalperiod of population growth and apparent prosperitythere is undeniable.

We must bear in mind, in considering the overall pat-tern of Seleucid-Parthian settlement, that the sites fromwhich it is reconstructed may have been occupied atany time during a period of perhaps five hundred years.More specifically, the Seleucid and Parthian componentsof it cannot yet be readily distinguished. Furthermore,there is no independent evidence as to the degree of con-tinuity of occupation that obtained in the smaller townsand villages across the countryside. With these reserva-tions, the general impression nevertheless is one of little

196

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

more than an intensification of a system that had alreadybeen laid down in outline in Neo-Babylonian andAchaemenian times.

Figure 40 shows clearly that extensive further canali-zation took place along the Tigris right bank belowSeleucia. Much of this depended on progressive extensionsof the nar malkha that had been brought from theEuphrates to that city. The lands along the Tigris meanderbelt levee thus served do not appear to have been settledor cultivated previously, and they must have been anattractive prospect for agriculturalists drawing in uponSeleucia (and in the Parthian period Ctesiphon) frommany parts of southern Mesopotamia. Over time, thetails of the nar malkha system gradually must have beenlengthened to take advantage of the fertile, well-drained(if also prone to flooding) levee terrain. Undoubtedly bythe Parthian period, and perhaps even earlier, the Neo-Babylonian-Achaemenian canal system 40 kilometersbelow Seleucia was annexed to it directly. It is only thegreater age of these sites farther downstream that indi-cates they were once part of an independent system an-tedating the arrival of Euphrates water. More recently,as figure 40 shows, the ubiquitous meander-cuttingprocess along the Tigris has cut into the tails of the narmalkha system at intervals so that some of its branchescan no longer be followed as a continuous levee.

We must assume that there have been course changesall along the middle Tigris, partly through localizedmeander-cutting like this and partly through more pro-nounced movements. In one such change the bed of theriver moved eastward from a position between Seleuciaand Ctesiphon to a present course probably adjoiningCtesiphon itself while passing through the ruins of thelater Sasanian city of Weh Ardashir or New Seleucia(Invernizzi 1976, p. 172).6 Similar shifts have occurredfarther downstream, although in most instances thepowerful aggradational effects of periodic flooding havemade it difficult to follow them in detail.

The pattern of change of the Tigris course was un-doubtedly complex and frequently reversible. The sub-stantial Islamic town of Dayr al-'Aqul, for example, wasdescribed in the tenth century as being situated on thesouth or right bank of the river, whereas today its ruinsare on the left bank (Adams 1965, p. 91). Yet study of theair photographs indicates beyond question that the irri-gation system supplying it during at least its earlier hey-day was an effluent of the Nahrawan canal. During theSasanian period and probably also the Parthian periodit must have been situated on the left bank, therefore,and its reportedly different position in classical Islamictimes was the result of no more than a temporary oscilla-tion. The ruins of Humaniya are also found on the Tigrisleft bank today, slightly farther downstream, but in thiscase the left bank ancient canal patterns provide no hintthat the town ever lay on the north side of the river before

its abandonment. Between the two towns, but 15 kilo-meters north of the Tigris course that they very roughlysuggest, lay a city directly on the lower Nahrawan canalthat in Islamic times was called Lower Uskaf (Adams1965, pp. 66-67). Plausibly, but not with certainty, thiscan be identified as the earlier Hellenistic city of Scaphae,shown in the Ptolemaic map of Mesopotamia as lyingdirectly on the Tigris. However, a reconstruction carryingthe bed of the Tigris at any one time past all of thesepoints would involve an enormous northward loop, en-tirely beyond the dimensions of the present Tigrismeander belt. Hence one must suppose that there hasbeen a greater latitude to the river's past movementsthan is suggested by shifts that would place individualtowns successively on its right and left banks. But un-fortunately there are as yet no other historically attestedpoints between Ctesiphon and Dayr al- 'Aqul with whichto confirm and extend our knowledge of these widershifts.

Still farther downstream along the Tigris right bank,about 100 kilometers below Seleucia, there are tracesof an additional new canal of Seleucid-Parthian date. Itsdistance from the apparent tails of the nar malkha systemof the time makes its direct dependence on a Tigris offtakelikely. Having recognized the effects of Tigris coursechanges, however, we must acknowledge that the head-works of this canal may have lain considerably fartherto the northwest than the bend in the river that at presentappears to represent the start of the canal. Quite possiblythe original offtake of this canal will one day be foundin the archaeologically unsurveyed region farther north-west along what is now the left bank of the Tigris. Therewas, in other words, a rapid filling in of the entire newTigris levee area once the shift of a few of the main urbancenters to its banks had been undertaken.

On a more limited scale, something comparable seemsto have occurred in the southwestern part of the centralEuphrates floodplain. Figure 43, illustrating Seleucid-Parthian settlement patterns within the intensively sur-veyed area, depicts a network of new canals west-north-west of Uruk that must have been fed by Euphratescourses by now situated far to the west of the formerBabylonian cities. Farther upstream, Gibson (1972, p.50) has noted a contemporary westward shift toward aEuphrates course in the vicinity of Babylon. Presumablythis was the course that also, at one or more offtakes farbelow Babylon, was the source of water for the newcanals above Uruk.

Thus there seems to have been an unprecedented dis-persal of Euphrates water. Some, by way of the narmalkha and the outskirts of Seleucia, was even beingapplied along the Tigris levee to the east. Another por-tion was following one or more channels comparable tothe modern Hilla and Hindiya branches in the westernpart of the alluvium. Yet, in spite of these diversions,

197

oi.uchicago.edu

0 "• -* 882

/

I ° -- 0901

A .

Fig. 43. Selucid-Parthian period settlement patterns in the intensively surveyed region.

198

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

figure 43 suggests that there was a further increase inthe density of settlement even in the central floodplain,north and east of Nippur. There is every reason to believethat, with an increasing geographic breadth of dispersalas well as density of settlement, consumption needs onthe lower plains were beginning to approach the limitsof what the Euphrates could supply with reasonablesecurity for those dependent upon irrigation agriculture.

As I noted previously, Seleucid-Parthian settlementstend to be irregular and discontinuous rather than com-pact. Only a few were of the proportions of major cities,a mere 25 percent of the total occupational area in siteslarger than 40 hectares as contrasted with 38 percent atthis time in the Diyala region and 55 percent in the samecentral Euphrates area during the Ur III period. Thelarge centers are fairly evenly dispersed, suggesting thatthey were performing central place functions for hinter-lands of fairly uniform size, but there are some discon-tinuities in the distribution of intermediate towns. Inplaces, as above both Uruk and Nippur, there are close-spaced clusterings of the latter that in most cases seemto go back to Neo-Babylonian or Achaemenian ante-cedents. Elsewhere, as in the newly canalized area fartherto the west-northwest of Uruk, there is a noticeably moreregular dispersal of small and intermediate towns in longlinear series of villages. This whole area gives the appear-ance of being a planned, large-scale development, notorganized by and around a central city (or at least notone that has yet been located through archaeologicalreconnaissance). If the survey provides any evidence ofSeleucid or Parthian state intervention in the developmentof the central alluvium, apart from the founding of thecapital cities and the canalization of their immediatehinterlands, it is probably to be found here.

Substantial, vigorous occupations continued through-the Seleucid period at both Uruk (Greek Orchoi) andNippur, but at both sites this is followed by a textual andnumismatic gap during early Parthian times until shortlybefore the time of Christ. Nor can this be filled by findsat sites elsewhere in the region, and it may well be thatthere was an interval of 130 years or so in which activityin at least the major centers was much diminished (Nissen1973, p. 82; Adams and Nissen 1972, pp. 57-59).

The silence is hard to explain. No direct testimony hasyet been found of destructive campaigns in the immediatearea as a consequence of intradynastic rivalries. Perhapsone should simply assume that military activities in-itiated by Charax under Hyspaosines (ca. 140-121/120B.C.) were focused on a handful of the major towns withimportant temple precincts, including Uruk and Nippur,to which it has already been noted that knowledge ofcuneiform writing was largely confined. Only in thesecenters would the use of clay tablets have persisted tomeet some functional needs for record keeping, whileelsewhere records in Aramaic were written on materials

that have almost entirely perished (Oelsner 1978, p. 106).Speculating further along this line, one could argue thatfairly limited and selective destruction thus would createfor us today the impression of a major abandonment.On purely a priori grounds, the case for continuity in-stead of abandonment is that the prosperity, periodicautonomy and even expansion of Charax under weakParthian suzerainty seems as likely to have had positiveas negative consequences for southern Babylonia in gen-eral. But the absence of evidence for economic or buildingactivity elsewhere in the region, even though based ona very inadequate body of observational data, remains un-explained with this hypothesis. At present it can only besaid that, for a substantial period, the region simply dropsfrom notice, perhaps with some accompanying move-ment of population (and especially the urban population)northward toward new centers of power like Seleuciaand Ctesiphon.

The recovery, in the first and early second centuriesA.D., is also puzzlingly silent in view of its impressivescale. Keall has argued, plausibly if not conclusively, thata "southern strategy" was pursued by Vologases I (ca.A.D. 38-60) and his successors. Presumably it was aimedat circumscribing the independence of the Greek tradersin Seleucia, imposing a Parthian governor in place of theHyspaosinid dynasty at Charax, and ultimately securinga greater return for Parthia from the Palmyrene tradewith India (Keall 1975, p. 625). It would have coincidedwith the first arrival by sea of Chinese silk, a costly luxuryfor which there was soon an enormous demand in theRoman world (Nodelman 1960, p. 102), and this mayeven have precipitated the decision to reassert greatercontrol over affairs in Mesene and Characene.

The founding of Vologasias to undermine the role ofSeleucia could well have been a part of such a strategy.At one time the new town was thought to have been situ-ated on the Euphrates, but a location on the Tigris belowCtesiphon and Seleucia now seems fairly certain (Maricq1959). An impressive fortification at Nippur, repeatedlyrebuilt on the ruins of the Temple of Enlil and occupiedbetween roughly A.D. 70 and 160, is seen by Keall asanother manifestation of the same objective. It was leftunfinished, and he believes it may have been intendedmerely to house a garrison placed there to help assurerural security on the plain south of the capital, ratherthan to withstand a siege (Keall 1975, p. 626). Yet atten-tion should be called to another impressive fortification,apparently contemporary and close to Nippur (site 826).Someone went to a great deal of effort to protect some-thing, in a region and period of time for which the avail-able sources give no hint of either military activity or in-ternal dissidence.

It is a distinct possibility that these installations areconnected in some as yet undefined way with the Palmy-rene caravan traffic. That traffic, already substantial,

199

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

greatly intensified in the second century, even afterTrajan's invasion and conquest of Mesopotamia (A.D.114-17) was followed only by Hadrian's abandonmentof Roman territorial aspirations there. The route betweenCharax and Palmyra was now marked with milestones,provided with a carefully organized system of wells andway stations, and protected by guard posts manned bycontingents of Palmyrene mounted archers. These wererecognized as auxiliary units of the Roman army, andPalmyrene commemorative inscriptions honor otherRoman military agents who not only also operated deepin Parthian territory but accompanied caravans all theway to Charax itself (Nodelman 1960, pp. 111-12).

The aspirations of the "southern strategy" were at bestbriefly attained, in other words. By the beginning of thesecond century, Parthian control had weakened to thepoint where its toleration of Palmyrene-Roman opera-tions in nominally Parthian territory constituted a passiveacknowledgment of Roman supremacy. Even if a fortlike Nippur may have been begun as an initiative of theParthian crown, therefore, its final rebuilding after thetime of Trajan may have been a response to differentstimuli. Architectural evidence for Zoroastrian religiouspractices in the phase III fort at Nippur rules out a Palmy-rene garrison, but not necessarily a local undertaking atPalmyrene behest. In the case of Uruk, too, Nissen sus-pects that its last period of growth and prosperity mustbe somehow linked with the caravan traffic (1973, pp.83-84). But in both cases the nature of the links remainselusive.

To summarize briefly these inconclusive leads, thecentral and southern Euphrates floodplain lay on the verymargins of the extensive but loosely integrated Parthianrealm. It does not figure in any official accounts yetknown, in spite of the dense population and extensivebuilding that went on there during the first and secondcenturies A.D. There is every reason to believe thatParthian control, and perhaps even interest, fluctuatedmarkedly. Observing that "the whole question of the rela-tion between Parthia and the small surrounding king-doms which are usually described as vassal-states remainsto be clarified," Nodelman stresses that Arsacid claimsto suzerainty "must have been at most intermitently en-forced" (1960, p. 103). Keall, too, concedes that imperialcontrol "was little more than a veneer" and that "notinfrequently the term 'Parthian Empire' is a misnomer"(1975, p. 620). It does not seem credible, in these circum-stances, to identify as the primary stimulus for develop-ments in our region the policies envisioned (and seldomconsistently followed) by the Arsacid dynasty inCtesiphon. For perhaps the last time, under the specialconditions of a power vacuum on the lower Mesopota-mian plain generated by Roman-Parthian military andtrade rivalry, the inhabitants were for a time left rela-tively undisturbed to pursue initiatives of their own. That

they did so with such marked success-at least as mea-sured in the extent and density of settlement-merelyconfirms the stimulus that similar conditions offered totheir remote ancestors four millennia earlier.

The Sasanian Period (A.D. 226-637)

The Sasanian period, as I have already pointed out inthe preceding section devoted to an overview of demo-graphic and urban trends, represented in many respectsthe apogee of ancient developments on the centralEuphrates floodplain. As I also noted earlier, however,the characteristics for which the dynasty is principallyknown were not by any means uniformly present through-out the period. Vigorous strategies of unification wereforeshadowed in a series of acts and aphorisms creditedto the early rulers in the dynasty but then were substan-tially reversed during much of the fourth and fifth cen-turies by the growing independence of the landed no-bility (Frye 1956, 1:319, 325). Only in the sixth century,in a series of moves characteristically combining militarysuccesses on the frontier with Byzantium and elsewherewith an administrative consolidation and fiscal reform,did the dynasty's centralized strength and capacity forlarge-scale, coordinated planning emerge once more inunmistakable form.

In an important sense, as Pigulevskaja has cogentlyargued, it is in the cities that one finds the most tangibleembodiment of the cumulative transformation theSasanians effected:

Comparative study of the ordinances of Hellenistic towns,those organized as a polis and others, and of conditionsin towns during the crowning epoch of the Sasanid state,leads to the conclusion that they had lost their inde-pendence of administration. They had nevertheless con-served and developed their original organizations ofcrafts, of corporate groups having their own representa-tives. The "imperial cities" began to occupy a new andimportant position, constructed and protected by theshahs, who interested themselves in them for economicas well as political reasons. [1963, p. 231]

Again, however, the reservation can be voiced that abrief summary like this tends to compress an irregularseries of developments with opposing tendencies, a kindof dialectical process, into a smoothly cumulative seriesof changes. It is not within the scope of this study to dealin detail with the oscillating balance of forces that insteadseems to have occurred over time, but even the archae-ological findings for the period cannot properly be un-derstood without further reference to the dynamicsbehind that shifting balance. Altheim has briefly sum-marized the process as follows:

The Sasanian economic landscape divides itself intotwo parts: on one side the domain directly under royal

200

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

rule, and on the other the domain of the landowningnobility in which the central power operated only in-directly. It was in the interest of powerful, far-reachingroyal control to increase the number of royal cities andtheir attached rural districts. This had the effect of con-verting indirectly ruled into directly ruled districts, andonly partly taxed districts into fully covered ones. Thehistory of the royal founding of cities thus also concernsthe struggle between royal power and that of the nobility.

Royal foundations, whether new or simply renamedformer settlements, were only possible on royal lands.It was Ardashir's intent in founding a new -dynasty todo away with the restraints on earlier kings by keepingin his own hands all the lands that he had won with thespear. In this fashion he succeeded in expanding theterritories under his direct rule, giving him the possibilityof founding or refounding cities in his own name. [1962,p. 220]

After Ardashir (A.D. 226-41) and his successor ShapurI (241-72), these policies were for a long time largelydormant. Only under Qubadh (488-531), in connectionwith a complex internal struggle over the Mazdakitereligious movement, were circumstances favorable oncemore for the royal power to assume the initiative. A broadreform of the fiscal base of the empire was begun withcomprehensive measurement and recording of cultivableareas subject to tax, and this was carried to completionby Khusrau I Anosharwan (531-79), with whose reignthe dynasty attained its greatest influence. Taxes onbehalf of the central authority, previously levied only onurban landholdings that were subject to direct royal con-trol, were extended under his rule to all other holdingsas well. A new nobility of service, without an independentbase and hence more responsive to royal direction, wasestablished both to fill sharply escalating administrativeneeds and to undermine the position of the landed nobility.Also on royal initiative, the construction or renovation ofirrigation works, roads, and bridges was undertaken onan apparently extensive scale. And always symptomaticof the accretion of royal power, we hear once more underKhusrau of the founding of cities (Altheim 1962, pp. 222-23; Altheim and Stiehl 1954, pp. 11, 133-34; Noldeke1973, pp. 164-65).

The tax reforms are of particular interest for the lightthey shed on the agricultural regime of the Sawad,roughly coterminous with the Tigris-Euphrates alluvium.Head taxes, levied three times a year on all males be-tween twenty and fifty years of age, ranged between fourand twelve dirhams according to means. Land-tax rates,

1applying only to fields under specified kinds of cultiva-tion, were based on crop type and productivity. Theywere set at one dirham per jarib (0.1592 hectare) forwheat and barley, five-sixth of a dirham for rice, sevendirhams for lucerne or clover, and eight for grapevines.Six olive trees or ordinary palms were also taxed one

dirham, but in the case of the finer Persian palms thatrate applied to only four. Provision had sometimes beenmade for the remission of taxes in the event of cropfailure, but now the needs of the state for an assuredlevel of revenue were firmly declared to take precedenceover the fluctuating ability to pay of individual agricul-turalists (Noldeke 1973, pp. 241-46).

The effect of the new approach represented by thesereforms was not merely to rationalize the preexistingsystem and make it more uniform but to introduce astructure of incentives that had previously been largelyor wholly absent. By specifying fixed amounts of taxper unit area regardless of yield, an inducement was givento landowners and tenant agriculturalists to maximizetheir own rates of return by cultivating more intensively.Apart from other forms of state intervention devotedlargely to the most capital-intensive portions of the irriga-tion system, this would presumably have fostered addi-tional local investments of labor and capital. Land level-ing and improved field canalization and drainage weresurely among the most productive forms this could havetaken, but quite possibly the rate schedule also favored anexpansion of summer cultivation or a shift to crops with agreater unit value. One is reminded of the Sasanian politi-cal precept that "the tax (kharaj) is the support of thestate; nothing increases it like justice and nothing reducesit like oppression." Enunciated already at the beginningof the dynasty, it was not only verbally reaffirmed byKhusrau I Anosharwan but seemingly brought nearer topractical implementation as a policy (Morony, n.d., chap.1, sect. 2). On the other hand, there is reference to such areform only for the Sawad, and it may not have beenextended to the realm as a whole.

Gross taxes received from the Sawad under Qubadhwere reported in later Arab sources to have been theequivalent of about 214,000,000 dirhams. No directlycomparable later figure has come down to us, but afterwide fluctuations state income from all sources is saidto have risen steeply under Khusrau II Parviz (590-628).If geographic regions can be assumed to have maintainedtheir relative proportions until the time of Mu'awiya(661-80), the income from the Sawad in Parviz's eigh-teenth regnal year would have been on the order of240,000,000 dirhams. Dividing gross revenues on the sameassumption, it might even have increased to more than340,000,000 dirhams by the end of his reign (Altheimand Stiehl 1954, p. 41). There is, of course, a very realpossibility that these numbers were corrupted during thelong and uncertain process of transmission to the Arabchroniclers on whom we depend.

Unfortunately, moreover, there appears to be no com-pletely credible approach to disaggregating the totalsin order to arrive directly at the scale of the agriculturalsystem and the relative magnitudes of its various com--ponents. If due allowance is made for exempted uncul-

201

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

tivable and fallow lands, it is difficult to imagine the tax aged more directly. But other than with regard to ruralbase (at any rate within the alluvium) in any given year security and irrigation, the crown's knowledge of andexceeding 25,000 square kilometers. This is less than interest in the countryside extended little further than16,000,000 jarib. If only the minimal rates for wheat, the extraction of revenue.barley, and rice were paid for at least three-quarters of We are fortunate in having one body of roughly con-it, which surely was the case, then a very heavy tax burden (temporary textual material with a quite different perspec-obviously must have fallen on orchard crops-and par- \tive-the Babylonian Talmud. It has serious limitationsticularly dates. Relying again on Neo-Babylonian density of its own as a source on agricultural practices and condi-figures (above, p. 186), and assuming equal numbers of tions, but, on the other hand, it furnishes us with a di-Persian and ordinary palms, taxes on date orchards verse, richly textured account filled with anecdotal detailcould have supplied a potential state income of 45 dirhams at the level of direct, personal interaction. There mayper hectare for no more-and probably considerably less even be an advantage of reduced bias and selectivity in-than 4,000,000 jarib. This suggests that the reve- that the Talmud reflects no conscious concern with de-nues credited to Qubadh are within, but approaching the picting the life of its times but instead is preoccupied withupper limits of, the bounds of credibility if we assume perpetuating and inculcating various rabbinical tradi-that the land tax was the major component in those Ntions. Newman's contribution (1932) in painstakinglyrevenues.7 The figures for Khusrau Parviz, or certainly winnowing and systematizing the mass of material is onethe one reported for the end of his reign, are then pro- for which nonspecialists must be grateful.gressively less realistic. The most that can be said is that To be sure, the full representativeness of the Talmudthey lend some support to an observation already made must be doubted on a variety of grounds. The presence ofon the basis of survey data, that there had apparently unbelievers is referred to at intervals, but its only directbeen an immense extension in both population and cul- concern was with a large, vigorous Jewish communitytivated area as compared with the most prosperous that in the main was responsible for its own affairs. Atperiods of earlier antiquity such as the Third Dynasty of least to judge from the available data, Jews lived in aUr. relatively small number of compact settlements, many

Official sources thus offer us a vague, remote, and gen- of them rather large, in which they constituted a heavyerally unsatisfactory picture of the agricultural regime, majority. They enclosed themselves, and were enclosedeven though an aphorism attributed to Khusrau I >by, a complex set of mechanisms for maintaining socialAnosharwan explicitly recognized the state's dependence boundaries. Even apart from some practices whose dis-on productive agriculture as the principal source of its tinctiveness rested directly on formal religious grounds,own prosperity (Adams 1965, p. 71). The scale and sophis- therefore, it is obvious that there may have been sometication of planning that was involved in the irrigation significant differences with regard to basic economic in-system assures us that there were competently staffed stitutions and matters of subsistence between them andengineering offices. Something is known of such offices other religious communities.for the Islamic period, at which time there were cadres In addition, there are spatial and temporal discontinui-of specialists in different phases of construction and elab- ties. The main concentration of Jewish settlements seemsorately codified techniques for leveling, water lifting, and )to have lain well to the north and west of the ancientwater allocation (Wiedemann 1906; Cahen 1947-48; lower and central Euphrates floodplain region primarilyBosworth 1969, pp. 151-58; Jabbar 1973, pp. 24-28). dealt with in this study. Neresh is apparently the onlyThe model on which they were organized was certainly named village or town that was regarded as a part ofof Sasanian derivation, at least in its main features. Other the Babylonian Jewish community, whose location onspecialists in considerable numbers would have been the Nahr al-Nars is likely to have placed it as far southneeded for cadastral surveys, and it is inconceivable that as the latitude of Nippur and on the central Babyloniantheir work could have been conducted without extensive plain. Most of the major towns instead lay on thewritten records that have entirely disappeared. By con- Euphrates north of Babylon, on the transverse canalstrast, the fiscal authorities may not have been especially connecting the Euphrates with the Tigris, or along thenumerous, although it was only their activities that seem Tigris above and below Ctesiphon (Obermeyer 1929, p.to have received the notice of the heads of state. Quite 242 and passim; cf. the convenient summary in Neusnerpossibly individual records were not necessary, with taxes 1966, p. 248). Furthermore, the Talmud must be under-being as a rule levied on a whole village, town, or dis- stood as a later compilation and redaction of materialtrict, collected by the community itself, and then turned covering a fairly long span of time. The greater part ofover as a unit to crown agents three times a year (New- Newman's evidence on agriculture seems to derive orig-man 1932, pp. 163-64; Lokkegaard 1950, pp. 139-40; inally from the encounters and disputations of the thirdbut cf. now Goodblatt 1979, pp. 259, 270). There were and fourth century rabbis, while the period of greatestalso crown lands, of course, that must have been man- interest here is the apogee of Sasanian settlement in the

202

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

sixth century. Although the differences may not havebeen large, it is troubling not to be able to control for theunconscious, later smoothing away of archaisms.

Herein lies the most pervasive limitation of the Talmudas a source, for it goes to the heart of the picture it canprovide even of the major Jewish settlements in theirown place and time. Neusner has succinctly, if perhapstoo pessimistically, summarized the difficulty:

What is most significant is the unreliability of Talmudicmaterials for the history of the Jewish community. Itindicates that the final collectors and editors were notconcerned about historical reminiscences, which musthave survived, but only about other matters mainly per-taining to events and opinions within the academies.[1969, p. 47]

We cannot too often remind ourselves that all we knowabout Babylonian Jewry consists of what the rabbis choseto transmit in their schools. Nowhere does that factbecome more striking than in the study of the life andculture of ordinary Jews. We have limited archaeologicaldata, the magical incantation bowls and the paintingsin the synagogue at Dura-Europos. Rabbinic literatureprovides little persuasive evidence about what the lattermay have meant, or what the former were used for. Ourconsideration of the external structure of the Jewishgovernment of Babylonia quickly came down to studyof germane rabbinic sayings and stories. It was the limitedusefulness of these data that become in the end the mostobvious and convincing result of our inquiry. [1969,p. 125]

If I find Newman's contribution more useful than thisseems to suggest is possible for the Talmud as a source, itis of course because Neusner's objectives are not thesame as those of the study. The purpose here is not a de-tailed reconstruction of institutions and a flow of historicevents, but merely a sketch of enduring features of routineagricultural life that can complement the fiscal and martialpreoccupations of the crown and the mute ruins of townsand canal levees. Page references are included in the fol-lowing account only where an observation by Newmanis directly quoted:

Wheat was apparently the staple crop, at least in thecircles in which the rabbis generally moved. It was clearlymore valuable than barley, although in the homes of thewealthy it might even be fed to dogs. Following wheatin order of importance were barley, spelt, rye, oats, rice,and millet. Bread was made not merely of flour fromgrain, but also from rice, millet, and even lentils. Pulseswere also of considerable importance. Seeming to followin order behind them was the cultivation of vines (blackand white grapes were distinguished), dates (Persian andinferior Aramean), sesame, flax, vegetables, and cuscuta(for brewing beer). Hemp was cultivated in some districts,both for cheaper varieties of cloth including shrouds for

the dead and for rope. Spices were much used, includingmost commonly pepper, ginger, and mustard, and per-haps were also cultivated locally.

Some features of this pattern seem more likely tohave been restricted to the Jewish community (or even toan upper stratum within it) than others. The dispropor-tion of wheat over barley is an example, although thismust be partly a result of geographic factors. Wheat andbarley both were kept as major crops through the entirehistoric record in middle and northern Babylonia, incontrast to the south, where barley almost completelyreplaced wheat by the second millennium B.c. becauseof its greater tolerance for salinity. To judge from early'Abbasid tax receipts, on the other hand, even in theadministrative districts falling within what had beenmiddle and northern Babylonia barley accounted forabout five-eighths of the combined production of thesetwo cereals (Jacobsen 1958, pp. 12, 46, 52-53).

Newman's suggestion that grapes preponderated overdates is a similar case. He notes that date beer was morecommon in the south and wine in the north, and it seemscompletely beyond question that the date palm has at alltimes been enormously more significant than the vine inthe agricultural economy. Again there may be a religious,class, or even ethnic explanation, the grapevine havingbeen better adapted to the Palestinian hills from whichthe Jews had been forced into exile. He does note, inci-dentally, that "in the list of permanent investments inwhich a man is advised to invest his wife's money, datepalms precede vines, the order being: land, houses, date-palms, other fruit trees, and lastly, vines" (1932, p. 99).Moreover, date palms are observed to grow untended inthe countryside, without the care of a gardener.

Apart from the production of foodstuffs for sale, linenwas the most important article of commerce. There werespecial urban markets for flax, and some town neighbor-hoods were known for their specialization in flax-soaking."Of such importance was linen to Babylonian trade that... public prayers, during which the Shofar was sounded,were offered when this commodity fell in price to three-fifths of its value" (Newman 1932, p. 23). Various formsof oil for lighting as well as cooking also must have beencirculated commercially. Sesame-seed oil was the staple,much more common than olive oil, although the latterapparently was not prohibitively priced. Also mentionedfor these purposes are cottonseed oil, pitch, melted animalfat, and fish oil.

Domestic animals have as yet received no mention.There were religious or customary prohibitions on theownership of herds, apparently related to conflicts overpasturage rights, but numerous references indicated thatJews at least participated in herd ownership and were ex-tensively involved in growing animal fodder. Moreover,Jews could and did serve not only as shepherds for othersbut as specialists in the breeding and fattening of animals.

203

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

Chickens, ducks, and geese were also raised, and there areeven many instances of fish-breeding in special ponds.

Cultivation practices that are mentioned include fallow-ing, crop rotation, intercropping of grain and vines (onewonders how this was classified for tax purposes!), andcross-plowing with oxen. Canals are ubiquitous. The com-mon use of lifting machinery in connection with some ofthese canals is suggested by religious conventions speci-fying which kinds of flowing water were suitable for wash-ing. Manuring of fields was widespread, usually by ar-rangement between a landowner and a herd ownercovering the maintenance of a flock in a particular field.Perhaps because of the extensiveness of this practice, fieldswere frequently fenced with stones, wood, or staves inter-twined with twigs. Handmills, donkey-drawn mills, andwatermills, in increasing order of size, are all referred to.All in all, the impression is one of a thriving, fairly inten-sive, surely diversified agricultural system. There had beena number of important introductions since the Achaemen-ian period (probably including rice, although it was pres-ent earlier in Khuzestan) and a market rather than a sub-sistence orientation may well have become dominant. Yetthere was a darker side that should not be lost sight of:" 'Ten measures of poverty came down to the world, nineof them were taken by Babylonia, and the one was dis-tributed among the other nations of the world,' is ananonymous saying of the Rabbis, and this was furtheremphasized, in the Gemara, in the ensuing discussion, bythe statement 'Babylonia is the place for poverty' " (New-man 1932, p. 24).

There is a third textual approach to the condition andadministration of the Sasanian agricultural economy,alongside the fiscal preoccupation of the major annals andthe elements of the subsistence background that occur inpassing in the Talmud. It consists of references to ante-cedent conditions that occur in Arab sources after theconquest, or of institutional and othlirpatterns that theArabs may be inferred to have taken over with littlechange from the Sasanians. There is little doubt that thisis potentially by far the richest source of all, but it is alsothe most beset with problems. Moreover, the scatteringand immense diversity of the material calls for the life-time work of specialists, not the summarization of sec-ondary sources to which a study like this is largely lim-ited. The sounder course here is to avoid doubtful mattersof judgment as to putative Sasanian antecedents of Islamicpractices and to deal with the latter only as they apply tothe specific periods of time in which they are attested.

Compounding the difficulty is the fact that there areessentially no surviving Arabic sources that deal at firsthand with the Mesopotamian alluvium during most ofthe first century after the conquest. The systematic re-cording of traditions began later, under the influence ofpartisan political and religious currents that make evenreferences to the seventh century often tendentious and

anachronistic. In particular there were strong inducementsto Persians to trace the main features of the new Islamiccivilization growing up around them to pre-Islamic ori-gins. To disentangle the thin thread of reality from a tan-gled skein of conflicting assertions in these circumstancesis perhaps as challenging a task as there is in Orientalistscholarship. Fortunately, Michael Morony (1972, 1976)has made a most useful beginning on it.

With regard to institutional continuity between Sasa-nian and Early Islamic times, Morony has shown thatthere was considerable differentiation both by region andby social status. Persian settlement had been heaviest eastof the Tigris and in certain garrison cities, but there wasalso a lesser aristocracy of dahaqin (sing. dihqan) who ei-ther lived on their rural estates or lived in the towns as ab-sentee landlords. Accounts of the loss of life accompany-ing the conquest are surely much inflated, and significantnumbers of the aristocracy not only survived but came toterms with the conquerors and initially retained theirprivileges. Over the longer run of sixty or seventy years,however, Morony concludes that many military and ur-ban defectors and their descendants were integrated intothe essentially different, Arab society of a new set ofgarrison cities like Basra and Kufa. The dahaqin, simi-larly, were first compromised and then increasingly dis-placed by a new class of Muslim landlords (1976, pp.46-47, 51-52, 56).

At first glance this suggests the fairly intact transmissionnot merely of Persian outlooks and life-styles but of largecomponents of the administrative system. Apart from thegeneral disappearance of the highest levels of the old socialorder and administration (with some significant excep-tions even there), Morony is at pains to stress the lengthof time that distinctively Persian elements remained inplace and the multiple channels of transmission that wereavailable. But it must also be said that ultimately a newand quite different integration was achieved-and thatthis was accomplished before any significant number ofstrictly contemporary accounts come to our assistance:

It should be noted that the disbanding of the asawira(Persian military contingents) coincided with the ruiningof the dahaqin, the conversion of the Magians at Hira toIslam, the change of the language of the tax bureau fromPersian to Arabic, and the coinage reform. These almostsimultaneous changes underscore the impression thatmost of the direct survivals from the Sasanian periodlasted for about sixty years after the conquest, until ca.700, before they either disappeared or were integratedinto a new Islamic civilization. [Morony 1976, p. 57]

Not to anticipate a fuller description of Islamic condi-tions yet to come, these observations have a direct bearingon the interpretation of Sasanian administrative practicesinsofar as they affected the settlement and irrigation sys-tem. Continuity in nomenclature is widespread. It has

204

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

long been known that most of the named Islamic towns,canals, and administrative divisions were Arabic deriva-tions of Persian, Aramaic, or even older terms. But howsafe is it to assume an essential continuity in the underly-ing demographic, economic, and territorial realities towhich those terms refer? There were symbolic as well asmaterial advantages for Persian partisans after the con-quest to claim that this was so-that aside from a newreligion, a new Arab elite (into which they would hopeultimately to be assimilated), and some new cities theformer patterns merely reasserted themselves after an in-terval of disruption. There are also advantages for thescholar, who only on this assumption can hope to estab-lish pre-Islamic geographic entities (leaving aside thelarger towns and cities) with any confidence. But is ittrue? And if reason can be found to burden with numer-ous qualifications the assumption that it is, what does thisimply about the smoothness and degree of continuity thatwas maintained through the Sasanian-Islamic transitionin other respects?

Here we must take cognizance of one of the centralfindings of the archaeological reconnaissance. At least tojudge from the intensively surveyed region, there was amajor reduction in the settled area at-very roughly-theend of the Sasanian period. Much of the lower centralfloodplain of the Euphrates disappears from the archae-ologist's view at that time and seemingly remained out-side the scope of sedentary life for perhaps as long as amillennium. This was, of course, the region known to'Abbasid geographers as the Great Swamp, but its descrip-tion in these terms probably should not be taken to implythat more than a limited portion of it was continuouslyunder water. Moreover, its designation as a swamp doesnot describe the etiology of those conditions. Can itsformation be understood solely as a consequence of floodsand perhaps other natural disasters, or only as the out-come of declining cultivation and maintenance capabili-ties that left the land more or less continuously open tosuch disasters?

Unfortunately, the timing and duration of the processof abandonment that in any case ensued must remainelusive inasmuch as archaeological surface collectionsprovide the major evidence. Therefore we must leaveopen the question whether we are witnessing the fairlyabrupt collapse of settled life. If this were so, it wouldnot entirely decide the further question of the respectiveimportance of the natural and social contributions to it.But it would lend weight to an event or series of eventssomehow connected with the conquest and its immediateantecedents. The alternative is that there was instead abroad regional-if not more general-economic declineor population shift, or both, whose effects gradually in-tensified during the first century or so of the Islamicperiod. What seems incontrovertible is only that there wasa substantial and to some degree permanent withdrawal

coinciding with Early Islamic, and perhaps terminal Sasa-nian, times.

A number of contributory factors can be suggested,most of them not mutually exclusive. Direct physical de-struction and population loss consequent upon the con-quest itself may well have been the least significant. Onthe other hand, it is much more likely that the canal sys-tem was profoundly disrupted owing to prolonged lack ofcoordination and maintenance of what had become ahighly complex and interdependent as well as large-scalenetwork. It should not escape our notice that the areasmost affected were the remote tails of the system, wherethe consequence of an attenuation of supplies farther up-stream would have been felt most directly. We must alsotake account of the cumulative effects of ecological de-terioration and Sasanian maladministration in the latersixth and early seventh centuries, including-but by nomeans limited to-the disastrous flood of 628. Perhaps,in fact, it was the conjunction of these factors, beginningwith prolonged disarray in late Sasanian times and cul-minating in the piecemeal but ultimately complete over-turn of the established order, that led to a disruption muchmore profound than would have resulted from any singleelement. What is significant here, however, is not so mucha fuller specification of the causes of the breakdown andabandonment; that must be left as a problem for furtherresearch. At this stage we can recognize only the massivescale of the demographic decline that ensued and its his-torically unprecedented abruptness at least relative to thatscale. It should therefore come as no surprise if, as a re-sult, the break with older patterns of administration andnomenclature was correspondingly large and abrupt whenreconstruction finally began in earnest.

More detailed justification can be found for this ad-mittedly tentative proposal. On grounds of size alone,Zibliyat (site 700) and Jidr (site 004) were two of theimportant cities of southern Iraq during the Sasanianperiod. Their ancient names must be among those towhich some reference continued to be made in Islamictimes. Yet an identification eludes us. 8 Dwarfing them,and virtually all else, in importance in the eyes of Muslimanalysts was the little buffer state of al-Hira. Its impor-tance in Arab eyes is understandable, both because of thepivotal role it played in earlier politics in the Arabiandesert and because it lay directly athwart their line of ad-vance on the Sasanian empire. But it was credited withonly six thousand head-tax payers on the eve of the con-quest (Hitti 1966, p. 391). Its yearly average productionof 30,000 kurr of grain (about 87,000 tons or 1,080,000hectoliters) would, at Achaemenian yields (above, p. 000),which are the latest available, require a mere 540 squarekilometers of cultivation (Kister 1968, pp. 151-52). For aprincipality of these modest proportions to loom so largein the consciousness of later Muslim historians, the latterwould have had to be quite out of touch with conditions

205

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

in the Sasanian empire at the time of its maximumstrength.

Let us further consider the gigantic canal system thatserved both Zibliyat and Jidr as well as Nippur. This waspresumably the Nahr al-Nars, for at least in the thirteenthcentury Yaqut placed Nippur along its banks. Yaqut at-tributes its construction to the Sasanian king Narses (293-302), and he and all the classical Islamic geographersspoke of its source as having been an offtake from theLower Sura canal somewhat downstream of the town ofHilla (Le Strange 1895, pp. 256, 260). But a mere glanceat the succession of canal and settlement patterns on theMesopotamian plain in Sasanian and Early Islamic times(figs. 44 and 47) shows that the upper part of this courseis a later renovation, almost at right angles to the mainaxis of the system. What had clearly happened was thatthe Sarat-Nil canal, dug only in the time of the great'Umayyad governor Hajjaj ibn Yusuf at the beginning ofthe eighth century (Le Strange 1895, p. 261), interdictedthe former canal's line of flow so that a new source hadto be found. This means in turn that no memory appar-ently had survived even among scholars and tradition-alists of the different upper course antedating the timeof Hajjaj.

No less in need of clarification is the pair of canalsknown as the Upper and Lower Zab, after which no lessthan three administrative subdistricts (tasasij, sing. tassuj)were named. In the Islamic period they fronted along theright bank of the Tigris, from above al-Na'amaniya toone day's march above Wasit (Le Strange 1905, pp. 38, 73).Easily 50 kilometers, this distance suggests either that thenames have proved elastic again or, more probably, thatthe canals were originally components of a fairly extensivesystem. Yet all our sources are vague or in conflict overwhere they lay. It is impossible to avoid the suspicion thatlittle more than the name and the region to which it veryroughly applied may have come through the Arab con-quest and the upheavals of population that followed it.

Similar doubts arise when we turn to the locations ofother administrative districts. As I noted previously, mostof the names betray a Persian origin. But some of themalso betray a distressing tendency to wander. There is thecase of the two tasasij that were both called Falluja andadministratively placed within the same kura, of UpperBiqubadh, one containing the ancient town of Falluja atthe very upper limit of the Euphrates alluvium and theother entirely detached from it, surrounded by tasasij ofa different kura, and a 100 kilometers downstream (Mo-rony 1972, chap. 1). Or, again, the tassuj of Barusma wasunambiguously attached to Middle Biqubadh and placedwell to the north of the Sarat-Nil canal by Ibn Serapion atthe beginning of the tenth century (Le Strange 1895, pp.255-56). Yet elsewhere (Fiey 1968, pp. 170, 253), withequal specificity, we are told that the same, presumablyrather modest subdistrict also bordered on the subdistrict

of Nippur that lay 100 kilometers southeast. A possibleexplanation for both these anomalies lies in district di-visions originally having been assigned by the Sasanians onthe basis of territories served by particular canals. Barusma(in Sasanian times Beth Rushme), for example, would thenhave applied to a long, narrow but integrated domain be-fore the advent of the Sarat-Nil system. But the point isthat the Muslim historians and geographers were clearlyunaware of the contrastive usages of an earlier time thatmight have helped to explain their own confusion.

These discrepancies have many common elements, buta common explanation cannot necessarily be found forall of them. Still less is it possible to reach any generaliza-tion that applies more widely, as to the extent of undis-torted administrative knowledge that occurred across thelengthy and confusing Sasanian-Islamic transition period.But the difficulties do seem to justify treating the canaland settlement patterns of the Sasanian period essentiallyin terms of the physical remains they have left, rather thanon the basis of what some Muslim authorities later saidabout them.

Five maps are relevant as we turn to a consideration ofthe settlement and irrigation patterns of the region be-tween the Tigris and the Euphrates. Figures 44 and 45deal with the immediately applicable findings of archaeo-logical reconnaissance, the first for the greater part of theMesopotamian alluvium and the second for the intensivelysurveyed region. Figure 36, already discussed in connec-tion with the initial presentation of data on demographyand the urban hierarchy, is a simulation of areas under cul-tivation in the Sasanian period from the data in figure 45and table 21. Finally, as was the case only to a lesser ex-tent in the discussion of the Seleucid-Parthian period, theLANDSAT imagery (fig. 6), and to an even greater degreethe complete recording of surface traces of ancient canalpatterns that is to be found on the base map, is of consid-erable importance.

The incompleteness of the settlement layout recordedin figures 44 and 45 must be stressed at the outset. Someof the factors involved have been previously presented,but others have not. Moreover, an understanding of theirinterrelations and combined significance is more essentialfor this period than for any other.

In the northernmost part of the plain the exceptionallynumerous and massive sites of the later periods were forthe most part not recorded (Adams 1972). By hindsightthat is regrettable, but the reconnaissance there was con-ducted in 1956-57, at a time when a focus exclusivelyon Sumero-Akkadian and Babylonian history still seemeddefensible. In any case, the application of a survey ap-proach to the unfamiliar ceramics and very extensiveremains of the later periods initially presented many diffi-culties. Fortunately, a mo1e intensive resurvey was car-ried out a decade later (Gibson 1972), covering some 1,500square kilometers in the district around Kish that is

206

oi.uchicago.edu

C" 4

I

\\ V

0 25 Km.

Fig. 44. Sasanian sites and watercourses on the Mesopotamian plain.

207

/'

oi.uchicago.edu

$cltp\c

0

0

0,

(4\

Fig. 45. Sasanian period settlement patterns in the intensively surveyed region.

208

+

· · '~ ·

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

fairly central to the northern part of the plain. But, ad-mittedly, a reduced rate of site recovery is unavoidablethroughout the upper part of the plain because of rela-tively widespread and intensive agriculture, large-scaleland leveling and drainage projects, road construction,and many other forms of disturbance associated with themodern development of the Iraqi economy.

Around Kish, as in the region that is the principal focusof this study, Sasanian sites seem to have been the mostnumerous. The demographic significance of this is evengreater, in that in at least some parts of that district sep-arate sites cannot readily be distinguished since settlementwas almost continuous (Gibson 1972, pp. 49, 57). Thesame generalization can be shown to apply to the Diyalaplains on the opposite bank of the Tigris. If we excludeBaghdad and Samarra, on the grounds that these wereimperial capitals drawing resources from an immenselylarger region, by all odds the time of most extensive occu-pation there, too, was the Sasanian period (above, table19; cf. Adams 1965, tables 19, 20). There are no reason-able grounds to doubt, in other words, that settlement andpopulation throughout southern Mesopotamia was at itshighest during that period. It is only an accident of thelocus of modern development in the country, and of thefocus and timing of archaeological surveys, that so littlecan be said of the northern part of the alluvium. Thecanals shown in figure 44, presumably the antecedentsof the Nahr al-Malik and the Nahr Kutha of the Islamicperiod, accordingly are located largely on the basis ofmajor surviving spoil banks and levee systems, and theycannot yet be shown to be of Sasanian date on any othergrounds.

We face somewhat similar difficulties in generalizingbeyond the effective limits of archaeological recon-naissance in the southern part of the alluvium. It was orig-inally argued in connection with the computer simulationof figure 36 that very large, continuous tracts must havebeen under cultivation in order to support the populationdensities found within the intensively surveyed area. Thesmulation program's output in fact coincided fairly closelywith the limits of the intensive survey itself (except inareas that seem to have been at most briefly used becauseof the growth of swamps). Since the limits of the surveywere for the most part the quite arbitrary (and alreadysuperseded) ones of the 1962 frontiers of cultivationshown in the air photographs, one may reasonably sup-pose that further extensions of intensive reconnaissancewould simply enlarge the pattern. Can this supposition betested further?

Here the base map becomes the most relevant source,considered in combination with the Sasanian site layoutrecorded in figure 45. It will be seen at once that mostof that map is covered with an unmistakable grid ofancient canals. Detectable also on the ground, those on thebase map have in the main been traced from air photo-

graphs. Segments of the grid are in use again today, theold levees offering obvious advantages for siting newcanals to maximize command of the land for irrigationpurposes. But the grid to be seen on the base map is notin any sense connected with the rapid expansion of culti-vation that has come in the last century or so.

This conclusion rests in part on the highly selectiveand discontinuous use made of the grid by modern canals.It is also shown by the irrelevance of the modern culti-vation frontier to the layout at every point. Still morestrikingly, it is shown by the fact that numerous branchesof the grid are interdicted by, and hence are obviouslyantecedent to, the Shatt al-Hilla branch of the Euphratesitself, both above and below the towns of Suq Lemlum(site 1474) and Rumaytha (on the southwest margin of thebase map). Lemlum was already occupied in essentially itspresent location at the time when the earliest Europeantravelers followed the line of the Euphrates (Longrigg1925, p. 2; Niebuhr 1968, 2: 251). Rumaytha may havebeen only a nineteenth century town (cf. Cadoux 1906, p.186), but in any case it existed on the Hilla branch at atime long before the presence of more than an extremelythin and marginal fringe of agriculture there. Clearly,therefore, the virtually continuous grid of canals that con-stitutes the bulk of the evidence recorded on the base mapis antecedent to the virtual abandonment of the regionthat had occurred by the Middle Islamic period and thatcontinued in much of it until the threshold of moderntimes.

Furthermore, the preponderance of Sasanian remains inall surveyed regions very strongly implies that this was thetime to which most parts of the grid must be traced. Somemay be of Seleucid-Parthian origin, especially those alongthe southern margins of the area that afterward were in-creasingly abandoned to swamp. But the virtually dou-bling of the aggregate area of recorded settlement betweenParthian and Sasanian times (cf. table 19) makes the laterperiod much more likely for most of it. Similarly, the evensteeper decline in aggregate settlement after the Sasanianperiod (cf. table 18), coupled with the northward retreatof the settlement frontier and the absence of substantialadditions to the pattern anywhere (cf. figs. 45 and 47),makes an Early Islamic date of construction for any sig-nificant part of this grid extremely unlikely. The con-clusion is inescapable that most of the virtually continu-ous ancient canal system to be seen in detail on the basemap dates to the Sasanian period.

Some further amplification of the Sasanian canal layoutin the still unsurveyed area adjoining the lower Euphratescan perhaps be gained by consideraing the LANDSATimagery (fig. 6). The detection of some of the ancientwatercourses suggested in this map may be questionable,since the minimal units of definition are many times largerthan in an aerial photograph (above, p. 33). But it is sig-nificant that certain canals shown in figure 6 correspond

209

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

rather closely with ones shown on the base map and infigures 44 and 45. In particular, a comparison of thesemaps suggests that a major lower section of what is nowthe Shatt al-Hilla was once a canal, and that its courseseems to form a further element in a grid with Sasaniancanals to the north. Also shown in figure 6 are a seriesof apparent canals that once crossed or intersected withthis course. At least to judge from the LANDSAT imagery,some of the water supplied to this area was routed to itfrom a western branch of the Euphrates-apparently theonly branch at the time-in a position approximating thatof the modern Hindiya branch. With a whole series ofmassive diversions farther upstream, it is not unlikely thatin Sasanian times the Euphrates entered the swamps notfar from the location of the modern town of Shinafiya onthe Hindiya branch with very little if any residual flow.

I should stress once more that we cannot view thisoverall pattern as the product of slow, cumulative trendsthroughout the Sasanian period. What is known of thedynastic history of Sasanian times suggests instead thatmajor agricultural expansion came primarily in the reignsof a handful of monarchs who somehow combined ex-ternal military victories with temporary dominance overthe landed nobility, and that the high point in the processwas probably achieved early in the sixth century. If this isso, then not only the canals but the settlements that orig-inally must have accompanied them would have been inuse for only a relatively short time. For the settlements,this implies in most cases a relatively slight buildup ofoccupational debris. Since much of the area where thisgrid occurs is once more in cultivation, and since rapidalluviation is therefore also continuing over much of it,the prospects of ever adequately recording the communi-ties that were newly founded to care for this immenseextension in the cultivated land surface are, frankly, dis-mal. But any comprehensive assessment of the scale ofthe Sasanian achievement in Mesopotamia must neverthe-less take their existence fully into account.

The area in which the new Sasanian canal grid can bemost clearly identified is a broad band running fromnorthwest to southeast. To the north, this band is perhapsbest defined by what is now called the Shatt al-Nil, amajor trunk canal (shown by a heavy line on the basemap) that continued into the Islamic period and that inplaces still forms a broad, discernible depression. LargeSasanian towns line the banks of the canal, but the mostimportant cities-at any rate Zibliyat and Nippur; Jidris badly obscured by sand-are situated at short distancesoff to the south. This may mean that the original trunkcanal, the one reputedly built by Narses at the end of thethird century, paralleled the later one but lay farthersouth.

In any case, one can see that the countryside for along distance to the south is divided into north-southstrips by a fairly regular network of parallel, unusually

straight and hence carefully laid out branch canals. Atintervals these are intersected by other canals, either atright angles or along diagonals, whose uniformity of orien-tation again indicates that they were laid out accordingto a large-scale, carefully surveyed plan.9 The effect wasto open up a very large new area for cultivation by divid-ing it into polygons of varying size, principally rectangles,trapezoids, and triangles of from as little as 20 or 30 to1,000 or more hectares. Forming the boundaries werebranch canals of varying capacity and, more important, ofvarying sources. In fact, the provision of multiple sourcesof water for essentially every enclosed field suggests aconscious element of "overdesign" in the entire system. Itseems very likely that an important criterion in the plan-ning process was the ability to substitute one water sourcefor another. This would have encouraged the balancingof supplies and requirements across a vast area whose in-dividual, cell-like compartments had no way to communi-cate quickly with one another.

The same considerations that define this canal networkas primarily if not exclusively Sasanian also apply to thecomparable network that parallels the right bank of theTigris. In the latter case there is independent support forthis dating since the central town of al-Nu'maniyah (nowal-Na'amaniya) and its surrounding administrative dis-tricts were reportedly founded by the Lakhmid an-Nu'manibn Mundhir (580-602) of al-Hira (Morony 1972, chap.1, sect. 1). But this implies additionally that a number ofmajor canals, shown on the base map and in the LAND-SAT imagery (fig. 6) to have derived their water from theTigris, are also very likely to be of Sasanian origin.

Here then we must recognize a system whose enormousgrowth had led to a transformation of its requirementsand mode of operations as well. It was shown much earlierthat the Euphrates alone was adequate for only some8,000-12,000 square kilometers of cultivation annually,principally because fall irrigation needs had to be met at atime when the river's volume was still low. Now itemerges that lands under at least periodic cultivation inthe Sasanian period were vastly more extensive than this,extending over the greater part of a floodplain five or sixtimes larger than even the higher of these two figures.Even with the omission of somewhat less than half of thisowing to the practice of fallowing, well over twice as largean area must have been simultaneously under irrigationas the Euphrates alone could supply. Consistent with thatobservation are the data on density of settlement (wherethey exist), on the planning for utilization of alternatewater sources, on the actual construction of offtakes thatcould have been fed only from the Tigris, and on theamount of cultivation that would be needed to supplythe state's reported revenues if the land tax were theirprincipal source. The Sasanians succeeded in introducingsuch extensive cultivation, in other words, that it presup-posed an irrigation system comprehensively planned to

210

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

use Tigris as well as Euphrates water to meet minimalirrigation requirements.

This should not be taken to imply that Tigris waterwas in widespread, continuous use throughout the centralfloodplain, or that Euphrates and Tigris supplies werenormally deployed there in unison and without geo-graphic restriction. The critical perod was during the falland early winter months, and this was also the time whenthe Tigris was relatively more manageable. Hence its mostextensive use is likely to have occurred at that time. As thespring flood approached, it was probably drawn upononly for cultivation along its own levee backslope, whichcould not be provided for as readily in any other way.Similarly, it would have been difficult and wasteful tobring Tigris water westward to areas close to the Euphra-tes. This suggests that there may have been a practicalbasis for the customary distinction made by Muslim ge-ographers like Ibn Serapion (Le Strange 1895) betweenlands irrigated by the Euphrates alone and other lands irri-gated by the Euphrates and Tigris together. Especiallyincluded among the latter was what had been the Sasaniancrown domain of Kaskar, across the river from whosemajor center the Islamic city of Wasit was later founded.In the late Sasanian period the kura of Kaskar extendedwestward to include Nippur as one of its subdistricts(Morony 1972, chap. 1, sect. 1), so that most or all thegreat plexus of Sasanian canals serving Zibliyat and Jidras well as Nippur lay in the area formally defined as hav-ing been served by both rivers.

In many respects it is the comprehensively planned andexecuted, large-scale approach to the use of the Tigris thatepitomizes the Sasanians. While there is no evidence thatit lay within their powers to place a weir directly across sobroad, swift, and dangerously variable a body of water,they were clearly prepared to make the massive invest-ments and to assume the risks that went with constructinglarge diversions from it. The existence of a policy of thiskind is most conclusively shown by the largest Mesopo-tamian canal of all times, the Katul al-Kisrawi-Nahrawansystem that served the entire lower Diyala region on theTigris left bank. Its long intake had to be deeply incisedthrough the headlands of Samarra, and it followed acourse that necessitated crossings of both the Adheim andthe Diyala rivers before it reached the major cultivableareas it was intended to serve. Apart from the technicalcompetence and uniformity of design displayed in manydetails of its construction, what is most striking is theambitiousness with which the Sasanian engineers plannedit to transform the basic drainage patterns of a vastalluvial landscape (cf. Adams 1965, pp. 76-79).

Yet while a grand design is clearly apparent behind theKatul-Nahrawan's general layout, we must also recognizethat it relied to a considerable extent on preexisting irriga-tion works. Figure 46, newly derived from an analysis ofthe air photographs, primarily shows the system of other

canals that existed before the introduction of the Katul-Nahrawan in the time of Khusrau I Anosharwan (A.D.531-79). The positions of the Katul and the Nahrawanare also shown with a dotted line, however, and it is evi-dent that much of the latter merely pieced together orlinked up-while of course also greatly enlarging-earliercomponents. After its construction the entire system wasmore or less rapidly rebuilt so it could be supplied fromthe Nahrawan as the major source of water, but howmuch of this had been accomplished before the end ofthe Sasanian period is not clear. Its fully reconstitutedform, with the Nahrawan as the unquestionably centralaxis, is to be seen only in the map referring to the EarlyIslamic period (fig. 51).

No systematic body of Sasanian surface collections orrecorded site observations is available comparable to thatfor the pre- and protohistoric sites discussed in chapter 3.But the ruins of the period are remarkable for more thantheir vast extent. Much of the Sasanian pottery is draband seemingly mass produced, and the proportion ofglazed wares is small and also unexciting (see appendix tothis chapter). Well-fired bricks, on the other hand, arefrequently found in profusion. Bits of copper or bronze aswell as iron are fairly common even on the surface. Ex-tensive if generally shallow looting of Sasanian cemeteryareas on older sites seems to indicate that valuable seals,signets, and perhaps coins regularly occur in graves. Large,carefully made basalt milling stones abound, presumablyhaving been brought down the Euphrates from Syria.Above all, there is a marked increase in the quantities ofglass to be observed on the surface as compared withParthian and all earlier sites.

In this connection, let me call attention to the series ofsites specializing in glass manufacture along a newly dugSasanian canal northwest and north of Uruk-Warka (par-ticularly sites 1532, 1533, 1534, and 092). The scale ofwhat can only be described as industrial production thereis suggested by mounds hundreds of meters long that ap-parently are composed mainly of glass slag. Numerousglass furnaces also can still be seen in place.

To understand the resource base that led to such spe-cialization, it is worth noting that in late Sasanian timesthe area south of this canal line probably had begun toinclude extensive swamps. Here would have been foundlarge accumulations of snails as a source of calcium car-bonate, for they still occur in seasonal and permanentswamps in this area today in almost unbelievable num-bers. Suitable water-laid sand might also have been sup-plied locally. Finally, naturally occurring plants in thesame area are reported to be a source of sodium carbon-ate.'0 There is little to suggest that these sites were ex-tensively involved in fashioning glass vessels from the rawmaterial they produced in such quantities. But the im-portant point is that, in difficult terrain at a great distancefrom the main centers of administration and consumption,

211

oi.uchicago.edu

0 10 25km.

S 0.1- ha.* 4.1 10 ha.

* 10.1- 20 ha.* 20.1-40 ha.

4 ,0.1-200 ha.

i 200. ha.

Nj

0luBAR-

0

0

Fig. 46. Sasanian watercourses and settlement patterns on the lower Diyala plain.

212

*

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

a large-scale industrial operation went on (probably intothe Early Islamic period) that implies a fairly high degreeof economic integration not merely in the cities, but ex-tending throughout the countryside.

Virtually all the features of the Sasanian system thathas been brought forward in the foregoing discussionhave two aspects. On the one hand, they reflect thedevelopment of increasingly complex integrative mecha-nisms, and on the other hand, they show a new depen-dence on those mechanisms that was not smoothly re-versible. The articulation of regionally specialized glassproduction facilities with distant urban markets is oneexample, but the irrigation layout is perhaps the mostcompelling one. Not merely its initial design but its con-tinuing repair and operation depended on knowledge andresources that simply could not be supplied by autarkiclocal villagers in the event of a breakdown.

The scale of the canal network, in the first place, meantthat breakdowns might occur far from the districts wheretheir effects were most serious. Second, the "overdesign"feature mentioned earlier frequently would have giventhose nearest a silted-up or breached supply canal a num-ber of alternatives by which to compensate for their loss,but an accumulation of a number of such problems ordeficiencies upstream would have left those nearer thetails of a canal network with critical shortages for whichthere was no simple remedy. Third, there is every reasonto believe that there were weirs, sluice gates, and per-haps other components of the system that were particu-larly vulnerable to damage (natural or man-induced).Their continued, effective functioning presupposed a cen-trally supplied staff of specialists, and perhaps the capitalresources of the imperial treasury as well. Nothing isknown directly of irrigation facilities in this immediateregion, to be sure, but the sophisticated Sasanian engineer-ing works on the Nahrawan canal in the lower Diyalaregion (Jacobsen 1958, pp. 87-91; Adams 1965, figs. 18-21) are virtually certain to have had close parallels here.In particular, there is a major fan of radiating branchcanals at site 1213 that surely indicates a weir of impos-ing dimensions and other control works as well. For allthese reasons, accompanying the whole program of agrar-ian expansion was an increased dependence on central co-ordination and control.

Yet we have also seen that political stability remainedas remote a goal as it had ever been. Faced with internalrevolts and powerful external enemies, the crown couldmake no realistic provision for the long-term continuityof agricultural management. Its tax "reforms" rigidifiedas well as increased the burdens on the agricultural popu-lation, imposing fixed demands that in variable naturalcircumstances could not always be met without extremehardship. Its preoccupations, in fact, were with maximiz-ing fiscal returns, on which its own immediate strengthdepended, and with the possibility of windfall spoils from

military forays into Byzantine and other foreign territory.Its urban and agrarian policies were pursued less for theirlong-term effectiveness than for their immediate effect onthe shifting balance of power between the crown and thelanded nobility.

For all the grandeur of the Sasanian economic achieve-ment, here lay its crucial weakness. Each new expansion-ary step reduced rural self-sufficiency and tended to placethe whole mechanism more and more at the mercy of de-stabilizing political forces. In the absence of any means ofcontrolling those forces, the possibility of sudden, deep,and tragic oscillations in the supply of the basic necessi-ties of life for great masses of the population loomed everlarger.

There were other built-in flaws as well. The develop-ment of a latticework of long, intersecting branch canals,each gradually building up a levee of its own, made eachenclosed polygonal cell an internal drainage basin. Actingtogether with an improvement in the supply of irriga-tion water, this greatly increased the dangers of salinity.The requirement of drainage if this problem was to bealleviated may have begun to be recognized, for the nameof one canal implies that it was at least partly designed tocarry off excess water (Morony 1972, chap. 2, sect. 1). Onbalance, however, the problem could only grow worse.The extensive band of new canalization south of Zibliyatand Nippur that I have already referred to, for example,had the effect of obliterating an important avenue ofnatural drainage. Additionally, as both the populationand the state's own fiscal demands increased, more landhad to be brought into cultivation regardless of the de-clining returns that might be expected from it. In time,therefore, we must assume that average productivity be-gan to decline. Output in those areas that were most ad-versely affected would have declined precipitously, andthe population dependent on them now would have hadfewer and fewer alternative areas to which they couldturn. There is no way to determine how much average liv-ing standards would have declined, but for some of thosecaught at the margins the decline must have beencatastrophic.

A further, potentially serious loss of flexibility involvedthe role of livestock as a subsistence alternative. Largeherds have always represented a form of' investment. Arelatively secure reserve in time of crisis, they provided ameans by which a displaced group could bridge a diffi-cult transitional period and take up life in a new area.Herds can be partly maintained on stubble, on volunteergrowth in fallow fields, and on limited, strictly controlledgrazing of young shoots of barley. These sources are notlikely to be adequate for the whole year, however, at leastif the number of animals is consistently kept near its upperlimit. Uncultivated or waste tracts (within or beyond thelimts of cultivation) thus play a crucial part in the keep-ing of herds, particularly seasonally filled depressions

213

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

whose shrinking margins can yield excellent forage in thelate spring and summer. This niche declined steeply insize as a consequence of the immense expansion of theSasanian cultivated area, however, so that an absolutedecline in herd carrying capacity seems almost certain.With the human population having grown rapidly inresponse to the extension of the irrigation system, there-fore, the ratio of this animal reserve to potential humanneeds must have fallen sharply. Dependence on cultivatedfodder was a partial substitute, whose importance we haveseen attested in the Talmud, but at some point thisbrought herds and people into more direct competition forthe same set of increasingly scarce resources. Again, theeffect would not have been uniformly felt and may noteven have entered into urban (i.e., recorded) perceptionsor conscious state planning. But it must have had an ad-verse effect on more marginal groups, especially duringtimes of crisis that may have been preciptated by quitedifferent causes.

Thus we can identify several features of a classic syn-drome: increasing population density, reduction of flexi-ble reserves with which to meet periodic crises, reducedproductivity at least in the more marginal districts, andincreased reliance on the state administrative structureeven though the latter remained as vulnerable to disrup-tion as ever. To these we can add a final factor, perhapslargely consequent upon the others. Epidemic disease canbe expected to accompany all the circumstances just de-scribed, and indeed there is documentary evidence for it.Early outbreaks of plague are chronicled in the Talmud,occurring in the third and fourth centuries (Newman1932, p. 24). Probably much more serious and generalwere those that began in the sixth century and continuedinto the eighth. As yet there is no basis for even guessingat.their impact within Mesopotamia, but it has been esti-mated that the loss of life elsewhere in the Middle Eastassociated with the "plague of Justinian" (541-44) alonewas 20-25 percent (Dols 1977, p. 17). And while this obvi-ously would have acted to reduce the negative effects di-rectly associated with overpopulation, the social disrup-tion accompanying serious outbreaks of disease mightalso have hastened the breakup of the administrativesuperstructure on which the whole irrigation systemdepended.

In summary, then, there is another side to the Sasanianachievement. As we have seen to be the case also in thesmaller realms of earlier antiquity, many of the most im-pressive gains were fundamentally compromised by theabsence of anything other than qualified, short-term po-litical stability. In addition, whether consciously or not,short-term advances in scale, complexity, revenues, andall the other convenient indexs turned out to have beenpurchased at the expense of increasing ecological fragility.In this sense the essential dynamic of the Sasanian demisewas an internal one. Its replacement by a new Islamic

civilization changed the course of world history, but theprimary explanations for its collapse are to be found inits own decay rather than in external pressure (cf. Adams1978, pp. 332-33).

There is a fundamental fragility and transience to theSasanian accomplishments, in other words, a set of deepinternal contradictions that the dynasty's predilection formonumental building only intensified. Something of thiswas perhaps even apparent to observers at the time.Writing only a generation after the advent of Islam, Isaacof Ninevah surely had in mind the utter dissolution oftheir imposing facade of power in a passage that stillreverberates today:

They have entered it as an inn for a night and left itas travellers on a journey over the whole earth, withoutthinking of return. Some of them kings, some governors,some wise, some honoured. Some of them scribes, someorators, some judges, some commanders of armies. Someof them possessors of riches, some lords of goods. Andnow after their death there is neither the order of theirdegrees, nor the crowns of their government; nor theirdreadful thrones, nor their lordly pleasures, nor the praiseof those who honoured them. [Quoted in Morony 1976,p. 56]

The Islamic Period (Post-A.D. 637)

For two quite different and yet connected reasons, onlya more limited discussion of the Early Islamic and 'Ab-basid periods is appropriate than that for the precedingperiods. On the one hand, there has been a rebirth ofspecialized interest in the economic and social history ofmedieval Islam that promises important new advances inour understanding within a short space of years. It wouldcontribute only confusion to indulge in a further rehashof secondary sources, many of them no longer responsiveto the issues at the center of current debates. On the otherhand, the archaeological contribution that can and mustbe made to those debates requires little space for adequatepresentation. It is essentially a negative finding, docu-menting a precipitate retreat from a vast central area ofthe Sawad that continued for centuries and has reverseditself only in modern times, and also calling attention tothe lack of references to the local details of this processin the works of the major chroniclers. Either of these con-siderations alone would argue strongly for brevity; to-gether, they compel it.

The steps in the headlong Muslim conquest of Iran, andthen the advance westward across North Africa and intoSpain, took place on a greater, more familiar stage thanthe one with which this study is concerned. They needno recounting here. Nor does the strife and successionunder the early caliphate that first saw the rise of the

214

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

'Umayyad dynasty in Damascus (A.D. 661-750) and thenthe assumption of power by the 'Abbasids leading to theestablishment of Baghdad and its growth into a metropo-lis of unprecedented size. But meanwhile the only signifi-cant break in what almost seems a veil of silence withregard to the south-central Sawad during that long seriesof epochal events is in connection with Hajjaj ibn Yusuf'sresolute, repressive governorship of the province of Iraq inthe late seventh and early eighth centuries. He was de-monstrably active on the immediate peripheries of theregion and probably within it as well; his founding ofWasit and construction of the Nil canal connecting theEuphrates with al-Nu'maniyah have already been noted(cf. Hitti 1966, pp. 449-50; Perier 1904).

If the golden age of the caliphate was the time of Harunal Rashid (786-809), then the climb to the summit ofworld power had been steep. The descent was even moreabrupt. It signifies that the consolidation and expansionof Islam was a phenomenon on a different scale fromthose considered here, the outcome not of a patientamassing of men and resources within a delimited area,but of a spreading conflagration that for a long timeseemed to know no limits. Nothing that went on con-cerning the agricultural heartland of the caliphate shedsmuch light on that momentous process of growth. Butthere is not much doubt that the mismanagement andconsequent deterioration of the heartland was one of thedecisive, converging forces contributing to the dissolutionthat quickly followed.

The effect on imperial revenues has long preoccupiedscholars, as indeed it preoccupied learned observers at thetime. A recent article by David Waines (1977) not onlysummarizes present understandings of the fiscal as well asadministrative collapse that had overtaken the 'Abbasidrealm by the later ninth century, but also traces some ofthe deeper rural roots of the process. His account makesreference to archaeological findings in the Diyala regioneast of Baghdad (Adams 1965, chap. 8), but in this con-text his appraisal of the documentary sources is more im-mediately relevant.

Conceding a thicket of problems that includes copyists'errors and the difficulties of distinguishing normative fromdescriptive lists of receipts, Waines shows that there wasa decline by about half in gross imperial revenues betweenthe time of Harun and the second decade of the tenthcentury. The Sawad, which had supplied about a fifth ofthe total, or 100 million dirhams at the outset, was sup-plying little more than 10 percent, or 20 million dirhams,at the end of it. Erosive processes were not limited to themore distant extensions of the empire, in other words,but were in fact concentrated around its heart. More-over, they increased in intensity, with most of the lossoccurring between the time of Ibn Khurdadhbeh's records(845-73) and those of Ali ibn 'Isa (915). In many strategicand formerly prosperous tasasij, losses of 90 percent or

more were recorded in this period of less than a singlehuman life span.

The causes underlying the decline of agricultural pro-duction in the Sawad are to be found within a broad setof relationships between the 'Abbasid ruling apparatusand the rural population. This sector, primarily the labourforce of the agricultural system bore in large measure theindirect effects of 'Abbasid imperial attitudes and the moredirect effects of their administrative policies and politicalfortunes. When the rural population finally turned againstthis constellation of pressures, it had only recently fallenvictim of a wave of disorder resulting from conflict withinthe ruling order itself. The destruction to rural lifewrought by this conflict was then compounded by theconflict of large segments of the rural population againstthe ruling order. The damage to the irrigation based agri-cultural system caused by several decades of unrest waspredictable and is starkly reflected in the figures of 'All b.'Isa's tax roll.

One factor implicit in the emergence of the crisismay be expressed abstractly in terms of the 'Abbasid viewof the imperial edifice which differed in a significantrespect from that of the Sassanian. The 'Abbasids werethe inheritors rather than the creators of the vast irriga-tion system which supported human life and cultivation inIraq. For the Sassanians, the core of empire had been anetwork of artificial canals connecting town and country;for the 'Abbasids, the empire was a network of highroadslinking the metropolis with urban markets throughouttheir domains. Thus, orders of priority were different. Tothe 'Abbasids the irrigation system was a constant and,although they attempted certain renovations and expan-sion projects, what proved more decisive to the long runstability of the system was their tendency to disregardits depreciation. It is possible, too, that the 'Abbasidsviewed agriculture in the Sawad with a benign indiffer-ence so long as revenues flowed into the treasure fromother sources sufficient to satisfy their own appetite forextravagance and to support a burgeoning bureaucracyand a demanding military. [Waines 1977, pp. 295-96]

No brief, general assessment of this kind can do justiceto the parade of abuses to which the agricultural pro-ducers were increasingly subjected. Leaving aside thosethat were more violent and without shadow of legalism,the tax system alone defies the imagination. The basicrate of the land tax had been increased to a 50 percentshare of the harvest under the Caliph Mahdi (775-85),but there were many supplemental payments ranging fromsalaries for the collecting personnel and surveyors tostorage fees, enforced gratuities, and even charges for thepaper used for records. Manipulation of the (Persian)solar and (Muslim) lunar calendars frequently led todemands for payment in advance of the harvest, and onoccasion taxes were even demanded for the following yearbefore the current year's harvest had been gathered. Thespreading institution of tax farming escalated the rates

215

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

imposed and multiplied other demands upon the pro-ducer, enforced by torture, while the revenues turnedover to the state by the officials receded. Herein lies animportant caveat in using state revenues as the only avail-able index to the state of the agricultural economy. Inthe long run the growing difference between collectionsand amounts forwarded to the treasury might be a serioussource of economic damage, but in the short run it mis-takes growing administrative ineffectiveness for generaleconomic decline (El-Sammarraie 1972, passim, chaps.4-5, p. 189).

In the sequel, the depredations of state officals, notto speak of even more direct plundering by Turkish mer-cenary forces, provoked a mounting tide of rural violencein response. The Zanj rebellion was put down with greatdifficulty only in 883, after fifteen years during much ofwhich large areas around Basra and Ahwaz had slippedentirely outside government control. The Qaramita move-ment that followed it as the spearhead of resistance provedstill more widespread and intractable, so that by the earlytenth century urban administrators could only view muchof the countryside as a sea of unrest and hostility thatcould never be completely subjugated. Violence fed onviolence, in other words. The perimeters of state controldrew irregularly inward, and prospects for any con-structive, long-term approach to the agrarian economydiminished to the vanishing point.

A useful indicator of the increasingly circumscribedarea within which agricultural operations received anyform of state support is provided by the lists of weirsassembled by al-Khatib al-Baghdadi and Suhrab early inthe tenth century. Thirty are referred to, some of themalready at that time on canals that were not longer sup-plied with water. But even if we assume that all had con-tinued in use until late in the ninth century, their locationsare striking. Almost all are concentrated in the environsof Baghdad: south of Samarra, along the 'Isa and Nahra-wan canals, and along other, even closer waterways. Onlythree lie at a greater distance, two along the Euphratesand its effluents in the direction of Babil and Kufa, and onebetween Babil and al-Nil (El-Samarraie 1972, pp. 31-32;cf. below, p. 236).

It is ironic-but not surprising-that this devastatingprocess of retrenchment should have given rise to a schoolof thought and popular movement glorifying indigenous"Nabatean" achievements, especially those connectedwith the spread of civilization and the improvement ofagriculture. Even as actual conditions were deterioratingintolerably, exhaustively detailed compendiums were ap-pearing with elaborate botanical nomenclature and care-ful specifications of all the procedures and requirements ofgood husbandry (Cahen 1971; El-Samarraie 1972; Fahd1977). Much of the practical experience and learning thusbrought together and codified as part of the Shu'ubiyyamovement, with implications of strictly contemporary dis-

sent, may well go back to Hellenistic times or even earlier.But its relevance to the needs and opportunities actuallyfacing the agricultural population was minimal.

With an archaeological partiality for la longue duree, Ihave crossed over too quickly from the inception of theIslamic period to the time when conditions had beengenerated that were ineluctably leading to the collapseof the rural Mesopotamian sedentary life. This may beuseful as a means of anticpating directional trends thatare largely hidden in earlier, more fragmentary data, buta fuller look at demographic and fiscal data for the inter-vening period also is necessary.

The Muslims did not wait long to reinstitute the fiscalsystem the Sasanians had developed. Within ten yearsafter the conquest of southern Mesopotamia, 'Umar (634-44) had appointed a commission to conduct a cadastralsurvey of the Sawad and to raise the land-tax rates as asource of provisions for the army. The competence ofthe surveyors (or perhaps of their informants) is in somedoubt, since they concluded that there were some36,000,000 jarib in the Sawad of Kufa alone (Hitti 1966,p. 426). Following Morony in using the smaller jarib ofabout 1,050 square meters, this amounts to some 37,800square kilometers, which is about 9,600 square kilometersmore than the maximum area (as indicated by planimetricmeasurement) that in Early Islamic times might havebeen defined as the Sawad of Kufa on the basis of havingbeen irrigated from the Euphrates. The reported calcula-tion is made even more unrealistic by its failure to takeinto account what must have been very large untaxableterritories including the extensive areas that had beenpermanently submerged after the disastrous floods in 628.

Note also that the instructions to the commissionsharply distinguished lands watered by the Euphratesfrom those watered by the Tigris. Perhaps this indicatesthat the Muslims were still ignorant of the combinationof water resources that had become a fundamental re-quirement of the irrigation system. But it is rather morelikely, after ten years of their control of the government,to reflect the breakdown of some of the more complexfeatures of the system. Quite possibly separate zonesdependent on each river had in fact reappeared, with acorresponding sharp decline in the gross area that couldbe cultivated.

Irrespective of the defects in their calculations and theshrinkage of the land-tax base, the commissioners arereported to have collected the surprisingly large amountof 100,000,000 dirhams in taxes from the Sawad of Kufaalone. Included in this amount were said to be headtaxes on some 500,000 men, probably about a third ofthe total non-Muslim population (Hitti 1966, p. 428).Particularly if it reflects only the territory of Kufa as thatwas later construed, the population figure is extremelyhigh; by way of comparison, it is between two and threetimes the estimate given earlier (above, p. 149) for the

216

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

population of the entire alluvium during the Third Dy-nasty of Ur. Perhaps it is only a symbolically roundedexpression of a high estimate that was not based onactual enumeration, although such an interpretationwould disquietingly undermine confidence in the usabilityof other figures as well. Perhaps it also reflects a substan-tial, and surprisingly rapid, population movement afterthe conquest from other parts of the Sawad into the Kufaregion. However, that also implies the massive disrup-tion of an enormous, finely tuned irrigation system, withno possibility at all that its productivity could be sus-tained without a lengthy transitional period of recon-struction.

The Muslim population at this time was still heavilyconcentrated in and immediately around the new garrisoncities of Kufa and Basra. In the time of Ziyad there arereported to have been 80,000 soldiers and 120,000 de-pendents at Basra and 60,000 soldiers and 80,000 de-pendents at Kufa. 'Ali's slightly earlier registration of theKufan army in 658 placed 57,000 Arabs (including 17,000youths) upon its rolls, but only 8,000 slaves and converts.(Morony 1972, chap. 2, sec. 3). Many of the women, to besure, must have been of Persian or indigenous stock. Butthese cities stand out as swollen foreign enclaves of anexceptional character, abruptly superimposed on the pre-existing urban hierarchy and only very gradually becom-ing an organic part of it.

If we assume that most poll-tax payments were in thelowest category, some 85 to 88 million dirhams wouldhave had to be supplied from the land tax on the Sawadof Kufa. There is no way to determine the real taxablearea on which the commissioners depended, but their newrates certainly simplified their problem. Assessments wentup in every category, and in addition they were newlyimposed on a number of crops (e.g., cotton and sugar-cane) that had not been taxed earlier. Wheat quadrupledto 4 dirhams per jarib, while barley doubled to 2. A newtax of 1 dirham for each 2 jarib of uncultivated landsignals Muslim concern over what must have seemed aubiquitous and disturbing process of abandonment. Thetax on vineyards rose more slowly, from 8 to 10 dirhamsper jarib, but the rate trebled on ordinary palms andquadrupled on the finer "Persian" variety. And steep asthese increases were, they failed to take into account thefurther effect of the reduction in the size of the jarib. Thisalone would have served to increase the real tax burdenby almost a third (Morony, n.d., chap. 1, sect. 2).

Following a procedure similar to the one introducedfor the Sasanian tax schedules (above, p. 202), we cantentatively set aside secondary crops to reach a judgmenton the orders of magnitude involved. Let us further sup-pose that wheat and barley were of equal extent andoccupied perhaps 40 percent of the taxable area. The ratefor unoccupied land can perhaps be supposed to apply toan equal total area of cereal lands left in alternate fallow.

Cereals (including fallow) then would have produced1.75 dirhams per jarib in taxes. Date orchards are moreof a problem. We are told that they paid between 5 and10 dirhams per jarib, which at the stated rates of 1 dirhamper finer palm and /2 dirham per ordinary palm seems toimply a density of only 95 palms per hectare. This isclose to the United States standard of 100 per hectare, butthat spacing is designed for a setting where water is plenti-ful and emphasis is on producing uniformly high-qualitydates for a competitive market. It contrasts sharply withmodern densities of about 450 per hectare in the Basraregion and with the Neo-Babylonian standard usedearlier of about 227 per hectare (Wilkinson 1977, p. 93;above, p. 186). Does it imply a deterioration in orchardmanagement after the conquest? Alternatively, the densitycould have been set at an unrealistically low level pre-cisely to stimulate date cultivation. Steeply reducing therate as calculated per unit area while increasing the rateper palm would, in effect, provide a powerful inducementto growers to intensify their date production by plantingmore palms within a given area.

Setting aside that tantalizing but for the present un-answerable question, we can further suppose that anaverage orchard income of 8 dirhams per jarib was repre-sentative of the remaining 20 percent of the land not de-voted to cereals. On this basis, average receipts wouldhave been 3 dirhams per jarib and the roughly 85,000,000dirhams that were reportedly raised in the Sawad of Kufa(after allowance is made for the poll tax) would haverequired a tax base of more than 28,000,000 jarib, orabout 29,400 square kilometers of cultivable land. Thisis so slightly above the maximum limits of the Sawad ofKufa as calculated from Euphrates drainage patterns asto lie within an acceptable margin of error. But if dueaccount is taken of the very considerable areas to whichthe land tax did not apply, it suggests either that the re-ported level of receipts was a formally recognized butunrealizable goal or that the rates applied in practicewere often still higher than those specified.

Perhaps reinforcing these speculative reconstructionsare the next figures available to us, breaking down byregion the state income during the caliphate of Mu'awiya(661-80). The whole of the Sawad at that time paid120,000,000 dirhams, little more than that reported earlierfor only Kufa. There are wide discrepancies in differentreports of the amount of taxes Hajjaj was able to collectfrom the whole Sawad at about the end of the century,but again under 'Umar II (717-20) we learn of 124,000,000dirhams.

Morony, who is inclined to accept the Muslim figuresbut to remain skeptical of the Sasanian ones, concludesthat the Muslims probably managed to maintain some-thing close to the earlier level of state income. He doesnot question that there was some decline in the grosscultivated area ultimately constituting their main tax base

217

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

but believes that they would have been able to compen-sate in part by driving up the rates. Citing Ya'qubi's re-port that annual income from state-held lands in Iraqand its dependencies during the reign of Mu'awiyaamounted to 100 million dirhams in addition to the landtax, he further observes that there had been a dramaticextension of state property and the importance of itsincome as a result of vigorous land reclamation policies.There is also reference to large reclamation projects thatwere carried out by private developers and that similarlyremained outside the land-tax system (n.d., chap. 1, sect.2). Clearly, it would be an error to use receipts from theland tax as a measure of all forms of agricultural ac-tivity. It seems quite certain that the land tax was leviedon a substantially higher proportion of Sasanian produc-tion than was the case under the radically altered condi-tions after the conquest, even if the extent of the subse-quent reduction cannot yet be exactly specified.

The area involved in Islamic reclamation projects ap-parently lay mostly on the lower Tigris north of Basra,and perhaps to a much lesser extent on Euphrates outfallchannels into the Great Swamp south of Kufa. Surveyshave not yet been conducted in either of those regions, sothat there is as yet no archaeological basis for estimatingthe relative proportions of Sasanian and Early Islamicsettlement there. The evidence for a very substantial de-cline in land use in the central Euphrates floodplain (andto a somewhat lesser degree in the Diyala region) appearsto lend additional support to reports of a level of Sasanianstate income considerably higher than could be attainedafterward. If this central area were fully representative,in fact, a serious decline in income would have beeninevitable no matter how repressively the Muslims raisedthe rates. But with future surveys (assuming they arefeasible in the Amara swamps and the dense palm grovesnorth of Basra) we may indeed find instead that for somecenturies after the conquest the absolute decline remainedrelatively modest, with the precipitous plunge beginningonly in the late ninth century. In that case the types ofstate income would merely have shifted in their propor-tions, with new categories being introduced to match afinal shift of settlement away from the age-old domainof cities along the tangled web of former Euphrates leveesin the center of the alluvium.

In the end, then, we must turn once more to whatarchaeological evidence is available at present. It hasalready been shown that the overall decline in the aggre-gate occupied area within the intensively surveyed areaafter the apogee of the Sasanian period had reached about94 percent by the eleventh century or so and that thelarger, presumably more "urban" centers were particu-larly hard hit. But the mapped sequence of changes pro-vided in figures 47, 48, 49 and 50 may carry some addi-tional impact of its own.

The first of this series summarizes what is known of

the distribution of Islamic settlements in most of ancientBabylonia. It omits the region between Baghdad andSamarra that became increasingly pivotal, the Diyalaplain fed by the gigantic Nahrawan canal system, andthe entire area east of the Shatt al-Gharraf and along thelower Tigris that included Wasit and Basra. Moreover,though Kufa itself is shown, the region around it still hasto be left virtually empty of contemporary recorded set-tlements. In other words, the surveyed terrain shown herefalls in the interstice between most of the major centersof Islamic development. Certainly figure 47 cannot beregarded as adequately representing what went on aroundthese centers. The same is even more true of the followingthree figures, which are confined to the still more limitedgeographical frame of the intensively surveyed region.But the scale of the decline that these figures do conclu-sively document is nonetheless of great importance. Inthem we see the virtual abandonment of an area of per-haps as much as 10,000 square kilometers, which for mil-lennia had constituted the vital hearth of a rich andancient civilizational heritage.

The sequence of steps by which this abandonment waseffected is left somewhat obscure by ambiguities in thedating of Sasanian and Early Islamic surface collections.There are unresolved questions, discussed in the appro-priate sections of the appendix to this chapter, that havea direct bearing on more general interpretations thatmight be offered here. Broadly speaking, however, fourphases can be distinguished even if their chronologicalboundaries remain somewhat elastic and questionable.The first and most extensive is the essentially late Sasanianpattern (fig. 45) that has already been dealt with. Theforegoing discussion of the disruption accompanying theMuslim conquest seems to imply that the most significantretractive step between that phase and the next (fig. 48)coincided fairly closely with the conquest itself-naturallyincluding the years that immediately followed before aknowledgable, effective, centralized agrarian administra-tion could be reestablished. The chronicles seem to over-simplify grossly the protracted series of steps by whichthe southern part of the alluvium gradually was con-verted to lightly occupied swamps, but numerous refer-ences to disastrous floods in 628 (Le Strange 1905, p. 27)may highlight an especially serious episode that overcamethe recuperative powers of the system just a few yearsearlier. In any case, the degree of specificity thus sug-gested is quite beyond what the surface collections canbe expected to show, at least at this methodological junc-ture. On the basis of the collections alone, it is equallypossible that extensive abandonment had already gottenunder way in the terminal years of the Sasanian period,or, alternatively, that there was a slower but continuingprocess of withdrawal that continued through the seventhcentury and much of the eighth.

With the second phase, here termed Early Islamic, a

218

oi.uchicago.edu

1 Originating in Sasanian period2 Early Islamic (Pre-Samarran)3 Middle Islamic (Samarran-Late Abasid)4 Late Islamic (Ilkhanid and later)Sites without numbers are Sasanian-Early Islamic() Reduced or doubtful occupation* Trace of occupation

2Q3) 23 3 3

1234 3

23' -

" "1 3(23)

\ /, i,(23).

'" - Nu'm niy2

., " ,23, ,24. »4 .S23 3"' 23 34 " 4\1 )234' 323 233 .23 -3 3.

3 .4 / 3)323 )'-

'23(4)

:23 2!\.^2 ;.4

"2 -

"33'-

34" ' 3'-,4 341j~

(234)

/(234)

0 25 Km.rii MEOW

Fig. 47. Islamic sites and watercourses on the Mesopotamian plain.

219

?

( 231.

123.

(34)-

(234)

-(123)

I

oi.uchicago.edu

'N; \x4/

//°

\ a.-"~, -

-\\0 0~*K\ Nbffdr\

I 0

J/32ON

r I

C

$

U)

loo

C

L \

I/ /

10*I - A. /

00

1*41

o/

/t

SN\

\ '-

\~*

\.

C

cl*z

C

N\

\I

I IpQL

Fig. 48. Early Islamic period setttlement patterns in the intensively surveyed region.

220

0

\

0

+

oi.uchicago.edu

e

4/•

0

C-•9

\cOi -

%r

0

0

C.

~I. \

N

iN

N\

+/

/ -- ~/ , N

I /'~ ii

N/,

&

4'

C'I

pG

Fig. 49. Middle Islamic period settlement patterns in the intensively surveyed region.

221

c

ZU)

0

!

\

oi.uchicago.edu

,.,~ .' N

a C4'

/ \0I..

'N

C-

1. a

Oa

I

/

N

\

// t-,/ / N

I N ,a, \,/ \

, \-· a<

1ý

Fig. 50. Late Islamic period settlement patterns in the intensively surveyed region.

222

C.

U

\

a

\

\\

\

.4 I

o" C/

00

0

00

a+

10

((a

N .4

N

0r

- a=

N

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

decline had been generally experienced but was by farthe most severe in the southern part of the region. Isolatedsettlements remained there but probably were limitedto the crests of old levees in a more and more extensivelyflooded countryside. Supply canals entering the area fromthe lower Euphrates to the west seem no longer to havebeen fully functional, and quite possibly the occupiedareas existed only as enclaves with swamps upstream ofthem as well as below them. Farther north, extending tosomewhat below the latitude of Nippur, the surface evi-dence suggests a substantial thinning out of settlementeven though there the general growth of swamps cannothave been the immediate causative agency.

Evidence is offered in the appendix to this chapter forthe view that this pattern prevailed for some time into theninth century. On numismatic grounds, one might expectthat the onset of the next massive phase of withdrawalhad occurred before the middle of the century. On theother hand, as Waines notes, the "hard evidence" forpeasants leaving the land dates only from the "civil war"and siege of Baghdad in 865. This, he maintains, "wipedout any notion that the government's reciprocal functionof protection could be honoured" and "clearly triggeredthe prolonged rural reaction over the next decades" (1977,pp. 298-99). Only three years later, the Zanj slaves un-furled their banners and took up weapons in the southernmarshes (Popovic 1976).

Comparison of figures 45 and 48 indicates that therewere significant changes in the sources of water for theintensively surveyed region. The bed of the Nil canal is tobe seen at the northern end of figure 48, continuing theline of the Great Sarat canal eastward to al-Nu'maniyah.As I argued earlier, the original feeder supplies for theentire region had come in the main from the northwest.That approach was cut off with the construction of theNil, at the beginning of the eighth century if its attributionto Hajjaj is justified.

Presumably the new source that was found on theEuphrates below Babil and the modern town of Hilla(fig. 47) was developed at the same time that the Nil wasdug. It is not clear why this was done, since the old Narsnetwork could have been supplied by joining it directlyto the Great Sarat-Nil canal, in the immediate vicinityof the city of al-Nil. Perhaps that was done for a time,with the more laborious offtake to the west being createdonly later. But the Arabic 1:50,000 maps covering theal-Nil area fail to indicate a levee supplying the inter-connection, and I have been unable to examine the rel-evant air photographs.

The third phase is combined with the second in figure48. Sites that are shown with a square instead of circularsymbol contained at least a few sherds of "classic" Samar-ran sgraffiato pottery. As is noted more fully in the ap-pendix to this chapter, this pottery seems fairly certainlyto have been introduced on a considerable scale while

Samarra was still the imperial capital; that is, before 893.On the other hand, it may well have been in use, particu-larly in what was now becoming a remote, depressedrural region, until at least the end of the tenth century.

The settlement pattern suggested by this "Samarran"pottery may involve some slight withdrawal farther north-ward. Primarily, however, there was a very rapid thinningout of settlement throughout the whole region. As table18 indicates, the Samarran pattern is comparable in den-sity to that of the late 'Abbasid period. Taken at facevalue, the figures for the tenth/eleventh century occupa-tion even suggest that this came nearest of all to totalabandonment. But the difficulty is that the Samarransgraffiato ware of "classic" pattern on which the identi-fication of the period largely depends is a trade ware oflimited distribution rather than a widely made (andimitated) utilitarian ware. Thoughly relatively commoneven on some quite small sites in the Baghdad region,it is almost always limited to one or two sherds in surfacecollections from apparently contemporary settlements insouth central Iraq. Hence its absence at a particular Earlyor Middle Islamic site within the intensively surveyedregion cannot be taken as a reliable indication that thesite was abandoned during the later ninth and tenthcenturies.

Little can be said about the following two figures, forthey illustrate no otherwise identifiable settlements andcannot yet be articulated with known events or streamsof recorded tradition. Sedentary occupation of the centralplain between the Tigris and the Euphrates, from A.D.1100 or so until the onset of substantial tribal re-settlement in the seventeeth century, had little substanceor significance. Continuity was all but broken off, cer-tainly in the urban and civilizational patterns of life thathad flourished there so precociously, but probably alsoin the much more rudimentary sense of continuous habita-tion of individual communities.

However, it would misrepresent the nature of laterachievements to conclude this account by focusing ex-clusively on the near abandonment of the central Euphra-tes floodplain. Coinciding with that process of almostcontinuous demographic decline and withdrawal was,after all, a lengthy and initially flourishing, if ultimatelyalso deteriorating, epoch of Islamic world culture underfirst the 'Umayyad and then the 'Abbasid caliphates.With the movement of the cultural as well as politicalcenters to other regions, there is some justification fordismissing the south central part of the Tigris-Euphratesalluvium as insignificantly marginal. It is to the hinter-lands of the greatest of the new centers, Baghdad, thatwe must turn for the culmination as well as the denoue-ment of the record of Mesopotamian urban settlement.

Figures 51, 52, and 53 summarizes this later, and sub-stantially different, chapter of the record. They are basedon the newly detailed mapping of ancient watercourses

223

oi.uchicago.edu

0 10 25km.

* 0.1-4 ha.0 4.1-10 ha.* 10.1 -20 ha.0 20.1-40 ha.

40.1-200 ha.

;,I,, @ 200. ha. o ·.° ° \R

/0

Fig. 51. Early Islamic watercourses and settlement patterns on the lower Diyala plain.

224

* ýýo

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

in the Diyala region (fig. 8) that has been drawn up largelyon the basis of the aerial photographs. Baghdad itself isonly schematically represented. At the height of its pros-perity early in the ninth century it probably dwarfed bya full order of magnitude, in area if not in population,any earlier or contemporary Mesopotamian city. Thisenormously swollen scale of growth may well have hada number of distorting effects on the surrounding region.There is a puzzling absence of subordinate small settle-ments around its periphery, for example, where one mightexpect intensive truck gardening to supply the vast, di-verse, and discriminating Baghdad markets. Baghdad'shyperurban character in many political and cultural re-spects notwithstanding, perhaps we must think of a sub-stantial element of its population as being commuters-in-reverse who were employed primarily in agriculturalpursuits (albeit commercial rather than subsistence-oriented ones) for considerable distances into the sur-rounding countryside.

The map illustrating the Early Islamic period (fig. 51)reflects the mature impact of the Katul-Nahrawan canalsystem. Apart from the buildup of Baghdad after the mid-eighth century, evidently there was also a gravitation ofsettlement toward more distant parts of the region andespecially into the middle-lower Nahrawan district. Inthe neighborhood of Uskaf bani Junayd (Diyala site 734),the largest urban center on the Nahrawan, a proliferationof villages and small towns indicates a type of coloniza-tion different from that of the Parthian and Sasanianperiods. Whatever its nature, the result was that here therural population density reached higher levels than ithad ever previously achieved. Yet the map shows that alarge area to the north of the upper Nahrawan was at thesame time extensively depopulated. The familiar lessonto be drawn, applicable to other periods and on manygeographical scales, is that it is hazardous to generalizeabout population trends from localized and hence oftenunrepresentative data.

The Diyala settlement and watercourse patterns of theSamarran and late 'Abbasid periods have been combinedon a single map (fig. 52). This was an era of retrenchment,punctuated by repeated major crises and minor episodesof restoration (Adams 1965, pp. 84-89), and importantcanal-cutting initiatives or extensions of the agriculturalfrontier could hardly be expected. Sites largely or whollyabandoned after the earlier, Samarran part of the intervalhave been designated E, and it is apparent from theirdistribution that there was a massive abandonment ofvirtually the entire middle-lower Nahrawan region notlong after Samarran times. The accompanying deteriora-tion of the irrigation system has been described in detailelsewhere (Adams 1965, pp. 99-105). Only along thetails of what later became known as the Khorassan canal,paralleling the Diyala to the east, do sites primarily datingto the post-Samarran period (designated L) give some in-

dication of a modest reopening of older channels and re-settlement of a district that had for some time been nearlyabandoned.

Only for the late 'Abbasid period can the bed of theTigris through the 'Ukbara region north of Baghdad bespecified with reasonable certainty. Numerous traces ofsuperimposed meanders, accompanied by substantialancient settlements, suggest that the main Tigris coursehad been in this vicinity for a very long period (cf. fig. 8).It has not yet been possible to work out the respective agesor even the relative sequence of the earlier meanders, butthe final bed that was abandoned when the Tigris movedeastward about A.D. 1230 can be distinguished from theothers. This is the one shown in figure 52. Severalmeanders in its upper part are far too small for the Tigrisand typical only for a stream like the Diyala. They sug-gest that for some time before the complete shift to thenew bed the course of the Tigris was bifurcated, withmost of the water following approximately its moderncourse while a vestigial flow was somehow retained inthe 'Ukbara channel.

Figure 53, finally, traces the general withdrawal ofsettlement that had been consummated by the Ilkhanidperiod, after A.D. 1258. After its sack in that year by theMongols under Hulagu, Baghdad had shrunk back tobeing a city of fairly modest size rather than a metropolis,and most if not all other settlements in the region arebetter classed as villages and towns rather than as cities.The Katul-Nahrawan system had altogether ceased tofunction, and with its closure the entire southern part ofthe lower Diyala region (excluding a narrow and dis-continuous fringe along the Tigris) disappears from thesettlement record. The pattern of withdrawal and breakupthat had all but obliterated the central floodplain as alocus of sedentary life three centuries earlier was by nowthreatening to engulf even the immediate environs ofBaghdad. And, apart from Baghdad, it was only in a fewdesperately poor, increasingly tribalized enclaves that aprecarious degree of continuity was maintained until thebeginnings of the modern era.

Running through most accounts of European travelerswho passed through any of these struggling settlementson the threshold of our times is a tone of disdainful su-periority. Partly this must reflect their consciousness ofan elemental disjunction between the glories of the pastthat they envisioned and the contemporary impoverish-ment that they unquestionably saw. But partly also theirtone must reflect cognitive constructs that were fashionedout of the acquisitive and expansionist aspirations of theirown societies (Said 1978). They were generally on a kindof pilgrimage to ruins of ancient cities that held a highplace in their own cultural and religious heritage, andso it should not be surprising that they idealized the moreremote past. With this, however, all too often went animplicit assumption that they were its principal heirs

225

oi.uchicago.edu

0 10 25km.

Fig. 52. Samarran-Late 'Abbasid watercourses and settlement patterns on the lowerDiyala plain.

226

oi.uchicago.edu

0 10 25 km

* 0.1-4 ha.* 4.1-10 ha.* 10.1- 20 ha.* 20.1-40 ha.

40 1-200 ha.00

S

0

Fig. 53. Ilkhanid watercourses and settlement patterns on the Lower Diyala plain.

227

*

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

and a disbelief that others could exercise any real claim toshare in it. And with such attitudes in turn there tendedto be a pervasive lack of interest in the achievements ofintervening ages that would have established lines of in-heritance fully paralleling their own. Still awaiting any-thing remotely approaching adequate study, therefore,are the immense irrigation, urban, and other achievements

of the Hellenistic era and early Middle Ages to whichmuch of this chapter has been devoted. Even in a regionthat we primarily identify as the oldest urban heartland,those later remains-and the civilizations responsible forthem-must be regarded in the end as no less impressivethan all their predecessors.

APPENDIX

THE SURVEY DATA BASE: NEO-BABYLONIAN-LATE ISLAMIC SITESAND CHRONOLOGICAL INDICATORS

The archaeological data employed in this chapter arefairly heterogeneous and in part of quite inferior quality.Published, securely dated sequences of common, utili-tarian ceramic types that would be especially useful fordating surface collections are on the whole conspicuouslyabsent. Much less archaeological attention has been de-voted to remains of these later historical periods, in fact,than to the periods covered in the preceding chapters.From Seleucid times onward, the availability of coinsultimately will prove an inestimable advantage for datingpurposes, even though examples found on the surfacefrequently are too corroded for even very partial or pro-visional identification. But at present, in any case, there aretoo few associations of coins with excavated ceramicassemblages to provide more than occasional indicationsof the full spans of popularity of types occurring in thelatter.

Apart from the widely differing degrees of certaintywith which surface collections can be assessed, there islikely to be a cumulative improvement in the proportionof sites recovered for progressively later periods. In partthis is simply a matter of more recent remains beingsomewhat less likely to be obscured by various laterforms of surface aggradation. This effect is not likely tobe large, since the complex of ongoing erosional as wellas aggradational processes has made it possible to trace indetail the remains of many much earlier sites on whichchapters 3 and 4 are based. Much more significant was thegradual acceleration in the use of baked bricks, not merelyfor public buildings but for at least the footings or lower-most wall courses of ordinary houses. Occasional surfacefinds of baked bricks date to the third millennium andeven earlier, but as late as the mid-first millennium theycan be described as common only on a handful of thevery largest centers. Their frequency increases perceptiblyin Seleucid-Parthian times, and by the Sasanian period itis rare to find even a relatively small site without someevidence of their use. Finally, in the Islamic period theylitter the surface profusely and often are sufficiently pre-

served in their original positions to permit the extensivetracing of building plans with little or no surface clear-ance (cf. site 1013 in chap. 7). The significance of thisincreasing use of brick is twofold. In the first place, itgreatly retards wind erosion and thus preserves moundsto a greater height than would otherwise be the case.Second, it adds a characteristic irregularity to the surfaceremains of former settlements that is noticeable from agreat distance in the desert. For these reasons the rateof recovery of later sites obviously must be greater.

Apart from these few generalities, an account of thesurvey data base for the later historical periods must dealwith conditions so varied that a discussion of their com-bined features would be misleading. It seems more appro-priate to deal separately with the catalog of sites pertain-ing to different periods or groups of periods, appendingto each separate section a discussion of the ceramic"index fossils" on which site identifications have beenbased.

Neo-Babylonian, Achaemenian and Seleucid-ParthianPeriods

Chronological indicators are especially ambiguous forthis long span of time. An attempt was made to differen-tiate successive periods within it on the basis of what isknown of ceramics associated with dated building com-plexes unearthed in excavations. But the dating of theceramics admittedly remains very slippery, as I havealready noted in the foregoing chapter. This is reflectedin our inability to prepare separate maps for settlementdistributions in the Neo-Babylonian and Achaemenianperiods and in considerable uncertainty as to the arrayof settlement data that properly should be assigned tothe Seleucid period.

With respect to the partial conflation of Neo-Baby-lonian and Achaemenian data, for most purposes it isprobably sounder to combine the sites attested for eitherperiod into a single list. Moreover, as is noted onceagain below, the possibility must be conceded that some

228

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

of the traits associated with either or both of theseperiods may have been introduced somewhat earlier thanthe political events by which the onset of the Neo-Baby-lonian period is normally defined. Fortunately, this lastdifficulty is likely to have a very limited effect on theinterpretation of settlement patterns for this period, sincemost of the vigorous growth that occurred would hardlyhave been possible during the later years of the Neo-Assyrian occupation.

Turning to the Seleucid-Parthian period, it should firstbe noted that well-attested ceramic indicators are in factlargely limited at present to the Parthian period alone.There is known to be considerable continuity betweentypes occurring in Achaemenian and Seleucid contexts,and again between Seleucid and Parthian contexts. Atthis juncture, lacking adequate corpora of excavated,securely dated Seleucid pottery from this region, wesimply cannot state with much conviction that Seleucidcollections are more similar to those of the followingParthian period than to those of the preceding Achae-menian period. Hence grouping "Seleucid" and Parthiansites together, as I have done consistently throughout thisdiscussion, may involve mainly Parthian remains and verylittle that is Seleucid in date. If that is so, Achaemenianand Seleucid instead should have been hyphenated andthe Parthian period might better have been treated inde-pendently. The uncertainty obviously affects any develop-mental as well as historical explanations that may beoffered on the basis of survey data, basically requiringthat very little be said about the Seleucid period. Again,however, there may be a slight compensating stroke offortune in that such numismatic evidence as there is fromoutlying sites in at least the Nippur region points to moresubstantial Parthian-and specifically late Parthian-thanSeleucjd activity.

In short, there is a lengthy span lasting from approxi-mately the seventh century B.C. until perhaps the secondcentury B.C. during which the chronological placement ofsites on the basis of their surface collections is not veryreliable. This is not to say that sites actually occupiedwithin this span have in general been assigned to a differ-ent one (or vice versa), but only that attempts to impose in-ternal time divisions within it may well be questioned. Iwill note in a later section of this appendix that thereare a variety of other problems with properly datingceramic indicators for the Parthian, Sasanian, and Islamicperiods. In at least a narrowly archaeological sense, how-ever, these problems are less serious in that the relativesequence is quite firm so that controversies arise onlyfrom the need to assign absolute dates.

There is a further difficulty with Neo-Babylonian,Achaemenian, and Seleucid-Parthian sites, affecting theareas that have been recorded for them. Not infrequently,the remains of all three periods are heavily blanketed bymassive, overlying Sasanian occupations, making esti-

mates of size problematical. There was a like difficultywith many of the sites relevant to chapter 4, dealt within that case by assigning fairly broad size categories ratherthan making specific estimates of their dimensions in agiven period. The same procedure is followed once morein the accompanying tabulation (table 20). But the diffi-culty with it in this case is that, as described in greaterdetail in the foregoing chapter, the character of settlementitself seems to have undergone a change during the periodswith which we are dealing.

In becoming normally less nucleated, more amorphousand sprawling, individual settlements frequently com-prised a number of scattered mounds rather than a con-tinuously built-up area. This difference is too easily over-looked when sites are described in terms of gross sizecategories rather than more precise areas, and, in particu-lar, the discontinuous, sprawling type of ruin is seldomapparent beneath a later overburden of Sasanian debris.Hence the settlement areas credited to such sites maybe disproportionately larger than those listed in chapter4 for others that were actually of comparable size.

The problem is probably relatively most serious duringthe Parthian period. Neo-Babylonian and Achaemeniansites are generally more regularly bounded and con-toured-more similar, that is, to the earlier pattern oftraditional Near Eastern "tells." Sasanian occupations,generally being the largest of all, are much easier to recordaccurately. The limits of Islamic sites can be easily fol-lowed merely by observing the presence of highly dis-tinctive glazed wares, which in any case generally traceout circles of declining circumference with little or nolater overburden. Even for the Parthian period, however,this is probably a source of fairly modest errors. Only19 percent of the "nonurban" Parthian sites (categories1 and 2 in the table) were masked by a Sasanian occupa-tion of equivalent or larger size, although in the "urban"categories the proportion increases to 36 percent. In anycase, a column has been added to the tabulation referringto "terminal" occupation in order to call attention tothose sites whose assignment to a particular size categoryis not open to doubt because of a larger, later occupation.

Turning to the ceramic indicators on which the deter-minations in table 20 are based, there is every reason tobelieve that individual traits attributed to the Neo-Baby-lonian period may already have appeared during the pre-ceding century or so of Neo-Assyrian hegemony and mayhave continued into Achaemenian times or even later.This is in fact what the ceramic sequence at Nippurstrongly suggests: there is not a single type of exclusivelyNeo-Babylonian date, and only one quite minor onethat even can be limited to the Neo-Babylonian andAchaemenian periods. On the other hand, no less thanthirteen types can be attributed to all or some part (in-cluding the Neo-Babylonian period) of the span repre-sented by the so-called Assyrian (something of a mis-

229

oi.uchicago.edu

'.s3U00

*43ulaids

'9eg-oaN

NV

E.

I .

n n I v

%J

In L 10 10 10 -r- w

w

ol -a, 0 0 00 4 14 r4

~ C

z In

In Ln

I A

lo lo ID I

rLM

L

n tn

) I r m

t

~ n Ln 'n L

n 'n o

n 'n Ln 'n

%D

'D

O ̀

I K

-

"" I "m

l" I ·

S. 3uoz)

-4 ?!(\r N

rl I 1 r

-l I

cn a r- a, IT

10 I j rI

m 1-4

m r- .4 co 0, a

en -4L

, 10 10 10 10 ID

r, r, r, w ol w

ol a ··

O n D

r 1P

V-

-4 1.11 V

%

Ct r

v%

nj Dj wl o ol l ;;41

r-4 1 -4 v4 CJI

fn C"I M

I MI rn 171

.? 1 1*1 \ V

-4 i ý4 ý4 4

ý ý -4 ý4 4 4 4 4 -

ý --

-4 -4 ý4 ýq ý4 4 ý

ý ý

r, (7

Q r \ 1 0 r( L

M

%D

r V

In 00 0%

N C

" I? .

, IS

1 m

L

n o4

N

10 14 rl

OD

h

4 N l V

n 0 cr1

0 C

I r 0 mý4 -4M M

cn '

Ln

~i 1010

1 1 SOw w

a, 1 0

0 144 ý

4. ..

Ln r r

*qugqa-oaN

r 4

C4(

4ý 4 -

4 ýW

. I nl vr r( vr h rflO -t ~~ 0

~ ý

In W

T LM

M·~

_o ll(I t I

7a-s 0r

tv rIm l

i O -

o 0 1N Ir

p 61 b ;-ii ý4 in

n b 0%

C

N~

, , I

rll 0%

1 oIr al 01 01

ý4 R

i 0rý

0 0 00

ý4 ý4 ý4 -4 -4 4-4

r4 -4 -

z 4

4 4

44 4ý44-4-

4 4-4ý

4 4

r

4 4

e4 r 4 -

4 4 M 4 ý4 ý 4 4 ý 4r

4 r- -4 4

ý ý 4 C

rl co 0

%0 go C

71 0 4

IT

0 -C

4 1- 0

CID

4 r 0 0 cn I

m ol (71

"1 10 1 11 o

Il .1 w

M"

t' -4 1 1 11 ~I

I I ~ I

I I I I I I I

P2 So.

w w:, 0-1 m r- o c 0 0

r4 r4 N en c" I I"-

CO

i l~m

v ~~Irl vl o

0%

a m

a a, a, , 0%0, M 0, m

0, m

0%

a, 0%C

Y,

, 1-tn

ww

w w

Oi0

10 I"

*11132d-,Ta

s ý4

ý4 .4 4 ý4 ý4

4 ý4 4 4 04 ý4 r4 ý4 .4 O

n Cn 4 ý4 ý4

C4 en q -4 r I

-4 -4 4

4 tN r4 ý4 4 f

*qug-oaN

1, 1- 1-P

A

r4r4ý 1-4 41ý

r I

"4 I

I I 1

-1 1

1 1 1 1

1I I I I -.

.. -1 1 1 1 1 1 1 1 1 1 1 1 1 1

i~·r ;·rl yol N

I -l 1 1

~

1 1 1 .ý I I ..

I .ý 1 -1o

0 ~ N m

~1-4 IT

10 CC

) 0%

0

r, OD

m

m 00 CA

y-l 1? 10 10 rý -4 r- OD

-I z'J ID

C-4 t' 0 10 0

11 10 w 14 11 co In 10 r, 0% 4

v, (

en M m ý? 't

.41 In LA

L

A

ýv 10 %0

r 00 (7, O'ý 42, 0 0

0 ýf r4 C I Y

c i fn cn m c) rn -Il ct m

I n, tn %

v a o vi D

o

r,r,

P. P,

r, r,

o l

N r, rI r, r _

I r, r- , r, 01 co co 00 OD

I 0 co co OD

00 cc G

o 00 00 CO

1 00 CO

00 00 0 O

D

00 00 CO

CD

O

cli

-1 -4-4 C

N

N M M

M M

M 'n10 10 I ID

%D

0 r, rr, r

, m m w

0%

m 0%

a, ao

o 0 0 -1NN

%0 10 %

0 W %

C

10 10 10 10 10 10 10 10 10 10 10 ID

10 1 10 D

10 10 %

O

D

1010'O 10

10 %D

10 %D

%

0~

~ h~

~rI VI h m

o oDIoI

r~lhtiLn

-l r( I~I (·? 12 01 01 ll ?I I \ ~IQ

~Ir~-1 -1 -1 C

, I

ri u n D

,ý4

4 1

q -11 -

N

j ý q

M ý T

-

4 I 4 C

4 ý

~I rlrlr( V

\ 1r(3

~N

r 11

~ji I n rl

rlNIr~ rlrlO

l ~ rl

cnrv·

-l Ln tn

M

-t 4ý4 -

, N

-i,

(, , , : , , , , ,r

ý7 IT ? :D

ý In Inn Ln

A

n n n ' tn Ln'm

n m Ln n 'n Ln n n n v 1%

6n kn

In in %D

y

,4i -4i ý4

-l ý4 .1 "1 ý4 4 -1

-41 1" 1 ý4 ý4 4 4ý

C4

I 1 1

4 ý4 4!

4 41 14 14

ieni 10 1%

' D

r;,: oi

m_

_ _,

ý*

00

GI

0,-0(9r4u

230

',uo:)*q3Juad-*

aS-mavqav

*qeg-oaNw.4,uo:)

*qvg-oaN

.4,ruo:)*qllud-*TOS

-MOV43V

,qvg-oaN

2, juo:)*431a-*TOS

-meo*qvg-OON

W..ld-~a

*%,.4uoo

,q~jed-*Iab

*qvg-oaN

i5a

~n

\o\t

3( C

VC

VN

Z

4~&ill

h--- iilii

ý4 ý4 4 C4

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

nomer), Neo-Babylonian, and Achaemenian periods,while apparently not extending outside this span (Mc-Cown and Haines 1967, table 2). Clearly, there is areasonably satisfactory group of dating criteria availableprovided one is content with a span of attribution not lim-ited to the duration of the dynasty but instead extendingover three or four centuries. This greater imprecision willbe found reflected in the discussion of the relevant surfaceremains in chapter 5.

Five ceramic traits were used during the survey as diag-nostic of the Neo-Babylonian period: Outflaring jar orbottle necks with rounded or rope rims, usually with oneor more sharp-edged horizontal ridges at the middle of theslightly concave neck or in the transition zone between theincurving neck and the outcurving shoulder (McCown andHaines 1967( pl. 102); small, irregular pierced-lug or rope-loop handles, attached either at jar rims or on slopingshoulders (McCown and Haines 1967( pl. 104); deep, flar-ing, thin-walled bowls with short, concave upper sidesand near-vertical rims, sometimes with traces of whitishor greenish glaze (Adams 1965, p. 129); thickened or roperims on narrow-diameter jug or bottle necks (McCownand Haines 1967, pls. 104, 105); and lamps, usually un-glazed (McCown and Haines 1967, pl. 102). Burials inlarge urns also make their appearance at about thistime, and fragments of the urns are often seen on thesurface.

Many if not most of these Neo-Babylonian types mustbe equally applicable to the Achaemenian period. Fouradditional traits were identified with an Achaemenian oc-cupation: subhemispherical or somewhat flaring bowls,carefully shaped of an extremely thin-walled, well-levi-gated clay (McCown and Haines 1967, pl. 103:13-14;McCown, Haines, and Biggs 1978, pl. 50:11); a variety ofstamps and medallions applied, usually in a single hori-zontal band, on the concave, frequently profiled uppersides of round-bottomed, flaring bowls (McCown andHaines 1967, pl. 103:9, 16); horse figurines, frequentlywith schematically modeled riders (McCown and Haines1967; pl. 141:10; McCown, Haines, and Biggs 1978, pl.72:12); and "husking trays" with interior pecking orscoring and frequent evidence of heavy abrasion.

Turning to the Seleucid-Parthian period, essentially thesame substantial distinctive set of characteristics that wereoriginally applied to surveys in the Diyala region (Adams1965, pp. 130-31) were applied once more. It is apparentthat greatly improved differentiation of the dozen or sorecognizable traits into shorter time periods will ulti-mately be possible. R. J. Wenke has in fact made an im-pressive advance in this direction even without excava-tions, applying multidimensional scaling and matrix-or-dering to intensive surface collections from the Susianaplain while relying on surface finds of dated coins forchronological placement of his types and assemblages.Wenke finds that comparison with Mesopotamian sites

such as Seleucia tends "to support the general outlines" ofhis own study some 360 kilometers east-southeast. Hegoes on to concede, however, that there are "vast differ-ences" between the pottery types characteristic of thetwo areas, with some major types in one having "no realcounterparts" in the other (1975-76, p. 78). Hence it isnot possible to use his findings as a basis for refining thesequence in southern Mesopotamia at present; he merelypoints the way to the great progress that a systematiceffort comparable to his would permit.

It is perhaps worth noting the purely subjective im-pression that both Parthian coins and contemporaryglazed wares of all kinds are significantly less common inthe region covered in this survey than in Akkad and theDiyala region. Quite possibly the greater propinquity ofboth of the latter to the capital at Ctesiphon was the majorfactor accounting for this difference.

Largely reiterating the criteria used in the Diyala regionfor the sake of completeness, they are as follows:

Broad-line impressed decoration in a sawtooth or chev-ron pattern beneath thin Parthian-green glaze. (Surfaceexamples of Parthian glaze are almost always green,while excavated specimens are frequently blue or bluegreen. Clearly, there is a factor of weathering that needsto be borne in mind.)

Carved, low-relief decoration, triangular excisions, andappliqued buttom decoration beneath Parthian-greenglaze.

Single or double "twisted rope" handles, generally cov-ered with a thin Parthian-green glaze.

Thin, flaring bowls with a slight projecting elbow orcarination below a simple vertical rim.

Dishes or shallow bowls with flaring sides and beveledor downflaring rims, usually with light greenish glaze.

Outcurled, superimposed double jar rims on verticaljar necks, with Parthian-green glaze.

Thickened jar rims with wide, pronounced grooves onexterior and upper surfaces. This rim form normally oc-curs in association with double-rope handles and onvertical buff ware jar necks that are decorated with groupsof vertical incisions.

Punctate decoration, usually consisting of fine comb-tooth impressions in chevrons or intersecting patterns ofshoulders of jars like those referred to immediately above.In many cases, comb-tooth impressions alternate withlarger circular impressions possibly produced by birdbones or reeds. Comb-incised meander decoration alsomay be present.

High inflaring jar necks with thickened rims. Parthian-green glaze characteristically is applied to entire interiorand upper exterior.

Sasanian Period

As I indicated earlier, we are on substantially bettergrounds in discussing Sasanian settlement patterns on the

231

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

basis of the findings of an archaeological survey than weare for most of the preceding periods. The intrinsic im-portance of the period-the maximum urban and agricul-tural expansion within the entire ancient record-wassuch that an especially systematic effort was justified.Moreover, the ubiquity of Sasanian remains, a corollaryof that achievement, enormously simplified problems ofrecording on sites where other periods were frequentlyobscured by later debris.

Table 21 presents estimates of occupational areas for allsites whose occupation during the Sasanian period seemsreasonably assured. Following the major listing by grada-tions of increasing size, additional site numbers in somecases are given in parentheses. This indicates a relativelyhigher level of uncertainty as to the size of the Sasanianoccupation, not a greater doubt whether it was occupiedat all. Generally the source of the uncertainty is that therewas an apparently reduced Sasanian settlement on a Par-thian or earlier site, but with the proportion of the reduc-tion difficult to estimate because of the preponderance ofdebris of another period. In some cases it means insteadthat the settlement was dispersed over a number of semi-detached mounds, with some, but an ill-defined, cor-responding reduction in population density. No size esti-mates are included in this table for sites at which aSasanian occupation was regarded as "possible" or "prob-able" rather than fully confirmed.

The ceramic indicators used for the identification ofSasanian sites have received curiously little systematicattention, considering the crucial importance of the pe-riod. There have been only a handful of excavations, mostof them of modest scale and duration, even though inmany areas an almost overwhelming array of dense sur-face material has long seemed to demand more carefulscrutiny. A further difficulty is that much of what atten-tion has been given to the period has been in a traditionof art history, concerned primarily with exceptional speci-mens and refinements of style. Mass-produced utilitarianwares have received virtually no study, although, sincethey alone occur in sufficient quantities, they are muchmore likely to be useful for surface dating. Primary reli-ance in this study has therefore been placed on two recentbut quite limited undertakings that directly articulatewith survey needs and objectives: stratigraphic soundingsin 1969 at Tell Abu Sarifa that were explicitly intended toelicit a sequence (Adams 1970), and architectural surveysand soundings undertaken by the Deutsche Archaiolo-gische Institut in 1973 and 1975 at a number of Sasanianand Early Islamic sites situated primarily along the desertfringes of settlement west of the Euphrates (Finster andSchmidt 1976). References to earlier work, as well as somecomparisons to relevant findings in a geographically muchmore inclusive region, will be found in these two publi-cations.

Two problems with these accounts are worth men-

tioning, apart from the admittedly limited amounts ofmaterial on which they are based. The first is that bothdeal largely if not exclusively with excavated samples thatderive from the later part of the Sasanian period. Thelate Parthian-early Sasanian transition accordingly is stillvery imperfectly understood, and little can yet be said ofstylistic changes within the period that would permitus to identify occupational subperiods from surfacecollections.

The second problem concerns the Sasanian-Early Is-lamic transition. In this case there is now a reasonableand fairly consistent series of archaeological samples towork with, but their precise relationship to the epochalevents of the Muslim conquest of Iraq is not yet clearlyestablished. In other words, the known typological se-quence remains very loosely attached to absolute chronol-ogy; features currently defined as largely Sasanian mayhave considerable extensions into the Early Islamic period-or vice versa.

In any case, the abruptness of the political shift appearsto have had no counterpart in ceramic wares. As oneauthority has observed, "There is always an uncomforta-ble silence when Umayyad pottery, the earliest possibleIslamic pottery is mentioned. On the whole, the first 100years of Muslim rule have so far not yielded much in theway of strikingly new ceramic material and the generalfeeling is that pre-Islamic local productions were contin-ued without any dramatic technical discoveries" (Crowe,1977, p. 264). This conclusion is derived mainly from thestudy of glazed wares, which presumably were traded overgreater distances, since they were more expensive to pro-duce. So-called alkaline glazing was the standard of thetime, ranging from a grayish white to a deep blue basedupon varying admixtures of copper oxides. The commonimpression that such surface treatment was "the normalpractice for every day bowls and jars" (Crowe 1977, p.265) arises from the selective approach characteristic ofNear Eastern archaeology for the later periods and iscompletely out of touch with the prevailing consumptionpatterns of the vast majority of the population. Still, thesame generalization applies to the much more common,utilitarian, unglazed types. Older vessel forms and modesof surface treatment in the main merely continued withprogressive modifications, gradually supplemented bynewer ones reflecting an increasing predilection-though,at least at Abu Sarifa, never reaching even the 10 percentlevel (Adams 1970, table 2)-for glazed wares. Faced witha poorly anchored, prevailingly gradual transition of thiskind, we must bear in mind that attempts to distinguishbetween terminal Sasanian and Early Islamic surface col-lections are to some extent arbitrary.

Finster and Schmidt provide a good working definitionof the surface aspect of Sasanian sites, recognizable uponeven very brief and casual inspection: " 'Spitzfusstopfe,'Fingerdruckmuster, Kerbmuster, undekorierte Scherben

232

oi.uchicago.edu

TABLE 21 Estimated Occupational Areas in Hectares of Sasanian Sites

Site No.Area

0.1.2

.3

.4

.5

.6

.7

.8

.9

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2.0

2.1

2.2

2.3

2.4

2.5

2.6

2.7

2.8

2.9

3.0

651 962 1288 1545

195 530 609 650 834 839 1048 1136 1151 1494 1611

432 550 754 788 923 1202 1444 1485 1536

008 241 562 664 733 1106 1119 1335 1374 1502 1510 1524 1568 1618

250 652 655 771 922 1132 1224 1398 (647)

014 978 1147 1155 1207 1281

608 866 1122 1140 1221 1223 1255 1260 1566

015 619 661 851 1624

159 560 565 612 1191 1201 1212 1443

043 518 543 570 672 755 814 816 825 910 934 966 1097 1104

1167 1182 1204 1328 1397 1428 (665 A261)

799 1016

577 708 857 890 965 1150 1225 1246 1334 1381 1524 1588 1603

1527

202 645 690 725 927 1234 1263 1300 1382 1472 1559

058 222 588 654 663 701 860 868 942 1187

568 674 789 876 1073 1126 1176 1186 1195 1279 1297 1356 1395 1541

1544 1581 1614 1621 1633 A264

089 969 1068 1238

596 638 1076 1427

091 823 1254 1298 1380 1384 1596

544 594 717 929 1252 1422 1433

533 838 941 1556

513 525 616 637 643 683 698 873 902 955 957 984 986 1133

1142 1277 1402 1519 1525 1535 1619 1623

778

059 908 916 1093 1125 1266 1620

653 681 687 743 780 1216 1622

171 526 540 963 989 1128 1222 1557 1632

038 726 1078 1149 1268 1313

096 194 728 763 1010 1521

456 762 968

017 184 597 846 874 911 1084 1239 1292 1352 1408 1523 1543 (625)

849 1115 1153 1593

636 716 735 872 909 973 1243 1507 (245 329 573 675 803 1393)

182 992 1053 1294

545 1012 1040 1344 1375 1492 1567 1577 1580 1617 1637

684 774 821 1092 1586 (131 149 585 666 738 1197 1206 1311 1410

1430 1572 1583)

125 505 522 547 659 660 751 772 800 919 1029 1094 1228 1394

1416 1496 1497 1539 1561 1578 1616

918 1514 1613 1636

523 641 723 767 779 843 864 915 1018 1033 1041 1052 1060 1062

1173 1275 1302 1453 1573 (1011 1083)

903

122 223 569 646 732 833 840 862 938 1190 1192 1285 1296 1493

(520 535 600 740 871 1066 1082 1113 1174 1386 1570)

12.0

Area

4.1

4.2

4.3

4.4

4.5

4.6

4.8

4.9

5.0

5.2

5.4

5.5

5.6

5.8

5.9

6.0

6.2

6.3

6.5

6.6

6.7

6.8

7.0

7.2

7.3

7.5

7.6

7.7

7.8

8.0

8.1

8.5

8.6

8.7

9.0

9.2

9.3

9.8

9.9

10.0

10.2

10.9

11.0

11.1

11.2

1424 1560 (064)

924 1185

229 1116

(603)

1146

668 1262

768 914 946 (676

1529)

Area Site No.

12.2

12.5

12.6

12.7

12.8

13.3

239

623

587

521

1276

073 1130

1135

233

Site No.

1028

601 853 1014 1058 1095 1470

859 1085

1457

141 501 703 759 904 988 1047 1138 1326

180 906

591 1110 1117 1183 1220 1226 1343 1378 1548

559 791 1036

019 192 682 719 1209 1316 1374 1441 (519 534 639 764)

052 905 1258 1465 1582

085 809

925 1530

1286 1595

607 990 1131 1200

1065

138 913 1148 1189 (711 730)

062 093 121 571 581 1241 1415 1565 1629

863 926 1289

598 1203

554 617 1227 1532

1141

948 1087

035 088 214 381 599 704 1013 1435 1518 (512)

094

1426

1057 1242 1468 1627 A221 (983)

217 1184

099 1107

558 613 696 1026 1080 1533

238 595 845 1631

1121 1487

1210

985

1265

036 065 079 098 511 1039 1274 1314

566

1144 1287

200

1409

3.1

3.2

3.3

3.4

3.5

3.6

3.7

3.8

3.9

4.0

Area

13.5

13.6

14.0

15.0

15.8

16.0

16.6

17.0

17.1

17.5

18.0

18.2

18.9

19.0

20.0

20.2

20.3

21.0

22.2

23.2

23.8

24.0

25.0

26.2

27.0

27.2

28.0

28.8

29.5

30.0

32.0

34.5

35.0

45.0

46.5

49.0

50.0

52.2

54.0

54.2

58.0

70.C

72.C

160

230

Site No.

589

225

1417

090

1569

002 037 1101

1120 1214

(1229 1280)

072

709

056

510 1432

1079 1495 1537

614

842

1549

(213)

847

503

045

031

082

1436

134

055 532

1213

604 798 (582)

882

699

(621)

030

(265 NIPPUR)

063

1534

549

(URUK)

592

1273

1193

183

452

1310

092

196

656

700

004

Note: Areas of site numbers given in parentheses can be estimated only with a relatively high level of uncertainty. Generally this meanseither an apparently reduced Sasanian settlement on a Parthian or earlier site but with the proportion of reduction difficult to estimate,or a settlement dispersed over a number of semidetached mounds with at least some corresponding reduction in density. No size esti-mates are included here for sites at which a Sasanian occupation was regarded as "possible" or "probable" rather than fully confirmed.

___ __

----- . .

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

mit blauer Glasur" (1976, p. 168). This summary can bemore fully specified for analytical purposes as follows:

Very large storage jars with incurving, thickened rimsand a characteristic base that has been variously describedas a "Spitzfuss" and a "torpedo-fuse point." Rough red-dish ware, generally with an unoxidized black core; theinterior is sometimes partly coated with bitumen. Sasanianexamples tend to be almost cylindrical in shape and ameter or more in length, and they frequently have anotched, appliqued ridge below the rim (Adams 1970, p.100, fig. 6ce).

A variety of large flaring plain bowls with accented,sharply profiled rims and generally flat bottoms. Greatlyenlarged club-headed and ledge rims are common, oftenembellished with scalloped ridges either on their uppersurfaces or immediately underneath (Adams 1970, p. 99,fig. 6av-ax, bg-bj; Wenke 1975-76, fig. 7:227).

Small, plain buff or reddish stoppers or bowls with flatbottoms and everted upper sides, occasionally with pull-knobs centered on the bottom of the interior (Adams 1970,pp. 99-100, fig. 6bv, cc).

Rounded or flared thin buff ware bowls. Groups arenot infrequently found inverted and whole, exposed bysurface deflation. Their interiors often bear traces of Ara-maic or Hebrew incantations, suggesting that they weredeposited as part of a ritual. It may be of some ethnicor religious significance that their occurrence is apparentlydiscontinuous even on large sites with essentially contem-porary levels of occupation. Nippur, for example, has verylarge numbers; Zibliyat (site 700) has few or none. Otherbowls that are similar in size and shape are decorated onthe interior with incised multiple-meanders, perhaps combimpressions (Ricciardi 1967, figs. 132-33; Adams 1970,p. 99, fig. 6bu, 10al, an, 17).

"Honeycomb" surface decoration, applied primarilyto large plain buff ware storage jars. In varying combina-tions this consists of broad bands of cellular impressionsand trailed parallel channels, apparently impressed in athick, wet slip with the fingertips or spatulas of some kindWenke 1975-76, fig. 12:277; Finster and Schmidt 1976,before firing (Adams 1970, p. 102, fig. lOaa-ac, ag-aj, 16j;pp. 92-93. Abb. 48a-c, Taf. 48a-c, e-g, 52-55,58, 60-61).It was observed at Abu Sarifa that the earlier examples ofthis technique, putatively antedating the end of the Sasan-ian period by at least a century, were more geometric inarrangement and more carefully executed than the morenumerous specimens in the succeeding level, and that thestyle of decoration then rapidly disappeared. It has morerecently been pointed out (B. Finster, pers. comm.) thatthe earlier impressions also tend to be deeper, that thelater ones can be summarily described as having a "de-generate" appearance in comparison with the former, andthat this distinction is of considerable significance inordering surface collections.

Other forms of surface treatment applied to large

areas include a variety of patterned effects that have beendescribed as "riffling" and "pattern impressing" (Adams1970, p. 102) and as "stichelartiges Ritzmuster" (Finsterand Schmidt 1976, p. 92, Taf. 48d, 54b, 56, 61b).

Certain stamp impressions, principally applied to plainbuff ware jar bodies, are of Sasanian date. Characteristi-cally they "consist of curvilinear symbols or representa-tional motifs within a non-circular field," but they arestrikingly less common-and hence much less useful fordating in a surface reconnaissance-in south-central Iraqthan in the Diyala region and other immediate hinterlandsof the Sasanian capital at Ctesiphon (Adams 1970, p. 101).

Small, flaring, flat-based, or rarely low ring-based, cupsor bowls with a slight carination just below rim, coveredwith a thin, whitish blue glaze over a yellowish, veryfriable paste (Ricciardi 1967, figs. 175-76; Adams 1970,p. 106).

Probably beginning in the Sasanian period but morecharacteristic of Islamic times is a deep blue allover glaze.At least in later periods it occurs on large, deep, ring-basedbowls as well as large two-handled jars, but reconstructa-ble profiles of securely Sasanian date are not known in thisregion. Slight horizontal corrugations in the soft, flaky,yellowish fabric suggest that vessels may have been atleast partly formed through a coiling process. Interiors,also glaze-covered, typically fire to a lighter blue or a dirtywhite (Adams 1970, pp. 106-7).

Finally, note should be taken of distinctively Sasanianbaked bricks, common on all of the larger and many ofthe smaller sites. They are typically large (33-34 centi-meters square, 7-8 centimeters in thickness), well and uni-formly made, and reddish brown in color. Apparentlythey were relatively well fired, for they resist the effects ofdesert wind scour much better than their Islamic counter-parts.

Islamic Period

Remains of the Islamic period are more or less immedi-ately recognizable from the varied and distinctive glazedwares associated with them, although it is quite possiblethat for the first century after the Arab conquest, or evensomewhat longer, changes were introduced only gradu-ally. Profuse accumulations of baked bricks are anothercharacteristic feature, and fragments of glass seem con-siderably more numerous than they had been earlier.However, these obvious features provide little assistancein refining the chronology of a site within the Islamicperiod. Necessary as internal chronological subdivisionsare for our purposes, the reader should be warned thatthere is no wide consensus on them.

The Islamic sites recorded during the survey are tabu-lated in table 22, on the basis of temporal subdivisionsthat I will outline presently. Size categories are indicated,instead of measured areas. There is little to prevent theuse of a more accurate, measured standard for the Middle

234

oi.uchicago.edu

TABLE 22 Islamic Sites: Size Categories and Periods of Occupation

No. E M L No. E M L No. E M L No. E M L No. E M L No. E M L No. E M L No. E M L

002 3 - - 538 - 1 1 653 1 - 1 808 1* - - 955 1 (1) - 1117 2 - - 1238 1 - - 1369 - - 2

004 2 - - 540 1 1 - 656 1 (1) - 810 - - 1 957 1 - - 1119 1 - - 1239 1 - - 1370 - - 2

006 - 1 - 541 3 3 .3 657 - - 1 811 1* - - 963 1 - - 1120 3 - - 1241 (1) - - 1374 1 --

011 - - (1) 558 2 - - 658 - - 1 813 1* 1 - 968 1 - - 1121 2 - - 1242 (1) - - 1387 - - 1

036 1 - - 559 2 - - 660 1 (1) - 814 1 - - 969 1 - - 1122 1 - - 1243 1 - - 1388 - - 1

052 2* - - 560 1 - - 662 - - 1 819 1*(1) - 974 - - 1 1125 1 - - 1246 1 - - 1391 - - 1

059 (1) - - 564 - 1 1 670 2*(1) - 821 1 - - 986 1 - - 1132 1 - - 1252 (1) - - 1 3 9 3 - - 1

062 1 - 567 - - 1 674 1 - - 823 1- - 992 (1) - - 1133 1 - - 1255 1 - - 1395 - - 1

063 2 - - 571 - - 1 681 1 - - 838 1 - - 998 - - 1 1135 3 - - 1258 2 - - 1396 - - 1

064 (1) - 572 - - 1 682 2 2 2 839 1 - - 1002 - - 1 1139 2 - - 1259 (1) - - 1402 1 --

065 2-- 575 - 1 1 683 1* - - 843 1 - - 1004 - - 1 1140 1

- - 1260 1 - - 1406 2 --

085 2* - - 577 1 - - 686 1 - - 847 4 - - 1013 2* - - 1141 2 - - 1262 3 - - 1408 1 --

092 3 - - 578 - 1 1 687 1 - - 848 4* - - 1014 2* - - 1142 (1) - - 1265 2 - - 1409 2 --

094 2 - - 582 4 - - 697 - - 2 849 1 - - 1015 2* - - 1144 2 - - 1266 1 - - 1415 2 --

096 1 - - 583 - 1 1 700 2 1 1 850 1 - - 1016 1 - - 1146 3 - - 1268 1 - - 1417 3 - -

099 2 - - 587 1 (1) - 701 1 - - 851 (1) - - 1017 1 (1) - 1147 (1) - - 1273 5 - - 1422 1 - -

116 - () - 589 1 - - 702 1 1 1 852 1* - - 1022 1 - - 1148 2 - - 1274 2 - - 1424 2 --

125 1 - - 591 2 - - 707 1* - - 853 2* - - 1023 1* - - 1149 1 - - 1275 1 - - 1427 1 - -

131 1 - - 592 5 - - 708 1* - - 855 1 - - 1026 2* - - 1150 1 - - 1276 3 - - 1432 3 - -

138 1 - 593 - -2 709 3 -- 856 1* - - 1028 1 - - 1155 1 - - 1277 (1) - - 1435 1* - -

182 (1) - 594 1 - - 710 1* - - 859 2*(2) - 1030 2 - - 1173 1 - - 1279 1 - - 1436 4 --

188 - - 1 595 2 - - 712 1 1 1 862 1 - 1031 1 - - 1174 1 - - 1280 3 - - 1441 (1) - -

194 1 - - 598 2 - - 717 1 - - 863 2 - - 1035 1 (1) - 1176 1 - - 1281 1 - - 1444 (1) --

195 1 - - 599 1 - - 719 2 2 2 868 1 - - 1036 2 - - 1182 1 - - 1285 1 - - 1453 (1) --

200 2 - - 600 1 - - 725 1 - - 873 1 - - 1038 1 - - 1183 2 - - 1286 2 - - 1465 (1) --

214 2 - - 601 2 - - 726 1 1 - 874 1 - - 1039 1 - - 1184 (1) -- 1287 2 - - 1468 2 (1)-

217 2 - - 603 3 - - 728 (1) - - 878 (1)(1)(1) 1042 2 - - 1185 3* - - 1288 1 - - 1470 (1) - -

221 1 - - 604 2 - - 729 1 - - 879 (1)(1)(1) 1043 1 (1) - 1186 1 - - 1289 2 - - 1472 (1) - -

238 2 - - 605 3 - - 742 2 (2) - 881 (1)(1)(1) 1045 1 - - 1187 1 - - 1296 (1) - - 1474 -- (4)

501 2 - - 607 1 - - 750 - 2 - 885 (l)(i)(1) 1047 1 - - 1189 2 - - 1298 (1) - - 1475 (1)(1)(1)

502 - 1 1 608 1 - - 752 - 1 - 886 (1)(1)(1) 1051 1 - - 1191 1 - - 1300 (1) - - 1478 (1) (1)(1)

503 4 - - 609 1 - - 756 2 - - 888 (1)(1)(1) 1052 1 - - 1193 (2) - - 1302 (1) - - 1481 (1) (1)()

504 - 2 2 610 1 (1) - 757 2*(2) 890 1 - - 1053 1 - - 1194 1 - - 1310 5 (2) - 1483 ((1) (1)

505 (1)-- 611 1 - - 760 4* - - 892 - - 1 1058 1 -- 1195 (1) - - 1313 1 - - 1487 (1) - -

506 - 1 612 1 - - 763 1 - - 893 (1)(1)(1) 1061 1 (1) - 1197 1 - - 1314 (1) -- 1494 (1) -

507 - - 1 614 3 - - 767 1 - - 900 (1)(1)(1) 1063 1 - - 1202 1 - - 1316 2 - - 1504 2 --

508 3 3 3 617 2 - - 769 1* - 1 903 1* - - 1066 1 - - 1203 2 - - 1319 - - 1 1505 - - 2

509 - - 2 619 1 - - 773 1 - - 905 2* " - 1068 1 - - 1204 1 - - 1327 - - 1 1506 - - 2

510 1 - - 621 2 - - 774 1 - - 906 - - 1 1078 1 - - 1210 2 -- 1328 (1) - - 1507 1 - 1

511 2 - - 622 - - 2 775 1* - - 910 1 - - 1079 2 - - 1212 (1) -- 1329 - - 2 1521 (1) --

512 2 - - 624 - - 2 777 1 - - 911 1* - - 1080 2 - - 1213 4 - - 1330 - - 2 1525 (1) --

513 (1) -- 626 - 3 3 778 1 - 1 915 1 - - 1084 1 - - 1214 (2) - - 1334 (1) - - 1532 2 --

518 1 1 - 630 (1) (1) (1) 779 1 - - 916 1* - - 1085 (1) - - 1216 1 - - 1336 (1) (1)(1) 1533 2 - -

519 1 - - 633 (1) (1) (1) 780 1 - - 918 (1) -- 1087 2 - - 1220 (1) -- 1338 - - 1 1534 (2) - -

520 2 2 2 634 (1)(1)(1) 782 1* - - 919 (1) - - 1092 1 - - 1221 1 - - 1352 1 - - 1536 (1) - -

521 2* - - 640 3* 1 1 783 1 - - 920 (1) (1)(1) 1093 1 - - 1222 1* - - 1360 - 1 1 1546 (1) (1)(1)

523 1 - - 644 2 2 - 784 - - 1 922 1* - - 1097 1 - - 1223 (1) - - 1361 - 2 2 1555 (1) - -

524 - 1 1 645 1* - - 785 - - 1 923 1 - - 1101 3 - - 1224 (1) - 1362 - 1 1 1556 (1) --

527 1 1 1 646 1 - - 783 (1) - - 924 3 - - 1104 1 - - 1225 (1) - 1363 - 2 2 1559 I - -

530 1 1 - 648 - 1 1 790 (1 - - 925 1 - - 1106 1 - - 1226 (1) - 1 136 - - 1585 (1) - -

534 2 2 2 649 - 1 - 796 2* - - 938 1 - - 1107 2 - 1227 (1) - - 136 6

- - 1 A259 -1 -

536 1 1 1 651 (1) - - 798 2 - - 940 (1)- - 1111 1 - - 1228 (1) - - 1367 - - 2 A261 3 3 3

537 1 1 1 652 1 - - 803 - - 1 949 - - 1 1116 3 - - 1229 (1) - - 1368 - - 3 Nippur 5 - -

235

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

and Late phases, but it will be immediately observed thatby these later phases there was little significant settlementin the region left to record. For the Early Islamic periodthe difficulty is that the gradualness of the transition fromSasanian times, as well as continuing debate over how itshould be recognized, makes a distinction between Sasa-nian and Early Islamic occupational areas very uncertain.To list estimated areas of occupation individually would,in these circumstances, tend to endow them with an airof precision that is probably not justified.

Difficulties in distinguishing Early Islamic collectionsfrom Late Sasanian ones are responsible for a substantialnumber of cases in which a site can be given that datingonly provisionally. In these cases table 22 includes a con-siderably more doubtful estimate of its size category inparentheses. Also in parentheses in the Early Islamic col-umn are sites on which the only available information istaken from the files of the Directorate General of An-tiquities (see above, p. 216). Although these were re-corded only under the designation "Islamic," the distri-bution by periods of more adequately recorded sites makesit extremely likely that the great majority of them wereoccupied primarily if not exclusively during the EarlyIslamic period. Information on size is not included in theofficial records, and some of them are doubtless largerthan the lowest category (0.1-4 hectares) that has uni-formly been attributed to them. Taking all these circum-stances into account, sites that are parenthetically listedcannot be ignored on the grounds of somewhat doubtfulattribution. Such sites were not considered in analyzingthe urban hierarchy and extent of settlement for the Sasa-nian period, but there is probably less distortion in com-paring Early Islamic with Sasanian remains if in this casethey are taken fully into account.

Some size estimates are also enclosed in parentheses inthe Middle and Late Islamic columns of the table. Wheresuch a listing is provided in all three columns, it indicatesthat the site was not visited during the survey and isknown only from the directorate's files. Where it is con-fined to the Middle and Late Islamic columns, it suggests,in most cases, a relatively small terminal occupation dur-ing these periods. Sometimes it may also refer to an ap-parent discrepancy between the dating implied by theobserved surface pottery and other, less reliable evidence(e.g., brick size), or an estimate based on the known his-torical importance of a town whose ruins were not visited(e.g., site 1474). Details on these individual cases will befound in the site catalog given in chapter 7.

Turning to the dating criteria themselves, many aspectsof the discussion of the Sasanian period are equally appli-cable to its sequel. There has been an identical lack ofattention to the more mundane aspects of ceramic pro-duction, and especially to the unglazed utilitarian waresthat at all times heavily preponderated. Hence the ceramicsequences on which a surface reconnaissance must depend

in order to assign dates of occupation to individual sitesare still inadequately developed. Moreover, wide dispari-ties exist between the absolute dates assigned to the floruitof different styles by authorities working on different as-pects of the problem, with a general lack of the kinds ofunequivocal, stratigraphically secure evidence that wouldallow a resolution of existing differences.

For this discussion, therefore, the Early Islamic periodis a somewhat indefinitely bounded span of time. Ambigu-ities as to its inception have been adumbrated in the pre-ceding section; comparable or even greater ambiguitiesattend its transition to the succeeding period. It is de-fined primarily by a set of characteristics, many of themof undisputed Early Islamic age if a precise specificationof the chronological meaning of this term is avoided, thathave been fairly uniformly observed to constitute thedominant surface component on several hundred sites insouth central Iraq. These ceramic stylistic features ap-parently signify the terminal occupation in such cases,marking a time of very widespread, and to a large degreefinal, abandonment of the region.

It is likely, though by no means certain, that this aban-donment was under way by or soon after the mid-ninthcentury A.D. One of the numerous large sites whoseterminal surface materials are of this generally EarlyIslamic character is the much more ancient city of Nippur,the name preserved in Arabic as Niffar and known as theruin of a famous place as late as the time of Yaqut in thethirteenth century. Islamic coins have been found inquantities on the surface of the mound by excavators whohave worked there for many seasons, and all that canbe identified antedate the end of the eighth century(McG. Gibson, pers. comm.). Similarly suggestive of amassive decline of the region at about this time is the ab-sence of reference to it in the works of the great geogra-phers of the later ninth century, which would be inex-plicable if Nippur and its contemporaries had continuedto flourish. On the other hand, bishops of Nippur contin-ued to be appointed until the late tenth century, at whichtime the title was joined with that of the later city of Nil50 kilometers northwest. If the coins and geographers areto be believed, therefore, one must suppose that becauseof its antiquity and former importance the title may havebeen retained for a time even after the seat of dominicalauthority was transferred. Such a possibility is at leastnot inconsistent with the employment of "archeologis-mes" in connection with bishoprics of the time (Fiey1968, pp. 250-51).

Supporting a ninth century dating is the list of weirsmentioned on p. 216. Compiled at the end of the ninthcentury and including some entries that had already beenabandoned at that time, it contains no listing for the Nahral-Nars, nor any other that could conceivably stand forthe weir that clearly existed at site 1213 (cf. above, p. 213).The surface ceramics on the lengthy ruins adjoining both

236

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

sides of the main canal above the weir are essentially thosenoted as "Early Islamic" throughout the region, indis-tinguishable from those at Nippur. It seems very likelythat the weir, the canal, and also the adjoining site musthave been abandoned for a considerable period beforethe list of weirs was compiled, in order not to be includedin it.

As an alternative, however, it is not impossible that thereligious titulary should be given greater weight and thetime span of the Early Islamic surface materials continuedwell into the tenth, and conceivably even eleventh, cen-tury. A minor but unmistakable component of those ma-terials is the use of underglaze sgraffiato drawings, andsome Islamic archaeologists currently maintain that these"were not applied to slip ware in the Middle East muchbefore the end of the 10th century" (Schnyder 1974, p. 90;cf. Naumann and Huff 1975, p. 190).

In one sense this difference is only marginally relevantfor the purpose immediately at hand. A set of widely re-curring dating criteria is not less valid, after all, if it spansfour centuries than if it is largely limited to a two-centuryperiod. But, in another sense, what is at stake is a frame-work not only for comparative dating but for understand-ing of profound change in settlement and demography, aframework that applies to succeeding periods as well asto this one. Hence some discussion of ceramic dating prob-lems is necessary here, although it must be conjoined withthe broader, interpretive discussion in the foregoingchapter.

The Abu Sarifa excavations were undertaken to helpclarify these problems, and the sequence given here isbroadly consistent with the sequence of stratigraphic su-perimposition at that site. But it cannot be maintainedthat those limited soundings provide decisive support onany of the crucial issues. Discussions of dating still in-volve the glazed pottery almost exclusively, and the quan-tities of glazed pottery found were altogether inadequateto establish the relative position and internal dynamics ofchange of any particular type. Adding considerably to theuncertainty, moreover, was the extensive disturbance ofthe site in Islamic times, caused by deep pits in which trashsubsequently accumulated. Where individual fragmentsof a single glazed vessel of a key type are found distributedover at least 1.25 meters of vertical depth, some hesitancymust be expressed about assigning its provenience to anylevel. Similarly, the penetration of a handful of Early Is-lamic glazed ware types into what seems conclusively alevel mainly occupied during the Sasanian period suggeststhat the Abu Sarifa evidence adds little to our knowledgeof the fine-grained temporal distribution of those types(Adams 1970, p. 91, table 2, fig. 14).

In fact, however, little is known of temporal distribu-tions on any basis. A series of entirely new glazed stylesspread over much of central and southern Iraq, at someill-defined period or series of periods after the new Arab

conquest. The absence of these styles in level III at Susa(Rosen-Ayalon 1974, pp. 207, 233, 251, 261, 276) and ata site like Tulul al-Ukhayder, apparently abandoned onlyabout the middle of the eighth century (Finster andSchmidt 1976, p. 113), argues that most of them were in-troduced during the second century following the con-quest rather than earlier. If so, the uncertainties aboutAbu Sarifa multiply. Level IV there, considered EarlyIslamic, contains more than sixty sherds of these newtypes (Adams 1970, table 2). Is this merely a further indi-cation of the distortions introduced by extensive pitting,or should the date of primary construction in the levelbe raised from the mid-seventh to the mid-eighth century?

Prominent among the new styles, and of crucial im-portance for dating, are wares with splashed underglazedecoration in green, amber, and manganese-brown or-purple. The putative source of the style, certainly verysimilar in appearance and differing mainly in the absenceof manganese as a tint, lay in T'ang Chinese sans-ts'aivessels of a type not made after about the middle of theeighth century (Watson 1970, p. 37). Following Fehervari,"it is tempting to argue that this type of pottery was in-troduced . . . at or soon after the middle of the eighthcentury," perhaps by Chinese prisoners (1973, p. 36). Butactual imports are not known. Moreover, the Chineseprototypes are known only as imperial funerary ceramics,"not the kind of ceramic ware which would have appearedin the bazaar at any time, least of all, be exported to theother end of the Asian continent" (Crowe 1977, p. 270).

A related type of san-ts'ai ware reappeared in northChina during the Liao dynasty (A.D. 918-1126), and a fewunquestionable examples of this have been identified atseveral widely separated Islamic sites that hint at itsimportation by sea. This is entirely too late to account forthe origin of all Middle Eastern splash-glazed wares, but,as Basil Gray observes, we are not thereby compelled toopt for the mid-eighth century alternative. Questioningthe present completeness of the Chinese evidence, he hassuggested that continuity in the production of the warebetween T'ang and Liao will yet be found. If so, theprocess of borrowing could have extended over a long in-terval. The celebrated gift of Chinese ceramics from thegovernor of Khurasan to Harun al-Rashid in 804, includ-ing twenty pieces of porcelain and two thousand of othervarieties, may well have been an incident of some im-portance in this process of commercial or ceremonialtransfer followed by local imitation (Gray 1977, pp. 232-33).

In any case, splash-decorated glazed wares, includingthose additionally decorated with underglaze incisions inscript and floral motifs, were present at Hira in contextsascribed to the later eighth century on numismatic grounds(Rice 1934, pp. 54, 70). They not only begin but proba-bly reach their highest level of popularity in level II atSusa, coinciding with the later eighth and ninth centuries

237

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

(Rosen-Ayalon 1974, p. 261). Because of their promi-nence among the finds at Samarra, it has similarly beentaken largely for granted that splash glazes and sgraffiatowere in full use, if not necessarily at the height of theirpopularity, by the mid-ninth century A.D. To be sure,the character of publication of the finds at Samarra doesnot permit them to be directly associated with the rela-tively brief period of its use as the 'Abbasid capital (A.D.835-93). But all the published pottery from the originalprogram of excavations is at least reported to derivefrom the palaces and public buildings that were the mainobjectives of the excavations and that certainly were occu-pied during the floruit of the city (Sarre 1925, p. v). Theunstable, centralized, oligarchic character of the govern-ment of the time makes it seem rather unlikely that thesewould have received substantial, continuing use after theseat of power was returned to Baghdad. On the otherhand, nothing associates the disappearance of the stylewith the abandonment of Samarra as the imperial capital;the town continues today, after all, as a local administra-tive, market, and pilgrimage center. Moreover, a jar in avariant of the splash-glaze style for which there are a fewparallels at Samarra, securely dated by the hoard of coinsit contained, was still in use at Susa in the mid-tenth cen-tury (Fehervari 1973, p. 36).

All these disparate considerations cannot be fully rec-onciled. It must be acknowledged that date provision-ally assigned to the Early Islamic sites recorded in thissurvey may need modification when Islamic archaeologyfinally receives the concerted scientific attention that isits due. But it does seem reasonable to conclude that boththe "Early" and the "Classic" Samarran sgraffiato styleswere at least partly contemporary with the floruit of Sa-marra as a capital. Where the former of these styles occurson smaller sites in south-central Iraq, therefore, it is re-garded in this study as signifying a late eighth or ninthcentury occupation. The latter also seems to have comeinto use not long after the mid-ninth century, but it isassumed here to have continued through the tenth cen-tury and probably for some time into the eleventh.

Thus it is very likely that the presence of the "Classic"sgraffiato style on a site is an indication that it was occu-pied until a significantly later date than other sites with agenerally similar, Early Islamic pottery conspectus. Forthis reason sites where the style was observed have beenindicated in table 22 with an asterisk.

Emphasizing once more the essential fuzziness of ourknowledge of the ceramic indicators, those that have beenemployed in this study for the Early Islamic period are asfollows:

Large, plain buff ware jars continue in a Sasanian oreven earlier tradition, although they become proportion-ately less numerous. Hole-mouthed, "torpedo-fuse base"forms are among them, but somewhat more common areglobular forms with three handles attached to prominent,

high vertical collars that are decorated with multiplehorizontal grooves.

Also continuing an earlier tradition are large plain buffware bowls with flat bottoms and flaring sides. The earlieremphasis on decorated flange rims and club-headed rimsdisappears, but a few examples with moderately bolsteredrims may preserve a remnant of that style (Adams 1970,fig. 6j-o).

Newly introduced, perhaps only in the mid-eighthcentury, coinciding with the onset of level II at Susa(Rosen-Ayalon 1974, p. 50, appendix II), are relativelysmall jars of a notably thin, well-levigated, seeminglytemperless, light buff fabric. Most are given individualityby having been wheel-modeled to produce multiple hori-zontal contour breaks, grooving, or both. Incised decora-tion is often applied to these vessels, in the form of cross-hatching, meanders crosscut by slashes, and leaf andtwining plant motifs that are often arranged in verticalpanels (Adams 1970, fig. 6a-i, p-w, ac-af, ai-as, fig.10a-x; Rice 1934, fig. 20). Again to judge from Susa, how-ever, most of these types of surface decoration were intro-duced somewhat earlier (level III) on jars of less stylizedform (Rosen-Ayalon 1974, figs. 39-40, 53-54).

Knobs, or stamped or otherwise embellished "turbans,"added to the upper surfaces of large jar handles (Adams1970, figs. 7c, 8).

Impressed stamps, generally "bull's-eye" or other geo-metric patterns within a circular field, commonly appliedin one or more horizontal rows to bodies of large plainware jars (Adams 1970, fig. 9; Finster and S6hmidt 1976,Abb. 49; Rice 1934, fig. 22; Rosen-Ayalon 1974, figs.138-39, 141-43).

Handmade, crudely shaped jars, stoppers, and espe-cially flat-bottomed basins with large volcanic grits andother irregularly shaped tempering materials, includingsmall fragments of glazed sherds. Often pinkish in color,this locally fabricated and only semiportable ware is oftenfire-blackened and perhaps burnished on the interior(Adams 1970, p. 95, fig. 5a-f).

Dark-faced orange pottery, a hard, thin, somewhatbrittle ware used in the manufacture of presumably im-ported profiled bowls and jars that were widely diffusedin small quantities. Fluted strap handles are common andoccasional sherds bear impressed rouletted chevron dec-orations (Adams 1970, p. 96, fig. Si; Fisnter and Schmidt1976, p. 111, Abb. 45, Taf. 50a, b, d, f).

Black steatite bowls or basins also were imported. Thecharacteristic shape was flat-bottomed and vertical-sided,often equipped with horizontal lug handles (Adams 1970,p. 96).

Turning to glazed wares, by all odds the most commonis a deep blue or blue green allover glaze on a soft yellow-ish fabric. It is already referred to above at the end of theSasanian section, although the period of its greatest popu-larity was clearly later-at Susa, in fact, only after the

238

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

mid-eighth century (Rosen-Ayalon 1974, pp. 162-64). Themost common form is a large, ring-based, two-handled(strap or lug) jar that is sometimes decorated with under-glaze meanders and other forms of incised decoration;appliqued medallions and low indented ridges were pop-ular on vessels of this type in the Diyala region but arevery rare here (Rice 1934, p. 70, fig. 23; Adams 1970, pp.106-7, fig. 7d; Finster and Schmidt 1976, pp. 111-12,Abb. 46).

Small, white-glazed bowls with a low ring base, occa-sionally with widely spaced vertical fluting, are also rela-tively common. The surface finish is usually described asa tin opacified lead glaze, but recent research has castdoubt on at least the uniform applicability of this descrip-tion. Complicating the problems of chronological assign-ment, apparently there was instead a long, fairly complextransition, in which the Early Islamic products of this typestill were mainly produced in the Sasanian, underfiredalkaline glazing tradition (Crowe 1977, p. 266). A softyellowish fabric is sometimes but not always used for thislong-lived category. Some individual examples appear tobe local imitations of Chinese celadon, but the type itselfmay well antedate the introduction of celadon and hencemay be misleadingly described if referred to as merely"imitation" celadon (Sarre 1925, Taf. 23-25; Rice 1934,p. 69; Adams 1970, p. 110, fig. 11aa).

Next in importance, though pronouncedly rare, arewares with splash decoration beneath a transparent glaze,occasionally miniature vessels but predominantly large,deep ring-based plates in putative imitation of T'ang pat-terns. Two styles were recognized in the fieldwork ac-companying this study, an "Early" one believed to slightlyantedate the Samarran period (although continuingthrough it) and a "Classic" one also much in evidence atSamarra but probably continuing for more than a centuryafterward. A partial succession of this kind, combinedwith a considerable degree of overlap for the two styles, isat any rate strongly suggested by Reitlinger's work nearKish. He found that something very like the "Early" styleas described here had essentially ceased to be produced,while the "Classic" style remained in use, at the time of areoccupation dated to the late tenth or early eleventh cen-tury (1935, p. 201). Sgraffiato underglaze decorations in-cised through the slip may occur on both but are morecommon on the latter.

The "Early" style occurs on fabric that varies fromgrayish buff, fairly low-fired, to finer examples made of ahard, reddish, sandy clay. Long radiating splashes formdense, allover patterns in green, manganese (violet),brown, orange, or a combination, over an orange yellowor white slip. Incised patterns, when they occur, are exe-cuted in a uniformly fine, often barely visible, line andinclude horizontal bands, loose scrolls, and thin, uncon-nected, nonrepresentational motifs. It is the splashes ofcolor that clearly dominate the ensemble (Lane 1947, pl.

6B; Sarre 1925, pp. 76-77, Taf. 32:4; Adams 1970, pp.108-9).

The "Classic" style is executed on vessels of the sameshape, although hard-fired examples are perhaps lesscommon and some cases of inferior workmanship areencountered. The long splashes of color are reduced tosparse mottling, mainly isolated rows or small groups ofbrown dots, or small daubs and rare splashes of green,on a white slip. Incised decorations become denser, morediverse, more suggestive of written characters and otherrepresentational motifs, more painstakingly executed, andincreasingly dependent for their effect upon wide varia-tions in the width of the line or pattern cut through theslip to expose the underlying color of the fabric. It is theseincised decorations rather than the daubs of color thatare the primary focus of attention (Saare 1925, Abb. 161,165, 168-70; Adams 1965, p. 109).

Related to the "Early" style of splash-decorated wareare vessels of the same shape and fabric that receive all-over green underglaze coloring. Examples occur with un-derglaze incisions in the same patterns (Adams 1970, fig.11v).

Rarely occurring, although fairly common in otherparts of the Mesopotamian plain, are flaring, slightlyrounded bowls with radiating patterns of long blue-glazeand/or manganese violet-glaze splashes, over a white-glazed field on the interior surface (Adams 1970, p. 110).

Finally, there are small, thin, uniformly well madebowls of "luster" ware affinity, with mainly olive-coloredgeometric underglaze designs painted on white or graybackgrounds (Adams 1970, pp. 110-11; cf. Rice 1934, p.70).

Early Islamic bricks in many cases seem to resemblethose of the Sasanian period. They are smaller (typically29-30 centimeters square) but are almost equally wellfired and tend to the same deep, red brown color. A furtherreduction in size continues, perhaps still within the EarlyIslamic period. It is possible, for example, that the Sasa-nian-Early Islamic tradition of large, well-made brickswas intended largely or exclusively for public buildings. Asthe use of bricks in private dwellings rapidly spread, adifferent, much less well made type of brick might havecome into existence that for a time supplemented but didnot replace the earlier one. At present this is admittedlyspeculative. All that is certain is that within four or fivecenturies the only bricks still being made were small(21-23 centimeters square), poorly shaped, and of a soft,low-fired, yellowish fabric. But the use of brick size aloneas an indication of dating within the Islamic period,which at one time seemed a promising lead (Adams 1965,p. 183), seems increasingly likely to obscure a complex,overlapping sequence. Note may be taken of the occur-rence of 30-centimeter brick in the Islamic tower at Zibli-yat (site 700), for which a radiocarbon determination nowsuggests a dating as late as the Samarran period.

239

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

The general comments that have preceded the Sasanianand Early Islamic periods apply with equal force to theMiddle Islamic period. Most specialists' attention remainsconcentrated on the encompassing, paradigmatic stylisticexpression, all too often deprived of archaeological con-text by the illicit manner of its acquisition. Little practicalguidance on dating continues to be available, therefore,that is relevant to the masses of mundane pottery encoun-tered during a surface reconnaissance. If the problem canbe more briefly disposed of than in the former cases, it islargely because the size and number of settlements dwin-dled so precipitously that the problem of dating losesmuch of its significance.

The Middle Islamic period, for this study, is assumedto begin at about the end of the tenth century, or evensomewhat later, and to continue until the profoundlychanged conditions following the Mongol conquest in themid-thirteenth century. Within this interval there is afurther, easily recognized ceramic horizon, the advent ofblue black painted underglaze designs, as well as a groupof floral designs in blue or black. These styles appear to-gether at Wasit at about the middle of the twelfth cen-tury (Safar 1945, pp. 41-42). Their very rare occurrencesin this area, individually noted in the site catalog, reflecta reduction of settled life to apparently its lowest levelsince remote prehistoric times. Before that time and cer-tainly after the Samarran period, however, a complex oftraits occurred in the uppermost levels at Abu Sarifa (in-cluding a surface level that has apparently been wind-deflated) and elsewhere whose individual componentsseem for the most part to have persisted throughout theentire Middle Islamic span of perhaps three centuries.More comprehensively, the useful surface characteristicsof the period may be described as follows:

Blue glaze continues as the dominant color, but fabric,form, and the glaze itself are radically altered. The softyellowish paste is no longer employed, and in this respectthe glazed and unglazed vessels now appear to form acontinuum. Also discontinued is the characteristic two-handled, ring-based jar with allover blue glaze that goesback to Early Islamic or even Sasanian times. Smaller jugsand especially deep ring-based bowls are the newly pre-vailing forms. Finally, surfaces have a coarser, sometimesmottled texture, and the glaze frequently has crackedaway from the fabric. Bowl interiors are completelycoated, and the glaze coating is carelessly applied to theupper exteriors as well.

Green and grayish white glazes also continue, butwithin the same framework of changes in composition andvessel form as those just identified for blue glaze. Consid-erably less common than blue, these colors may be re-served largely or even exclusively for bowls.

The sgraffiato tradition likewise continues, while ap-parently declining progressively in quality. Splashes oreven isolated small flecks of color are frequently not in

evidence on fairly large sherds, and underglaze slip is ir-regularly applied and sometimes missing. Incised patternsare increasingly broad-line, crude in technique, and slap-dash in effect. The transparent overglaze also is lesssmoothly applied, and on many surface sherds it hascracked away almost completely. One is left with the im-pression that a highly stylized, near-luxury product of aformer period had become only a distant prototype for aproduct of widespread, localized manufacture.

Newly introduced for bowls is glaze of a very deep(manganese-) violet, sometimes appearing almost black.Application to the exterior is also careless and limited tothe upper part of the vessel, but interior surfaces are nota-bly smoother and more even in color.

Finally, there are the twelfth century (and continuing)traditions of black and blue underglaze painting. Gen-erally more common in rural central Iraq are black linearor reserved designs under dark green or blue lead glaze,but there are also floral patterns in blue or black beneatha white or pinkish glaze (Safar 1945, pp. 41-42). Both aretypically applied to relatively well-made, thin-walled, flar-ing-sided, ring-based bowls.

Little systematic attention was devoted to ceramics ofthe Ilkhanid and more recent periods, here designated asLate Islamic. Sedentary life appears to have been ineclipse within the area of intensive survey until towardthe end of this seven-century span, and mounds whoseelevation reflects a continuing buildup of debris from oc-cupation during the greater part of it are rare. Equallyrare even on those mounds, moreover, are glazed ceramicsof documentable antiquity, traded in from urban centersaround the peripheries of a region mostly given over totribal dissidence. Surface reconnaissance is largely tied,therefore, to highly localized, utilitarian and autarkic tra-ditions of pottery manufacture whose thin remains on adiffuse scatter of sites would be very difficult to place inchronological order.

Some use of blue glaze, primarily for large-bowl in-teriors, continued until the threshold of modern times.The glaze on more recent examples tends to be thin andrough to the touch and to flake away exposing an under-lying fabric that is pinkish and granular. Green glaze anda grayish white lead glaze that usually has a curdled orpitted appearance also occur very sporadically. Hand-made, low-fired, unglazed basins and jugs with large,irregular grit inclusions probably represent another con-tinuing tradition. Fragments of this type are widely re-ferred to by local residents today as "bedouin" pottery,and unquestionably it would be an appropriate article forfabrication in a bedouin encampment out of touch withany market center. The same kind of ware, particularlytaking the form of deep bowls, is frequently found withfugitive traces of painted geometric designs that have beendesignated "pseudo-prehistoric" (Safar 1945, p. 38; Ad-ams and Nissen 1972, pp. 67-68). Not a few sites on

240

oi.uchicago.edu

Culmination and Collapse of an Agrarian Base and Urban Superstructure

which this pottery is found also have spent cartridges, uni-form buttoms, and the fragments of china teacups thatsignal the reopening of the region to at least indirect and

intermittent contact with distant centers by the later nine-teenth century.

241

oi.uchicago.edu

Conclusion

It would be redundant if not misleading merely to sum-marize at a more abstract level an account that alreadyincludes a fairly large measure of synthetic reconstruction.As I noted earlier, no single, distinctively Mesopotamianparadigm of settlement has emerged whose physiognomybecomes recognizable only in the heartland of the region.To distill a single harmonious and consistent pattern outof the many unstable and discordant elements that havebeen described would not only do violence to many ofthem individually but would have the more pervasiveeffect of suppressing one of the main sources of dynamismaccounting for the region's historic creativity and im-portance. So the foregoing chapters will be left without abriefer recapitulation, contributing something, it is hoped,to an enhanced understanding of the Mesopotamian rec-ord of demographic, agrarian, and urban development.

Yet at the same time this study cannot merely acknowl-edge the virtual end of its supply of primary survey datain the Middle Ages and fade away into the silence of theGreat Swamp along with the dwindling number of thoseremaining who were inhabitants of the central Euphratesfloodplain. Research domains worthy of exploration maytwist and unfold or come together and take new forms,but since they are not self-enclosed they never simplyterminate. In entering them at all, therefore, we assumesome responsibility to suggest their continuing and widerresonances. Other themes of comparable importancecould perhaps be cited, but there are three larger issues inparticular on which this study may conclude by at leastdrawing together diverse earlier strands into a more co-herent position.1

The substantive foundation of this study, though it hasranged widely (and perhaps incautiously) into other dis-ciplines, has been the findings of archaeological surveysdevoted to settlement and irrigation patterns. Two prin-cipal categories of data have been assembled: an array of

ancient settlements that can be described in terms of sizeand periods of occupation, and the traces of at least someof the ancient watercourses that made those settlementshabitable. What connects the two in the most narrowlyempirical sense is that they accompany one another ac-cording to an obvious and inescapable least-effort prin-ciple. In an otherwise relatively featureless and arid land-scape, human communities either gravitate to their watersupply (and incidentally, best available means of bulktransport under preindustrial technological conditions)or redirect that supply to their preexisting centers if it isdiverted elsewhere by natural or human agencies. Theirrigation patterns greatly enrich an understanding of thesettlement patterns and in turn cannot be understood ex-cept insofar as periods of use can be assigned to com-ponents of the watercourse network on the basis of thedating of adjoining archaeological sites.

The first of the broader themes arises from considera-tion of a common characteristic of these two kinds of data.It is in the nature of both that marshaling them contributesto a reorientation-happily, already under way-of thepremises and priorities with which research on ancientMesopotamia is carried forward. Indexes like settlementsize admittedly may be only roughly related to many moreinteresting political and economic institutions. The timeperiods it has been possible to distinguish in the archaeo-logical surface collections are exasperatingly long fromthe viewpoint of anyone concerned with finer-grainedstudies of social process. But these are entities that nowapproach a status of having been assembled compre-hensively within a fairly inclusive region, and they canbe quantitatively expressed. The systematic counting isimportant, not diversionary. As Georges Lefebvre waswont to say, "Pour faire de l'Histoire, it faut savoircompter" (Cobb 1971, p. 1527). Methods can and willbe refined, but the introduction of measured aggregates

242

oi.uchicago.edu

Conclusion

not only permits "testing" and consequent refutation orimprovement, but also encourages the generating of newhypotheses. Crude and inconclusive as the first steps maybe, they must be taken if we are ever to move beyond the-ories that remain permanently tangential to one anotherbecause they rest on each authority's subjectively chosenmetaphors and examples. If there is ever to be a disciplinethat brings archaeology and Oriental history togetheraround a common subject matter, it will have to be builtnot on the best-preserved, most vivid, or most evocative,but on what has painstakingly been established to be themost commonplace.

By an extension of the same principle, the use of datagathered in a more or less uniform way tends to unify theentire span of Mesopotamian history. In particular, ithelps to break down chronological divisions that, throughtheir association with fairly coherent bodies of linguisticand textual source material, have become barriers to com-munication between modern specialists. One need not bea fervent believer either in complete cultural continuityor in the lack of it to see the advantages of reducing thesebarriers. Only when genuine comparisons across long in-tervals of time are possible, transcending the specializedbiases and limitations of particular archaeological as-semblages and genres of texts, can we begin to under-stand why certain patterns developed cumulatively whileothers went in a different direction and flourished morebriefly.

The other two broader issues require a more extendeddiscussion. Both begin once more with the dependence ofthis study on the ancient canal- and settlement-patterndata, but each proceeds from there in a different direc-tion.

ANCIENT MESOPOTAMIA AS ANIRRIGATION SOCIETY

The existence side by side of these two distinct bodiesof data focusesattention on the relationship between aprevailing mode of subsistence-in this case one depen-dent on canal irrigation-and the social institutions thateither were superimposed on it or were generated in. re-sponse to its internal stresses and requirements. The an-cient settlements that were located during the surveys in-cluded an unusually large, and uniquely early, proportionof cities. Not to depreciate later achievements, the re-peated reminder is useful that we are dealing here with theremains of the world's earliest literate civilization. It is notpossible within the scope of an already extended study tosurvey even the most important and enduring features ofthe institutional structure that characterized that civiliza-tion. Nevertheless, we should not wholly ignore the ques-tion whether, and in what manner, some of the social con-comitants of irrigation exercised a wider influence on thecourse of societal development. What was the relation-

ship, if any, between a precociously early and long-per-sisting growth of urban life and major reliance on irriga-tion agriculture?

Probably the most fundamental advantage of irrigation,here as elsewhere, was the relatively high agriculturalproductivity it permitted. It was unquestionably of greatimportance that yields were high, in relation to land aswell as labor inputs. To the degree that environmentalchallenges were successfully met, large surpluses thuscould be mobilized above the needs of the primary pro-ducers. Additional labor services could beextracted fromthe agricultural population during at least a •atof theyear, supplemented by those that could also be furnishedby pastoralists living in a generally symbiotic relationship"with the cultivators. As one consequence, relatively highpopulation density was attainable even at a fairly simple,only slowly improving level of technique. At least equallyimportant, considerable proportions of the populationcould be maintained in specialized pursuits rather than inagriculture itself. All this is obvious and long understood,and needs no more than brief acknowledgment. Perhapswe can describe these conditions as tending to be neces-sary in some degree for the growth of an urban civiliza-tion. anywhere, but they do not significantly advance ourunderstanding of the major structural features character-izing this one in particular.

Still less useful as an explanatory framework for at leastMesopotamia are the sweepingly systematic views of K. A.Wittfogel (1957). Insofar as conditions denoted by hisconstruct of a rigidly authoritarian, centralized, bureau-cratic "Oriental despotism" ever were even remotely andbriefly approached in this historical and geographic set-ting, they were never a norm and instead might be de-scribed as a kind of asymptote or limiting case. UnderlyingWittfogel's view is an insistence that the management andcoordination of an irrigation system of any size requiresa despotic political regime that completely suppresses thescopotaction and autonomy of all other societal group-ings within its compass. It should have become apparentin the earlier chapters of this study that in Mesopotamiathat kind of unrestricted power was conspicuous by itsgeneral absence, and that even in the conduct of irriga-tion the hand of the state was only selectively and episodi-cally applied.

Another difficulty is that Wittfogel has chosen to focuson a single, somewhat questionable or at least contingentvariable in a complex, interdependent set. A central find-ing of chapter 1 was that the most decisive factor was theirregularity and unpredictability of the water supply. Con-flict and competition over it was unavoidable, since notechnical means existed by which a position of advantagewith respect to access to water could be permanently as-sured. Water scarcity was thus likely to be at the root ofmany efforts at political consolidation in the hands of newdynasties, regional coalitions, and other groupings, and'

243

oi.uchicago.edu

Conclusion

it was just as likely to produce the coalitions that ulti-mately were their undoing.

As we have seen, the hydrological characteristics of thetwo rivers were such that only the Euphrates was a sig-nificant source of irrigation water for most of Mesopo-tamian history. Both its modest aggregate flow and itswide month-to-month as well as year-to-year variabilitymeant that substantially less than half, perhaps even lessthan a third, of the alluvial plain between the two riverscould have been reliably irrigated until the Tigris wasbrought extensively into service in Hellenistic or earlymedieval times. In earlier antiquity the problems of deal-ing with the Euphrates were exacerbated by its naturaldivision into an inherently unstable network of intercon-nected branches.

In the largest sense, Mesopotamian cities can be viewedt

as an adaptation to this perennial problem of periodic,unpredictable shortages. They provided concentrationpoints for the storage of surpluses, necessarily soon walledto assure their defensibility. The initial distribution ofsmaller communities around them suggests primarily lo-calized exploitation of land, with much of the producingpopulation being persuaded or compelled to take up resi-dence within individual walled centers rather than re-maining in villages closer to their fields. Tending to con-tradict a narrowly determinist view of urban genesis asmerely the formation of walled storage depots, the draw-ing together of significantly larger settlements than hadexisted previously not only created an essentially new basisfor cultural and organizational growth but could hardlyhave been brought about without the development ofpowerful new means for unifying what originally weresocially and culturally heterogeneous groups.

Also denying the existence of a smoothly deterministicsequence are temporal and regional variations in the proc-ess that were described in chapters 3 and 4. The historicpattern of competitively coexisting city-states is fore-shadowed in the Nippur region already in the Early-Middle Uruk period, for example, but farther to the southUruk seems to have remained for a long time afterward adifferent kind of center with a larger, heavily populatedhinterland. As is suggested by the apparent prominence oftemple architecture there (one might wish that its exca-vators had given more attention to contemporary do-mestic quarters), its leadership was perhaps legitimatedin relatively less pragmatic or political and more theo-cratic terms. Similarly unexplained by any simple, de-terministic reconstruction is the prolonged period dur-ing which sedentary occupation seems to have been ex-tremely limited, followed within the space of no more thana few centuries by the appearance of differentiated urbanhierarchies. Immigration may well have occurred, but italso contributes little to an explanation, since we canfind no source for a major stream of colonists who couldhave brought urban institutions with them. A fortified

storage depot model of urban origins simply does notwork very well unless there has been a long-continuingbuildup of population to generate competition over scarceresources. And, at least from what is now known, it doesnot appear that occupation in southern Mesopotamia be-came dense enough before Early Dynastic times to makecollisions of interest between major population cluster-ings unavoidable.

Still a further difficulty is that the proportion of urbansettlement never attained a constant level suggestive ofsome ecological determinant. It reached a remarkably highlevel for a time in the mid-third millennium, when wehave seen in chapter 4 that almost four-fifths of the popu-lation of the central floodplain was apparently crowdedinto large urban nuclei each covering more than 40 hec-tares. Thereafter it declined steadily, however, reachinga low of less than one-sixth at around the beginning ofthe first millennium B.c. before once more beginning toclimb. Clearly, there were important considerations atwork in determining the extent of urbanization other thanmerely the concentration and defense of surpluses as anadaptation to the uncertainty of the water supply. But itseems equally clear that the long-continuing existence ofcity-states as elementary building blocks of Mesopotamianpolitical and economic life, as well as the abrupt changesin political fortune that individual cities repeatedly under-went, cannot be understood without reference to the in-stability of the scarcest and most critical resource in theirenvironment.

Broadly speaking, contention over water divides up-ystream and downstream competitors. The cumulative ad-

vantage lies with those farther upstream, but its effectua-tion came piecemeal and was repeatedly altered anddeferred by the unstable Euphrates regime of multiplenatural channels. Moreover, the initial possibilities for de-veloping techniques of irrigation appropriate to dispersed,relatively small and weakly organized communities layeither along the Zagros flanks (and hence largely outsidethe framework of this discussion) or in the lowermostportions of the alluvium. The advantages of the latterwere that reduced stream volume and velocities and rela-

y tively low levee topography tended to simplify canalconstruction. Hence it is no surprise that the early citieswere largely concentrated in the south, or that later therewas an irregular but cumulative northward progressionin the centers of population and power concluding onlywith the founding of Baghdad as the 'Abbasid capital.

Major upstream-downstream rivalries for a long timetook the form of politicomilitary contention between city-states, and considerable evidence that maneuvers overaccess to water played a part in this can be found inchapter 4. But localized tendencies of the same kind musthave been no less important, even though the limited pur-poses of the existing textual sources generally were notsuch as to deal with them explicitly. On any canal long

244

oi.uchicago.edu

Conclusion

enough to accommodate different groups of cultivators,users nearer its inlet are prone to protect their own fieldsagainst the uncertainty of future water supplies by irrigat-ing excessively rather than leaving an equitable share forthose below them. Inadequate upstream maintenance alsohas little effect in that area initially, but it quickly reducesthe proportion of flow available to those dependent onthe tails of the canal for their cultivation. Furthermore,opportunities for intensive summer cultivation, includingthe all-important growing of dates, accrued to those withthe best-assured access to water-that is, with landseither along the major channels or near the offtakes fromthem. Herein lay an inconspicuous, slowly developing,but very powerful source of internal social and economicstratification. It may well be that differentiation alongthis and similar lines had as much to do with the appear-ance of complex, hierarchical state and urban institu-tions as the much more frequently and explicitly chron-icled rivalries that took an interurban, overtly militaryform.

It has been convenient heretofore to express the chal-lenge presented by this ecological setting in terms of thesingle threat of water shortage. That should be under-stood as vastly oversimplified, although it does seemlikely that the availability of water was usually the mostdifficult and immediate problem. Locusts and others pests,crop diseases, and late spring flooding before the cropscould be harvested are merely illustrative of other en-vironmental dangers. Nor were human predators, in-dividually or in organized bands or armies, any less de-structive and difficult to deal with. Soil salinization isperhaps representative of the intersecting human andnatural vectors that must have been most common. Al-though the rise of saline groundwater was a naturalphenomenon, it was hastened by overirrigation and wasselectively most damaging to those cultivating depressionsnear the tails of a canal system. What these agencies allhave in common is that they could be neither predictednor substantially controlled with existing technical means.Hence in response to all of them the husbanding of sur-pluses under hierarchically organized institutions witheven a modest, strictly relative assurance of continuityrepresented the most broadly advantageous course ofaction that was available. This further implies, of course,that irrigation should be seen not as an independentlydecisive force engendering the development of cities, butrather as one of the most important of a group of suchforces that tended to pose social challenges of a particu-lar kind.

Three successive, broadly contrastive patterns or con-figurations are evident in the stream and canal networksthat have been described and mapped for differentperiods. The first, applying to the Uruk and Jemdet Nasryperiods, appears to have been an essentially natural sys-tem of bifurcating and rejoining river branches. Only

quite limited portions of it can be recognized directly,except for a fortunately preserved, extremely sinuousreach north of Nippur, but linear alignments of sitessuggestive of lengthy, wholly artificial canalization areat best very rare. Irrigation must have been accomplishedby damming the natural distributaries and then conduct-ing water to the fields by relatively small-scale canal con-struction and localized flooding. The depth of the belt ofcultivation was probably seldom more than a few kilo-meters, so that almost all settlements are likely to haveremained as close to the parent watercourses as factorslike localized dangers of flooding would permit. Whilepotentially of great interest, nothing further can be saidat present of this most ancient pattern.

The second configuration applies to the third, second,and early first millennia B.C. For most of this span popu-lation levels were considerably higher than earlier, whilethe number of channels had been sharply reduced (pos-sibly with increasing climatic aridity as a factor). Thuseach of them was accompanied by a deeper belt of culti-vation, and around the major cities there must have beenstill further widening. At times, as in the mid-third mil-lennium, the number of settlements was relatively small,and their average size was correspondingly large. Hencealmost all sites at that time could be accommodated alongthe parent watercourses. At other times, as in the latethird and early second millennia, there appears to havebeen a breaking away from this intense stress on urbanismand a dispersal of smaller settlements along branch canalsleading laterally away from the major channels. But thedepth of the belt of cultivation, insofar as it can be de-duced from the location of often quite minor sites placedalong the branch canals, probably still did not extendmore than 15 kilometers or so to either side.

The major watercourses, though derived from whathad initially been natural channels, gradually assumedan increasingly canallike regime as the inhabitants of thearea undertook to dike, straighten, and deepen them, pri-marily to assure the passage into the cities of barges withbulk foodstuffs and other riverine commerce. Substantialbrick-lined dams and reservoirs also were within therealm of construction capability. Evidence for the actualconstruction of large facilities is rare, however, and mostof the technical terms associated with them seem to havebeen introduced no earlier than the mid-third millennium.This was five centuries or more after the appearance oftrue cities by any definition. It suggests that, if anything,large-scale, complex irrigation practices were a derivativeof the prior development of urban and state organization,rather than vice versa.

One should bear in mind that this second watercourseconfiguration was not for the most part a product ofplanned, systematic construction but instead was an out-growth of many centuries of small-scale modificationsand improvements. The individual branch canals, which

245

oi.uchicago.edu

Conclusion

were the operating units of the agricultural economy, laywell within the capabilities of local groups to constructand maintain (Adams, n.d.). Nor is there anything tosuggest state involvement in the processes of water-sharing and allocation at the local level, or more than rarestate intervention in the resolution of small-scale, localwater disputes. A number of royal inscriptions proclaimstate initiatives in canal construction, to be sure, and itcan be shown that the necessary funds for some of themcame from state treasuries. But the execution of suchprojects was generally left in local hands, subject to vary-ing degrees of state supervision. There are references tocorvee labor that must have been applied in part to im-provements in the river or canal system, although at leastin existing sources other civic projects seem to have re-ceived the lion's share of it. Moreover, textual evidencesuggests that much of the basic construction as well asmaintenance often was assigned to local contractors ofhired labor who acted as at least semi-independent en-trepreneurs.

All in all, the irrigation system and the major institu-tions of the urban society seem to have coexisted withlittle to suggest their close interaction or interdependence.Rather than the state using its necessary control overirrigation as a source of wider political leverage, theevidence (admittedly limited and not conclusive) couldbetter be interpreted as indicating that involvement inirrigation was one of the means by which the state some-times sought to extend its control over the countryside.And characteristic of this second configuration was a stilllimited scale of network development, essentially confinedto enclaves having a single city and its outlying dependen-cies as the service area rather than building up a systemof primary, secondary, and tertiary canals that servednumerous cities and had an interregional character.

At least at the time of its clearest and most extensivearticulation in the Sasanian period, the third configurationwas dramatically different from the second. However, itsvery massiveness and geographical all-inclusiveness thenhas meant that there is little opportunity to observe thesteps by which it may have been more gradually intro-duced from Neo-Babylonian times onward. Moreover,there must have been at least foreshadowings of it alreadyin much earlier times. If our information were more ade-quate, the Ur III irrigation system might be found to havein some respects closer relationships with the Sasanianconfiguration-in complexity, certainly not in scale-than with the preceding one to which it is assigned partlyon chronological grounds. It also seems not unlikely thatthe intensive development of the northern part of theMesopotamian plain that began with the First Dynastyof Babylon could have been characterized by similardevelopments.

However, the possibility of a significant break or turn-ing point before the introduction of an essentially different

configuration in the late periods cannot be dismissed.The increasingly large-scale use of new types of moreeffective lifting machinery to supplement gravity-flowcanals, after roughly the middle of the first millenniumB.C., would have been perhaps the most significant ele-ment of change. There is no way yet to quantify its effects,but we have seen that the ubiquity of such machinery byno later than the early Sasanian period is well attested inthe Talmud. Similarly, movement from the Euphrates tothe Tigris as the vital axis of political and commercial life,not to speak of the vast extension of the Sasanian canalsystem that depended on a considerable if supplementaluse of Tigris water, would have been unthinkable withoutthe new devices. At a quite different level, the evidenceof agricultural intensification as early as the Achaemenianperiod provides other hints of a new approach. In particu-lar, the relatively very high charges then levied for whatamounted to the use of water from state-owned canalsseems to imply an awareness of possibilities of economicexpansion that were tied to an ascending rate of stateinvestment in the irrigation infrastructure.

The main features of the new configuration are clear,even if their antecedents thus remain somewhat doubtful.Very large districts, covering not much less than the en-tire alluvial land surface that was not rendered unculti-vable by swamps or other problems, were systematicallygridded with canals of large dimensions. Their straightcourses, fairly even spatial distribution and geometriclayout, and various technical features reflect centraliza-tion not merely of planning but of at least some phases ofconstruction and maintenance. Major artificial arterieslike the still-impressive bed of a canal known as the Shattal-Nil north of Nippur could be undertaken virtuallywithout limitation of size or length; the Nahrawan canal-in the Diyala region is merely the largest and heretoforemost extensively studied example of this (Adams 1965).

The ambitiousness and complexity of these schemesstrongly suggests that components of the irrigation systemof the time were planned and constructed according touniform standards by full-time, specialized, technicallyvery competent cadres. This presupposes a new and moreintensive degree of state involvement in initiaty con-structing, and probably also in maintaining, the majorarteries of the irrigation system. Perhaps the most signifi-cant achievement under the new auspices, and one thatplaced the greatest demands on administrators who couldno longer operate exclusively within a framework of localor parochial interests, was the integration of Tigris andEuphrates water supplies in order to maximize the extentof cultivation. What is more difficult to establish, how-ever, is how far central control extended into the sec-ondary and tertiary networks of canals that were fed fromthe main arteries.

Existing accounts make no reference to state officials'taking up residence in the lesser settlements and rural

246

oi.uchicago.edu

Conclusion

districts, as they would have had to do if the entire irriga-tion system were under their active supervision. The morenormal pattern, as in the collection of head taxes andland taxes, probably was to make at most a few visita-tions a year while leaving the implementation of thestate's demands to local notables. Together with the con-siderable autonomy of internal administration on the partof the religious communities, this suggests that state in-tervention was sporadic and largely nominal, except fora handful of the larger towns and strategic control pointson the major roads and watercourses. Applied to irriga-tion, this almost certainly would have meant that waterin only the main canals was under state control. It wouldthen be supplied to even the major secondary canals onlyfor specified periods on some rotational principle, andwithin the terrain served by each of them water allocationand the settling of disputes over it would have been leftentirely in local hands.

Thus the administrative veneer seems on the wholeto have been fairly thin. Not reaching deep into thefabric of rural life, the irrigation administration was notan effective lever of wider political control. Cutting offthe supplies of whole districts was too general and self-destructive a weapon to be wielded against even the mostmenacing coalitions of opposing forces, and there is noevidence that the state ever resorted to its use. Similarly,although the regime was in a position to enforce severediscipline among its own cadres of irrigation specialists(and reportedly did so during the flood disaster in 628),its control over masses of corvee labor would have had tobe more indirect. How extensively corvees were used inconnection with irrigation construction is therefore quiteuncertain. Al-Hajjaj's employment of corvee labor forthis purpose in the Early Islamic period seems to havebeen regarded as a noteworthy and hence rare if notexceptional instance, at least from the perspective of histimes and quite possibly with reference to the precedingSasanian period as well. Corvees or their functionalequivalents must have been convoked by local notablesfor service on the lesser components of the system forwhich they were responsible, as they still are today bytribal shaykhs or their sirkals in southern Iraq. But thiscan in no sense be regarded as a manifestation of despoticstate power stemming from its control over the irrigationsystem. The state's best alternative accordingly was todevelop its own corps of specialized workers for serviceon the larger projects of irrigation canal (and perhapsother) construction. That practice is well attested for the'Abbasid Caliphate, as we saw in chapter 5, and its oc-currence then seems likely to be merely a restoration ofthe Sasanian pattern.

Once more, therefore, we are left with a general pic-ture in which the institutions of central government andthe irrigation system exhibit only a limited, loose couplingwith one another. Irrigation agriculture was explicitly

recognized as a principal prop to the prosperity of therealm and hence to at least the fiscal strength of the state,but the areas of direct intervention that were consideredfeasible were very circumscribed. The strategy of thecrown, during the fairly limited periods of its relativeascendancy over the landed nobility and other particular-istic forces that generally opposed it, was not so much totransform existing social and economic structures as todeploy its strength outside of them-founding new cities,campaigning across the Byzantine frontiers, and so forth.The instituting of a nobility of service under Khusrau IAnosharwan may represent a departure from this practicefor a time in the later Sasanian period, in the direction ofcountering the nobles' strength with greater administra-tive or bureaucratic "density." But in the progressivebreakdown of central power that ensued under his suc-cessors the balance tipped once more in the oppositedirection.

Impressive as some of the technical achievements con-nected with this third configuration certainly were, fur-thermore, we should not lose sight of the fact that theyfailed to enhance the fundamental stability of the state.While greatly expanding agricultural production and pro-viding some protection against minor, local variations inoutput, they even intensified certain other problems. Thegiant new canal networks disrupted earlier, natural drain-age patterns, for example, and hence must have beenquickly followed by serious difficulties with salinization.Other declines in productivity would have resulted fromthe expansion of the canal system into areas previouslyjudged less fertile and rewarding than those already incultivation. Mention was also made in chapter 5 of thevirtual certainty that the proportion, and probably alsothe absolute number, of livestock fell, with the consequentpartial loss of this important reserve subsistence source.

Against these negative factors there was, to be sure, agreat extension in the aggregate cultivated area and pre-sumably also some improvement in the supply of waterto the individual cultivator. But the continuation of thiswas now increasingly dependent on the effective func-tioning of a specialized corps of engineers and administra-tors. If the state's executive authority was compromisedby dynastic intrigue or dissolved in chaos, this corps alsosoon abandoned its work and began to drift away fromits posts. Nothing had changed to give greater assuranceof survival to the state itself. But in the event of its collapsethe new irrigation facilities of unprecedented scale andcomplexity fell back on local groups without the re-sources or experience to maintain them.

The relationship between irrigation and many of themost pervasive features of ancient Mesopotamian societythus seems to have been very substantial, but it is alsorelatively generalized and hence difficult to document.A dense population was sustained on the basis of irriga-tion agriculture, of which a high proportion could be

247

oi.uchicago.edu

Conclusion

engaged not in primary subsistence pursuits but in bur-geoning crafts, commerce, cults, and administration-theactivities we rightly think of as constituting the hallmarksof any ancient civilization. The uncertainties inherent inirrigation provided an important stimulus to the growthof cities and of hierarchically organized institutions withinthem that could serve to husband and redistribute sub-sistence and other resources. Also inherent in it wereinequalities of access to the main factors of production,and these must have at least accelerated the growth ofsystems of social stratification. But on the other hand irri-gation seems to have had relatively little direct influenceon the growth, powers, forms, or legitimation of the in-stitutions of centralized authority. A more indirect in-fluence in this realm as well is not to be denied, of course,since it is difficult to imagine the existence of any of thoseinstitutions except in the politically unstable, stratified,urban context that in southern Mesopotamia irrigationwas primarily responsible for making possible.

ANCIENT MESOPOTAMIA AS AN URBAN SOCIETY

The other empirical foundation of this study is thephysical remains of ancient settlements, urban and other-wise. These have been the primary objects of discussionin many of the foregoing chapters, so that the recon-struction of a developmental sequence paralleling thatjust given for the irrigation system would be unwarranted.Analytically, however, they present a very different typeof problem. In the narrow sense, the clusterings of brokenbrick and potsherds that the archaeologist finds aremerely nodes of concentration in a series of overlapping,but not identical, subsystems of activity: residence, gov-ernment, exchange, worship, redistribution, and so forth.But in a larger sense they are symbolically and practicallyoften almost coterminous not merely with the exerciseof political power but with the outstanding culturalachievements of the civilization. Clearly, it is beyond thescope of this study to deal with ancient settlements,particularly urban ones, as cultural exemplars in thislarger sense. But, once again, we cannot justifiably evadethe question whether, and to what extent, the urban basisof Mesopotamian settlement exercised an enduring, dis-tinctive influence on the whole of the region and on thesociety there.

One aspect of this theme has already been adumbratedin chapter 3, in connection with a discussion of states andcities as the principal focuses of alternative researchparadigms or strategies. States, capable of being formu-lated in terms of specified structures of activity whosepresence or absence is presumably "testable," offer whatsome scholars think are decisive advantages. Urbanismis indeed a much less rigorous construct to investigate,and its significance either as an influence on the largerinstitutional structure or merely as a characterization of

state structure is more likely to vary according to con-text. But particularly when defined as part of an essen-tially archaeological framework of investigation, the statetends to become an artifact of the measurements takento determine its presence or absence rather than a recog-nizable form of governance. What is likely to have beenfor a long time a changing, indeterminate mixture ofdifferent modes of social integration is too easily seen inonly its narrowly operational, political-administrativeaspects. The probability that the latter were of some im-portance even at the very outset of urban life is certainlynot to be denied. However, the flux that has been evidentthroughout this account of the Mesopotamian sociallandscape suggests that politically or administrativelyimplemented hierarchies of power often led a fluctuating,

> contingent existence. Cities, meanwhile, constituted themost visible and enduring realities of social life. That iswhy it has seemed more worthwhile here to regard themas the primary object of study.

In a way, Mesopotamian cities can almost be said tobe coterminous with the underlying conditions for civil-ized life as it materialized there. At the heart of cities werelarge, interlocking institutions administered by relativelyself-conscious elites, whose thoughts and activities as in-dividuals alone have come down to us through the claytablets they left. Only in the cities were to be found themonumental repositories of religious belief and tradition.Also virtually limited to an urban context were capabili-ties for abstract thought and hence for complex, long-range decision-making, insofar as these presupposedliteracy and corporate, transmissible memory embodiedin written archives. Finally, as I have already noted,within the capabilities of ancient technology it was citiesthat offered a strictly relative but still decisive degree ofprotection against natural disaster and military threat.

But to define all these as urban phenomena may implytaking for granted that the role of urbanism was all-in-clusive, rather than formulating an open question towhich a qualified or even a negative answer is conceiv-able. No matter how the question is asked, however, itis very difficult to provide a satisfactory answer whenthe available documentary materials rather narrowly re-flect the views and preoccupations of a limited upperstratum of the population wholly committed to an urbanway of life. And archaeologists have unfortunately abettedthis distortion by an almost complete lack of concernfor excavating smaller, outlying settlements that werecontemporary with the historic cities. At least until ashift in archaeological priorities makes a more balancedoverview possible, the major means for correcting theexisting picture lies in a comparison of it with the onlyavailable, although much more recent, works that doprovide some insight into the nonurban setting. Theseare the accounts that left by travelers and sojourners inthe region, part of whose attention was directly given to

248

oi.uchicago.edu

Conclusion

the countryside. Fortunately, much of this variegatedtestimony that has accumulated over the past two cen-turies or so (including a regrettably still very small pro-portion of ethnographic research) is fairly consistent inthe underlying patterns it identifies. It can be broughttogether with relatively more scattered and tenuous leadsin the ancient archaeological and textual evidence toform a coherent reconstruction, even though for a longtime to come any such reconstruction can only be highlyprovisional.

There are two principal areas of pronounced differencein emphasis between the analytical approach this studyhas taken, partly with this reconstruction of the enduringrealities of rural life in view, and the city-centered viewof all the primary ancient sources. Also contributing di-rectly to the first was the research methodology of archae-ological reconnaissance, for among its findings that can-not be ignored are gradations of size among ancient sitesoccupied during the same periods. There exists a largeand respectable body of literature in economic geo-graphy demonstrating that, at least in contemporary situa-tions where the question can be directly investigated, sizetends to be highly correlated with number and complexityof urban functions. The first approach that the data (demands, therefore, has been one that uses the shapeand differentiation of the urban hierarchy as an indirectbut important indicator of the relations between the majorancient urban centers and their hinterlands.

As I noted in chapter 4, however, the observations ofpatterned behavior that support this quite rudimentaryanalytical step seem to have had essentially no counter-part in the cognitive world of the ancient Babylonian.There is little to suggest that urban size was considered asignificant variable in its own right. The concept of city-ness apparently was essentially an all-or-nothing one,with individual cases either simply included within orexcluded from the terminological category. Subsidiarysettlements falling outside it, meanwhile, received strik-ingly little attention, in proportion not only to their con-siderable numbers but to their undoubted importancefor at least certain aspects of urban well-being. It seemsquite impossible that there were not regular, routinizedforms of active urban intervention in the affairs of manyof them, and rather likely that in some of the larger onesthese had developed to the point where scribes and minorofficials were sent on regular tours of inspection or eventerms of duty. But while some of the subordinate townsmust have been involved in this way, those below urbanstatus were rarely seen as parts of an articulated systemfor regional administration. The same is even more truewhen we posit the necessary existence of other networksinvolved in the flow of goods and services, including (butby no means limited to) those concerning tribute or taxes,exchange, and redistribution. What is instead apparent isthe lack of importance of settlements that were not de-

fined as having attained urban status, even from the nar-row economic vantage point of their urban overlords,beyond the initial fact of their subordination to a particu-lar city.

How can this curiously narrow, even dysfunctional,urbanite attitude toward smaller, dependent communitiesbe understood? In part, surely, it must reflect the rela-tively much greater security and depth of tradition, notto speak of physical amenities, to be found within thelargely self-enclosed world of the city-dweller. To thatextent it merely confirms the modern analyst's expecta-tion of an extremely circumscribed viewpoint in ancientliterary testimony. But perhaps also it partly reflects theinstability of relationships once they extended outsidethe walls of the city. Outlying dependencies, that is tosay, received so little attention because there was littleassurance of retaining any durable relationship with thoseliving in them. The lesser towns were merely pawnscaught in the continuing flux of interurban rivalries.

On the other hand, the existence of broad temporaland regional variations has already been invoked in adenial that there was any single paradigm. Therefore, asingle explanation probably should not be expected tocover the general lack of concern for the functionalaspects of urbanism that are responsible for the existenceof an urban hierarchy. We have seen that Uruk had many>outlying, smaller neighbors, most of them presumablyin some sort of dependent status, until at least the lateEarly Dynastic I period. Small settlements were signifi-cantly more transitory in the Nippur-Adab region, andin addition many of them are so situated as to makeallegiance to one or another of the nearby city-states amatter always open to disputes between the latter. Laterperiods of Mesopotamian history may offer other kindsof contrasts. As I noted in chapter 4, extremely smallcenters-"villages," in fact, according to the classificationfollowed here that is based exclusively on size-were onoccasion centers of complex administrative, craft, andeven intellectual activity in at least the northern part ofthe alluvium. Later still, with the spreading use of iron,currency, glass, and many other products binding citiesand rural regions into tighter, more continuous networksof exchange, it also seems likely that the insularity ofurban attitudes at least must have begun to soften.

The second area of broad difference between the ap-proach consistent with archaeological reconnaissancedata and the perspective taken by most of the textualsources is also related to the smaller settlements. It in-volves the later, however, not as descending links in anunbroken administrative or commercial chain but asrepresentatives of the rural hinterland more generally.As Diakonoff puts it, stress is placed on the rural com-munity "as a mechanism of self-defence and co-operationof the free and more especially the free rural populationoutside the state sector" (1975, p. 130). Underlying this

249

oi.uchicago.edu

Conclusion

view is the perception that the prevailing relationship ofcity to countryside, in all periods, was one of exploitationwithin limits primarily imposed by interurban powerrivalries. From the urban viewpoint, the assumed naturalorder was a continuing inward flow of resources up toor sometimes even beyond the capacities of the country-side for sustained replenishment. But while the vulnera-bility of the countryside is undeniable in most circum-stances, its impoverishment could not have been absolute.In disclosing for the first time the very considerable extentof rural settlement in most periods, archaeological sur-face reconnaissance confronts us with suggestions ofrural capabilities for rapid population growth and move-ment at various times-and hence of the deep-rootedtenacity and even vitality to be found there. It is onlynatural to inquire what adaptive strategies permittednumerous small groups of rural countrymen somehow tohold their own not only against the natural hazards anduncertainties that were common to all of Mesopotamiabut against predatory urban inroads as well.

To judge from the only window on the countryside wehave, with all necessary caveats over interpreting themore ancient past on the basis of tribal responses to condi-tions of Ottoman oppression that admittedly may consti-tute a kind of limiting case, an essential feature of thisrural adaptive strategy generally has been the mainte-nance of corporate social groups that pooled labor andmilitary potential while sharing risks. Kin-based groupsof this kind repeatedly come more or less clearly withinthe peripheries of urban concern, as reflected in the cunei-form record, from the beginning of the third millenniumonward. At times they appear in connection with thetransfer of titles to land, but it is to be assumed that theyoccupied more or less well defined territories irrespectiveof the changing legal formalities of ownershp in anurban context. Again to judge from recent practices,tendencies toward fission or fragmentation as a result ofongoing economic differentiation often may have beenheld back by the dispersal or rotation of individuallyassigned parcels of land. Also promoting the solidarityof such groups were reciprocal social and economic ob-ligations between the leadership and agnatically relatedlines of followers. "Tribal" groupings of the type, it istrue, could seldom offer more than scattered, temporaryresistance followed by flight or passivity when statepower was vigorously expanding. They may even ap-pear from time to time (at least in the perspective of stateadministrators) as formally sanctioned, centrally con-trolled aspects of the state's own system of functionaries.But it is clear even in the ancient, urban-oriented textualevidence that they repeatedly remerged with renewedstrength and independence when urban-based powercoalitions lost their capacity to hold in check the manyconverging forces that were always working to erode theirauthority.

This same fluidity of rural adjustment must alwayshave characterized subsistence practices. Defying tradi-tional but misleadingly rigid categories like "sedentary"and "nomadic," rural social groups in Mesopotamia haveessentially no choice but to maintain carefully balancedoptions of fairly nonintensive cultivation and pastoralism.Herds that can be driven as a group moves provide acomplementarity of resources that gives some protectionagainst crop failures, other natural disasters, and stateincursions. Moreover, the existence of herds permitsthese groups, even when conditions favor their prolongedsedentism, to maintain a larger population as both amilitary and a labor reserve than is possible on the basisof irrigation agriculture alone.

The crucial attribute is rapid adaptability in the faceof either social or environmental pressure. By shiftingback and forth in their relative emphasis on more mobilepastoralism and more sedentary cultivation, componentsof the rural society can manage to survive, and sometimesto prosper, by filling whatever niche is available. Eitherthey provide the lowermost, rural echelons in times ofstrong urban organization, or they take a more active,oppositional role when conditions for this are favorable.

1 There was, in other words, another side to the coin ofMesopotamia's extraordinary precocity in the develop-ment of cities. This appears to have been the deep-rootedpersistence in its countryside of essentially defensiveformations of semisedentary, tribalized peasantry.

This is not intended as a denial that in most periodsthe urban forces decidedly held the upper hand. Theirstrength was firmly founded on the internal articulationof their complex, large-scale institutions and reserves.But there remained a significant niche on urban periph-eries and in the countryside for more egalitarian societiesfollowing a strategy that stressed not stability but re-silience (Adams 1978), and some of them continue inthis strategy today. Adding to the difficulties of fully rec-ognizing the role of their counterparts in antiquity isthe fact that written sources tend to be least abundantduring periods when dissident rural forces found theirgreatest scope for action. The reverse is also true, so thatthere tends to be a disproportionate historical emphasisgiven to the-relatively rare and brief-periods whenthere was a firm and undisputed urban or imperial powerin place. But the more common condition throughoutmost of antiquity must have been a shifting mosaic of,/hierarchical, stabilizing tendencies associated with theascendancy of one or another city, and leveling, frag-menting ones that extended deep into the countrysidewhen its powers were seen to diminish.

The discussion in this section heretofore has certainlynot been intended to deny the generally accepted proposi-tion that Mesopotamia was a "land of cities." It has,however, introduced a few important qualifications tothe all-prevailing dominance and uniformity of urban

250

oi.uchicago.edu

Conclusion

conditions that may have seemed implicit in such a de-scription. There is one further step of the same kind tobe taken, involving the spatial delimitation of how muchwas in fact urban within the historic and geographic en-tity of Mesopotamia. Archaeological survey has pro-ceeded to a point at which there is more or less detailedinformation on ancient settlement patterns in most of thearea that on textual grounds was of any real importancebefore the latter part of the first millennium B.C. Someregions can be generalized about with more confidencethan others, to be sure, and there are problems andlacunae remaining in all of them. But the largest lacuna(the ancient kingdom of Lagash, most or all of whichlay east of the modern Shatt al-Gharraf in terrain stillawaiting an intensive topographic study), while clearlyvery important, is of relatively modest size in comparisonwith the area that has now been studied.

It is no longer possible, therefore, to imagine that theonly limits of urban settlement in the pre-Hellenisticpast were ever those of the alluvium itself. The same con-clusion, one will recall, has been reached on the basis ofa discussion of the irrigation potential of the Euphrates.We now know that the nucleus around which urban lifeformed and prospered was for a very long time a rela-tively limited part of the alluvium, and that even withinthe more limited zone there were large differences andperiodic shifts in the pattern of urban size and density.With all the caution that the still rather heterogeneous aswell as incomplete data make necessary, it may be worth-while to outline briefly the growing as well as changingpatterns of internal zonation that should be kept inmind when the idea of a "land of cities" is invoked.

The southermost part of the alluvium seems to haveplayed the key role during the earliest period of Holocenesettlement, probably extending back into the sixth mil-lennium. This may well have centered on the regionarounil Ur that is covered in Henry Wright's concludingstudy, but it probably extended upward for some distanceinto the Uruk region as well. Only future study can de-termine whether it also included part of the kingdom ofLagash to the northeast, but it would not be surprisingif the Gulf shoreline receded from much of this areaonly at a considerably later period.

Apparently quite abruptly and without obvious sourceor explanation, a part of the central floodplain thenmoved to the fore soon after the beginning of the fourthmillennium, in terms of both settlement size and density.Not long afterward, sites had appeared that are un-ambiguously to be recognized as cities. But even afterseveral centuries of further growth it appears that westill are dealing with a population for the whole alluviumof fewer than a hundred thousand persons, and that thesmall, scattered enclaves in which they lived were stilllargely confined to a relatively narrow band across itssouthern portion.

The increasingly dense settlement of the Uruk area inthe Late Uruk period initiated a new period of populationgrowth combined with migration, in some as yet unde-fined proportion. The southern part of the central flood-plain remained the major locus of relatively high popu-lation density, although there was some thickening of verysparsely distributed settlements in other areas as well.Probably the first significant wider dispersal occurred atabout the beginning of the third millennium, as is sug-gested by Kish's having moved into a position of politicalprominence well to the north. The lower Diyala region,on the other hand, continued for several millennia moreto be relatively lightly occupied and hence less important.

There is no evidence of a substantial alteration of theseregional relationships through most of the third millen-nium, although almost everywhere the pattern becamedenser as well as more continuous. At about the end ofthe period, during the Third Dynasty of Ur, populationappears to have crept upward past the half-million mark.The whole northern region of ancient Akkad, as well asthe Diyala region around Eshnunna across the Tigris,remained distinctly tarriant in terms of maximum settle-ment size, average settlement size, and population density.

The rise of Babylon substantially altered this in theeighteenth century B.c. There was a cessation of scribalactivities, if not necessarily of all accompanying occupa-tion, in a number of the southern cities, and a corres-ponding extension of settlement along what may well benewly constructed canals in the region north and east ofBabylon. Salinization that had driven down agriculturalproductivity in the south may have set the stage forthis, but it seems inadequate as the immediately pre-cipitating agent for an event so abrupt and general. Thefirst of an irregular series of westward shifts of theEuphrates also may have been a factor. This in turnopens the possibility of substantial new settlement inregions south of Babylon where survey is made inconclui-sive by massive, more recent Euphrates levee deposits. \Hence it is almost impossible to hazard an estimate foraggregate population in the later second and early firstmillennia. To judge from the central floodplain, it hadshrunk back to mid-fourth millennium levels, but thisdecline may have been partly compensated for by thefounding of new dependencies of Babylon in districts tothe south of it.

From this time onward there are increasing difficultiesin providing even a rough and impressionistic overview.Part of the effect of the shift to the Tigris, completed byHellenistic times, can be plotted with some accuracy inthe immense growth of new settlement in the lowerDiyala region. A similar but less marked process may wellgo on in part of northern Akkad, held back by the collapseof Babylon and the probable decay of its hinterlands andperhaps also by the destructive effects of repeated Romaninvasions. But as yet we have no basis for even guessing

251

oi.uchicago.edu

Conclusion

at the corresponding developments along the lower Tigris,in Mesene and Characene. If what went on there cor-responded to the Diyala plains behind Ctesiphon, thenthe Parthian period may have been a time of more rapidextension than any. It was, after all, a time when thedensity of occupation in the old urban nuclear zonealong the southern levees in the central floodplain alsoincreased very significantly.

The impressive extent of irrigation in the Sasanianperiod has already been discussed and presumably eitherwas initiated by or led to further demographic growth.Some formerly cultivated land was lost to swamps in thesouth, but the intensification of canal construction thatcan be traced almost everywhere must have more thanmade up for this. Certainly there is evidence of a greatincrease in settlement numbers and density, and what isknown of the history of the period does not suggest thatconditions could have been different along the lowerTigris than in the regions already surveyed. This, it ap-pears virtually beyond doubt, was the apogee of the an-cient settlement and irrigation record in every respect.Yet, as I suggested in chapter 5 and again in the preced-ing section, there is persausive evidence that the agri-cultural economy had begun to deteriorate seriously wellbefore the onset of the Arab raids that culminated in theIslamic conquest. Extrapolating from what is known tothe many important regions that are not, the populationof the Tigris-Euphrates alluvium at the apex of the dy-nasty's strength and prosperity seems certain to have beenwell over one million persons and perhaps even ap-proached two million.

The problems involved in estimating Early Islamicpopulation have also been treated in chapter 5. On theone hand, much of the central floodplain was either en-gulfed by the spread of the Great Swamp or abandonedfor other reasons. There was a considerable decline alsoin the Diyala region, to the north and east of the middleTigris. On the other hand, documentary sources indi-cate that this was a time of some real growth based onswamp drainage and reclamation, principally along thelower Tigris above Basra. How nearly these two processesbalanced out is quite uncertain from textual sources. Italso may not be easily resolved by future archaeologicalsurveys, since most of the region today is given over tointensive palm cultivation. What is not in doubt, how-ever, is that by no later than the late ninth or early tenthcentury, and perhaps somewhat earlier in a few regionslike the central floodplain, a general retraction of thecultivation frontier got under way. Within a few centuries,population levels had been carried back almost to whatthey were in the mid-fourth millennium. By the time ofthe fall of the caliphate in the mid-thirteenth century,the limits of sedentary life were being more and morenarrowly drawn around the outskirts of Baghdad, Basra,and a handful of other towns. And it was only these few

towns that precariously managed to carry the continuoustradition of urbanism, in the land of its origins, throughuntil the threshold of modern times.

The conception of a "land of cities" thus can be con-sidered most nearly coterminous with the whole Mesopo-tamian alluvium as a physiographic region during theParthian and Sasanian periods. At other times, later aswell as earlier, it is applicable only to considerably lessinclusive areas. Particularly during the fifteen hundredto two thousand formative years, what might be thoughtof as its nuclear zone grew only slowly from a very fewthousand square kilometers to a figure hardly more than

y twice as large. Considered against the 50 to 60 thousandsquare kilometers that were potentially irrigable andcultivable (uncertainties over the position of the Gulfshoreline, as well as the extent of swamps, make a closerestimate unrealistic), this is a modest figure indeed. Evenwithin the nuclear zone, moreover, the computer-gen-erated simulations of land use in chapter 3 indicate thatuntil almost the middle of the third millennium the cul-tivated areas in most cases were of limited individual sizeand were kept relatively distinct if not actually isolatedfrom one another.

Considering the entire third millennium as well as thefourth millennium, the cumulatively substantial demo-graphic growth that did occur probably was accommo-dated more through the expansion and coalescence ofmany of the internal enclaves than through marked orrapid extensions of the nuclear zone itself. The contrastbetween the apparent fluctuations of occupancy evenin the immediately adjoining Ur region and the relativelysolid, entrenched continuity in the central floodplain maybe especially significant in this regard. Important as Urwas in religious, political, and commercial terms, thereis a suggestion here that it was somehow compromisedby a position well to the south of the demographic as wellas urban core.

This is the sense, then, in which it still seems apt andaccurate to speak of a "heartland." There was prodigi-ously creative phase in the growth of a civilization duringwhich its developing centers somehow interacted syn-ergistically in spite of their rivalries. Many other con-siderations seem to have favored the alternative path ofwide dispersal, but the outcome was the consolidationof a region of at least relatively massive, interlocking, andirreducible demographic and economic strength. Yet itis important to remember also that the pattern was un-dermined and finally destroyed utterly in this particularsetting, although only after the initiation of its subse-quent, worldwide spread. We can reasonably concludethat it was not generated by any unique propensities ofthe landscape, and that we must look instead to the hu-man forces that were harnessed in the building of thecities themselves.

252

oi.uchicago.edu

7General Site

Catalog

(For descriptions of sites 001-466 see Adams and Nissen 1972, Appendix.)

501 - 250 NNW X 180 X 3.2. Little Parthian,mainly Sasanian-Early Islamic.

502 - 180 E X 140 X 3. Middle/Late Islamic.503 Tfilfil Jumali. 700 NW X 200-380 (see base map

for irregular outline) X 4.5. Sasanian-Early Is-lamic.

504 Tfilfil Jumali. 220 diam. X 5. Middle/LateIslamic.

505 Perhaps 80 diam. X 0.6, but site boundariesare blurred by adjacent and underlying canal spoilbanks. Sasanian-possible Early Islamic.

506 130 NW X 80 X 0.8, extending over a por-tion of the former bed of the Shatt al-Nil. LateIslamic.

507 - 90 NW X 40 X 1.8, situated between twoparallel branches of the Shatt al-Nil. Late Islamic.

508 - 350 diam. X 4, bisected into E and W com-ponents by the broad bed of the Shatt al-Nil. Early/Middle/Late Islamic.

509 - 190 NNE X 120 X 1.4. On opposite bankof ancient levee is a slightly smaller mound, 150diam X 1.4. 300 m WSW is a third, 130 diam. X2.3. Late Islamic.

510 - 500 NE X 350 X 6.5. Sasanian, a littleEarly Islamic.

511 -- 300 diam. X 4.5, consisting of a loosegrouping of mounds with some plain-level debrisrather than a nucleated settlement. Sasanian-EarlyIslamic.

512 - Scattered small mounds on both sides of theShatt al-Nil, forming a loose group 350 diam. X2.5-3. Sasanian-Early Islamic.

513 180 diam. X 3.5. Sasanian-possible EarlyIslamic.

514 -- 200 E X 140 X 3.5. Achaemenian-Parth-ian.

515 180 NNE X 130 X 1. Parthian.516 The two arms of this site, meeting at an

oblique angle as shown on base map, consist ofcompact masses of debris of fairly uniform eleva-tion except where crosscut by a canal. WNW arm700 X 240 X 4.5; ENE arm 550 X 300 X 4. Greatquantities of black kiln fragments and chunks ofblack basalt from milling stones litter all portionsof the surface. Mainly Neo-Babylonian-Achae-menian, some Parthian.

517 Ishan al-Wa'y. The area of perceptible elevationis approximately 500 diam. The distinct mound,coinciding with substantial sherd concentrations,is 340 N X 240 X 5.5. Trace of Uruk. Achae-menian-Parthian.

518 - 130 diam. X 4.5, with smaller, lower out-liers immediately SE and SSW. Parthian sherdsmay be only strays from neighboring site 517.Mainly Sasanian-Early/Middle Islamic.

519 180 N X 100 X 5. Farther S along old leveeare smaller mounds up to 90 diam. X 3.5, inter-spersed with lower debris forming a narrow butfairly continuous ribbon of settlement. The mainmound is Patthian-Sasanian. The smaller moundsto the S are Parthian-Sasanian-Early Islamic, butpreponderantly Sasanian.

520 Ishan Abi~ Judu. 260 diam. X 6. Smaller, lowmounds are adjacent S and NNE. The S moundis Parthian, the NNE mound Sasanian, and themain mound Early/Middle/Late Islamic.

521 - 900 ENE, a strip settlement along an an-cient canal levee. The WSW part of site is less than100 width, very low; the ENE approaches 200width and rises to 2 m. About 700 m W of thecenter of the site is a second, 130 NNW X 80 X0.9. The WSW end of the main mound and the

253

oi.uchicago.edu

General Site Catalog

second mound are only Sasanian. The ENE end ofthe main mound is also Sasanian but continuesinto the Early Islamic and Samarran periods.

522 - 80 diam., mostly plain level, highest hum-mocks of debris only 0.3. Sasanian.

523 90 diam. X 0.8. Sasanian-Early Islamic.524 220 NW X 100 X 3.5. Middle/Late Is-

lamic.525 Probably 180 diam. X 1.4, although di-

mensions are partly obscured by a superimposeddune. Sasanian.

526 240 E X 150 X 2.5. Sasanian.527 120 diam. X 2.2, with clusters of debris at

plain level extending SW. Early/Middle/Late Is-lamic.

528 350 N X 150, rising in widely spaced sum-mits near N and S ends to 2.6. Parthian.

529 130 diam. X 0.3. Middle Babylonian.530 130 NNE X 90 X 2.8. Sasanian-Early/

Middle Islamic.531 170 E X 90 X 2. Neo-Babylonian-Achae-

menian-possibly limited Parthian.532 At least 1,000 NNW X 250, with numerous

semidetached mounds rising 3.5-4.0 and a fewsmaller, conical mounds to 5 m. Sasanian.

533 - 220 E X 140 X 0.8. Parthian-Sasanian.534 Ishan al-Khor. The main mound is 340 N X

240 X 6.5, but continuous debris and low out-lying summits that may be integral with the siteextend 100 m N and 500 m SSE. The S end of thesite is Parthian-Sasanian, the N end Early/Mid-dle/Late Islamic.

535 - An area of debris at plain level that may besomewhat arbitrarily defined as 220 N X 120; oc-casional hummocks rise to 0.4 m. primarily OldBabylonian-Cassite, some Sasanian.

536 70 diam. X 1.3. Early/Middle/Late Is-lamic.

537 170 diam. X 2.2. Early/Middle/Late Is-lamic.

538 80 diam. X 0.6. Middle (-Ilkhanid)/LateIslamic.

539 - Perhaps 200 NNW X 100, but with onlysparse surface debris at plain level so that limitsare hard to define. Plentiful shells probably sug-gest recent, rather than necessarily ancient, swampin this vicinity. Late Uruk; note extreme rarity ofclay sickles.

540 - 240 NW X 150 X 3. Sasanian-Early/Mid-dle Islamic (-Ilkhanid).

541 Imam Najmi. 550 NW X 350 X 4.5, with an-cient canal bed separating S end from major partof mound. The well-known shrine has been mostrecently discussed by Paolo Costa (1971). Early/Middle/Late Islamic.

542 Probably this is the mound named by Kiepert(1883) Tell Abfi Rusiyat. 900 NNW X 350 X6. Possible Neo-Babylonian-Achaemenian, mainlySeleucid-Parthian.

543 - 140 NE X 110 X 1.2. There is also a smallsettlement of low, discontinuous mounds on theopposite, left bank of adjacent ancient water-course. Sasanian.

544 254 NNE X 120 X 0.4. Sasanian.545 60 diam. X 0.4. Sasanian.546 360 N X 280 X 5.5 Achaemenian-Parth-

ian.547 90 N X 70 X 0.6. Sasanian.548 110 diam. X 1.6. Adjacent W is a small

conical mound that may be an ancient brick kiln.Parthian.

549 Eight or more individual mounds, looselygrouped in an area 700 NW X 500, as shown onbase map. Generally they are small, less than 100diam. X 1.5. Surface debris over the entire areais dense and continuous. Parthian-Sasanian.

550 - 140 N X 90 X 1.4, although a line of dunesoverlies the N end so that the mound may continuefarther in that direction. Sasanian.

551 - 300 ENE X 220 X 3.5, with low debris ex-tending an additional 120 m E. Possible Achaemen-ian, mainly Seleucid-Parthian.

552 Scattered, generally low mounds and fairlycontinuous plain-level debris within an area 700diam. The most prominent summit is W centeral,180 diam. X 4.5. Achaemenian-Parthian.

553 - 90 diam. X 2.6. Neo-Babylonian-Achae-menian-Parthian.

554 - 220 WNW X 180 X 2.2. 100 m NNE is asecond, 220 WNW X 120 X 2.2. Much brick-kilndebris. Sasanian.

555 420 diam. X 4.5. In addition, there is a stripof debris 30-80 m wide extending WNW along oldlevee; more scattered hummocks also occur ESE.Parthian.

556 130 diam. X 2.4. Parthian.557 90 diam. X 2. Some pottery-kiln debris.

Parthian, possibly also Sasanian.558 280 diam. X 3. Sasanian-Early Islamic.559 - 130 NW X 100 X 2. 150 m N is a second,

190 diam. X 1.8. Sasanian-Early Islamic.560 - 160 E X 120 X 1.8. Sasanian-Early Is-

lamic.561 - 190 NNE X 140 X 0.6. Parthian.562 - 160 NNW X 90 X 1.2. Trace of Uruk.

Sasanian.563 - 550 NNE X 300 X 3.2. Parthian.564 - 180 E X 120 X 1.8. Middle/Late Islamic.565 - 190 ENE X 100 X 1. Parthian-Sasanian.566 - 500 NE, varying in width from 100 to 250

254

oi.uchicago.edu

General Site Catalog

m, ht. to 2.5 m. 250 m NW is a second, 70diam. X 0.6. Sasanian.

567 160 NW X 90 X 2.2. Late Islamic.568 220 NW X 120 X 2.2. Sasanian.569 130 NW X 80 X 1. Sasanian.570 140 diam. X 2.2. 30-31 cm sq. bricks. Sa-

sanian. The brick size may imply terminal Sasanianor even Early Islamic, but none of the pottery re-garded in this study as diagnostic of the latter wasidentified.

571 - At least 250 diam. X 2.4, but limits of siteare obscured by drifting sand. A canal from NWbisects the site but is obviously later, since its spoilbanks are continuous through and beyond it andrather prominent. Trace of Parthian. Mainly Sa-sanian. There is a little scattered Late Islamic pot-tery along the spoil banks of the canal both NWand SE of the site itself.

572 130 NW X 60 X 0.3. Late Islamic.573 110 N X 60 X 0.4, although the N end has

almost certainly been cut away by the trench andaccompanying embankments of the main drainagechannel of the Mussayib irrigation project. Inten-sive collection made in circle of 5 m radius, center.Trace of Ubaid III. Mainly Early/Middle Uruk;Late Uruk limited to S end. There are a few latergraves, at least some of them Ur III-Larsa, and alittle Sasanian pottery also.

574 170 m SSW of 573. N mound 70 diam. X0.2; S mound, 130 NNE X 30 X 0.2, may only bea long tail to the N mound or may be slightly de-tached. Surface debris generally fairly sparse, al-though rich in artifacts of excellent craftsmanship.At least one apparent kiln was observed near Nend. Intensive collection made in circle of 5 mradius near S end. Middle/Late Uruk. Note thinand exceptionally well made stone bowls, muchchipped flint, numerous spindle whorls.

575 - 110 diam. X 0.4. There are very extensive,thick deposits of mussel shells NE and SE ofthe site, probably reflecting a relatively recentswamp in this vicinity. Middle (-Ilkhanid)/LateIslamic.

576 - Low hummocks of debris within an area 80NNW X 10. Cassite-Middle/Neo-Babylonian.

577 - 240 WNW X 90 X 2.5. Sasanian-Early Is-lamic.

578 - 90 WNW X 60 X 0.5. Middle (-Ilkhanid)/Late Islamic.

579 - Irregular, small area of debris at plain level.Parthian.

580 - 400 E X 240 X 4. Possible Achaemenian,mainly Parthian.

581 -- 250 diam. X 5. Sasanian.582 - A string of semidetached mounds along an

ancient levee, 1,900 NW X 200-370 X 2.5-3.5.Sasanian-Early Islamic.

583 250 WNW X 100 X 1. Some Middle Is-lamic, mainly Late Islamic.

584 Hummocks of debris suggesting at most avery small settlement, along the spoil banks of anold, low canal levee. Middle Babylonian, possibleNeo-Babylonian. Trace of Parthian.

585 Brick-kiln debris. Parthian-Sasanian.586 90 E X 70 X 0.2, with another still smaller

(20 diam.) outcrop 200 m ENE. Site 585 is 30 mW. Ur III-Larsa.

587 700 WNW X 180 X with two 4.5 m sum-mits spaced about 200 m apart along the axis ofthe site. A few 21 cm sq. brick observed. ProbablyParthian, mainly Sasanian. Apparently there wasa small, continuing Early Islamic settlement, andthe 21 cm brick may reflect some building evenlater.

588 160 WNW X 120 X 2.5, with a smallermound adjoining ESE. Sasanian.

589 More than 500 ENE X 250 X 6, with mainsummits at opposite ends. Lower debris tails offseveral hundred m farther ENE. Parthian-Sasan-ian. Some Early Islamic pottery is confined to ENEend.

590 180 E X 80 X 1. Seleucid-Parthian.591 220 diam. X 2. Sasanian-Early Islamic.592 From its size and location, this is possibly

to be identified with the Talmudic Pum Nehara(Obermeyer 1929, p. 96). Approximately 800 NWX 450 X 5, but cf. base map for irregular outline.There is a NE outlier, 350 NE X 300 X 4.5.Sasanian-Early Islamic.

593 Ishan Abfi Khalesi. 300 NE X 180 X 2.5. LateIslamic.

594 220 N X 100 X 2.5, with additional smallmounds 150 m N and S of main one. Sasanian-Early Islamic.

595 400 N X 200 X 7. Sasanian-Early Islamic.596 200 N X 140 X 4.5. Parthian-Sasanian.597 220 NW X 180 X 5. Sasanian.598 380 ENE X 170 X 5.5. Sasanian-Early Is-

lamic.599 - 320 NNW X 220 X 6.5. Limited Parthian,

mainly Sasanian, limited Early Islamic.600 - Dispersed, low mounds and fairly sparse

debris at plain level cover an area at least 300 m indiam. The greatest elevation, 2.5 m, is at SW.Mainly Parthian, some Sasanian-Early Islamic.

601 - 180 diam. X 3.5. 250 m SE is a second,100 diam. X 0.6. Sasanian-Early Islamic.

602 - 280 ENE X 100 X 5. Seleucid-Parthian.603 - 420 E X 320 X 3. Within these limits the

site consists mainly of dispersed mounds, but de-

255

oi.uchicago.edu

General Site Catalog

bris is fairly continuous and most of the area is atleast slightly elevated. Sasanian-Early Islamic.

604 Triangular base on W, apex pointing E. 900E X almost 600, composed of adjoining individualmounds rising to 4.5 m ht. Sasanian-sparse EarlyIslamic.

605 A loose aggregation of mounds within anarea 900 N X 400; cf. base map. Maximum ht.4.2 m at N end. Sasanian-Early Islamic.

606 120 diam. X 1, with a band of low debrisextending 300 or more m E. Seleucid-Parthian.

607 200 NNW X 120 X 4.5. A second 30 m E,of similar size but slightly lower. There is also alarge, low N extension of the first mound. Themounds are Sasanian, but on the N extension thereis also Early Islamic.

608 190 NNW X 90 X 2.2. Sasanian-Early Is-lamic.

609 110 diam. X 1. Sasanian-Early Islamic.610 170 diam. X 2. 140 m W is a second, 100

diam. X 1.8. Many 21 cm sq. bricks. Possible Sa-sanian, mainly Early Islamic. The small bricks sug-gest a continuation into Middle Islamic times, butthis could not be confirmed from the surface pot-tery.

611 320 E X 200 X 3, with the summit locatednear W end. Mainly Achaemenian-Parthian. Alittle Early Islamic.

612 - 240 E X 70 X 1.8, three small mounds inline. All are covered with brick fragments, and aneven smaller brick-pile occurs 80 m SSE. Possiblythe remains of irrigation canal headworks? Potteryvery sparse. Sasanian-Early Islamic.

613 280 diam. X 4. Sasanian.614 - 520 WNW X 350 X 7. Parthian-Sasan-

ian-Early Islamic.615 160 diam. X 1.6. Parthian.616 180 diam. X 1.4. Sasanian.617 300 E X 220 X 3.5. A little Achaemenian.

Mainly Sasanian-Early Islamic.618 450 WNW X 240 X 3. Achaemenian-

Parthian.619 - 180 NW X 100 X 2. Sasanian, rare Early

Islamic.620 Ishan MadhriAb. Slightly dispersed mounds oc-

cupy an area 1,100 NW X 350, but the principalmound bearing this name is 520 NE X 240 X 8.5.Achaemenian-Parthian.

621 Ishan Medar. 1,200 WNW X 480, dispersedmounds rather than a nucleated settlement.Mostly low, but ESE end rises to 4.5 m. Neo-Baby-lonian limited to low hummocks at the W foot ofthe site. Mainly Sasanian, including a small, iso-lated pottery kiln on a mound W of the main partof the site. Some Early Islamic.

622 160 NNW X 120 X 4. 80 m E is a second,slightly smaller and lower. A third, 140 diam. X4, is 900 m SSE. Late Islamic.

623 300 E X 250 X 5, with low, small moundscovering an area immediately W that is almost aslarge. Sasanian.

624 Tell A'isha. 280 NW X 180 X 4.5. Late Islamic.625 Tell Abui Z6fr al-Abyadh. 350 NW X 220 X

6. 50 m NW is a small mound, perhaps 50 diam.,and low hummocks of debris extend SE along oldlevee. Mainly Seleucid-Parthian, some Sasanian.

626 Tell Abfi Z6fr al-Aswad. 320 diam. X 3.5. 200m SW is a second, 160 NE X 90 X 2.2, andthere are smaller mounds S and SSE. Middle/LateIslamic.

627 Tell 'Amra. From Inspectorate of Surveys rec-ords, file 295, register 440. Recorded as Akkad-ian-Ur III. (N.b. This dating should be treated withgreat reserve. Early datings also were assigned inthese records to sites 878, 882, and 891, none ofwhich was found to have been occupied earlierthan the Neo-Babylonian period.)

628 Tell Umm al-Buni. From Inspectorate of Surveysrecords, file 39, register 1587. Recorded as Neo-Babylonian-Parthian.

629 Jemdet 'Alka. From Inspectorate of Surveys rec-ords, file 26, register 1587. Recorded as Neo-Babylonian-Parthian.

630 Tell Abu Ghadara. From Inspectorate of Surveysrecords, file 25, register 1587. Recorded as Islamic.

631 Tell Mezerir. From Inspectorate of Surveys rec-ords, file 41, register 1587. Recorded as Parthian.

632 Tell Abfi Sudayra. From Inspectorate of Surveysrecords, file 27, register 1587. Recorded as Neo-Babylonian-Parthian.

633 Tfilfil al-Tarhiya. From Inspectorate of Surveysrecords, file 282, register 440. Recorded as Islamic.

634 Bwaydhat al-Rubaydha. From Inspectorate ofSurveys records, file 28, register 1587. Recordedas Sasanian-Islamic.

635 Tilfl Tuwaymat. 170 diam. X 6. 200 m S is asecond, 160 NW X 110 X 6. A small, low,parallel mound is adjacent S of the latter. Seleucid-Parthian.

636 - 40 diam. X 1.8. Sasanian.637 --- 180 diam. X 2. Achaemenian-Parthian-

Sasanian.638 - 200 N X 140 X 2. Sasanian.639 - 1,000 NE X 650 X 2.5. Many bricks 37 X

16 X 7 and 30 X 30 X 7, numerous foundationsor walls in place. Small, underlying Uruk, Akkad-ian settlements. Overwhelmingly Ur III-Larsa.There is also a little later (Old Babylonian, Neo-Babylonian, Parthian, Sasanian) pottery, but it canbe characterized as a city with dense, continuous,

256

oi.uchicago.edu

General Site Catalog

and prolonged occupation only during the Ur III-Larsa period.

640 Tell Nuwaija. 450 E X 270 X 4, surroundedand partly covered by dunes. Early Islamic-Samarran, the W end continuing into Middleand Late Islamic times.

641 - 80 diam. X 0.7, tailing off ENE and WSWalong right bank of pronounced old canal levee.Sasanian-Early Islamic.

642 40 diam. X 0.7. Parthian-Sasanian; theceramics here suggest that the site may terminatebefore the occupation of site 641.

643 - In 1968 this site was noted as 180 NNE X100 X 5, its NNE end cut away to some extent bythe excavation of the main drainage channel of theMussayib irrigation project. Subsequently the re-mainder of the mound was almost entirely de-stroyed by the excavation of a parallel channelto substitute for the original one. A second mound,several hundred m SW, is 120 diam. X 1.1. Itsupper levels may have been scraped away, how-ever, to serve as embankment material for the sub-stitute channel. Parthian-Sasanian.

644 220 diam. X 2.5, but debris at plain levelmay define a larger area of settlement. PossiblySasanian. Early/Middle Islamic (-Ilkhanid).

645 - 140 diam. X 3, with smaller, lower moundsimmediately N across old canal levee. Sasanian-Early Islamic-Samarran.

646 100 diam. X 0.4. Sasanian-Early Islamic.647 160 NW X 100 X 1. 250 m SSW is a sec-

ond, 110 diam. X 1. Mainly Parthian, some Sasan-ian.

648 100 diam. X 4. Samarran-Middle/Late Is-lamic.

649 240 NW X 100 X 3.5. Middle Islamic(Late 'Abbasid-Ilkhanid).

650 150 E X 60 X 1.2, on left bank of old canalbed. Lesser clusters of debris also occur on op-posite bank, to the S. Parthian-Sasanian.

651 140 NNW X 80 X 1.2. Sasanian, probablyalso some Early Islamic.

652 140 E X 110 X 1. Sasanian-Early Islamic.653 140 diam. X 5.5. Immediately S is a second,

100 diam. X 5.5. A third, smaller and lower, is Nof the first. Trace of Uruk. Sasanian-Early Islamic,Late Islamic.

654 - Strip settlement along the right bank of anancient canal, 500 NNW X 30-70 X 0.8. Sasan-ian, probably a later subphase than that seen atsite 655.

655 - Strip settlement along the right bank of anancient canal, 300 NNW X 30-70 X 0.8. SmallUruk occupation. Sasanian.

656 Ishan al-Jihariz. 1,200 NW X 350 X 6, al-

though only the central mound reaches this ht.Continuous debris at plain level also extends 250m E of the main mound to outlying, lower sum-mits. A modest Parthian occupation, mostly Sasan-ian. Early Islamic debris is confined to the highest,central area. A few 21 cm sq. bricks there mayindicate a very minor later continuation, althoughthis cannot be confirmed from the observed surfacepottery.

657 80 diam. X 1. Late Islamic.658 80 diam. X 1, directly overlying and hence

postdating a major ancient canal branch. LateIslamic.

659 80 diam X 0.7. Sasanian.660 80 diam. X 1. Sasanian-Early Islamic. A

few 21 cm sq. bricks may indicate a somewhatlater continuation on a small scale.

661 180 E X 100 X 3.8. Parthian-Sasanian.662 40 diam., plain-level debris. 150 m SE is a

similar area 80 m diam. The first is Jemdet Nasr,the second Late Islamic.

663 200 E X 80, the central portion rising to3.8. Outer limits of site are obscured by dunes, butseveral smaller mounds occur 300 m SW. Parthian-Sasanian.

664 160 N X 90 X 1. Sasanian.665 See sketch map, with ht. of individual

mounds shown in m. Achaemenian-Parthian, aminor Sasanian occupation also.

. ..

666 - 140 diam. X 2.5. The limits of the site aresomewhat arbitrary, however, since surface debrisextends well beyond the limits of the perceptibleslope. Mainly Parthian, Sasanian confined to Nend.

667 - 100 diam. X 1.2, but with debris at or justabove plain level extending several hundred m Nand NE. Small Uruk settlement. Ur III-Larsa,Achaemenian. Some Parthian pottery also is pres-ent but may derive from graves only; slipper cof-fins of this date are being exposed above thesurface of the site by wind erosion.

668 - 300 NE X 140 X3.5. A second, 200 mSE, is 350 NNE X 200 X 2. Sasanian.

669 - 80 diam. X 0.8, although only the centralpart rises perceptibly above plain level. Sparse de-bris at plain level extends WNW and ESE. 400 mNNE is a second, 90 diam. X 1.2. Immediately Eof the second is a third, 70 diam. X 1.2. Achae-

257

oi.uchicago.edu

General Site Catalog

menian-Parthian. Probably the latter period is pre-dominant on the first, the earlier period on theother two.

670 220 WNW X 100 X 3; collection was lim-ited owing to wind-laid sand deposits on mound.600 m NE is a second, 90 diam. X 2. AdjoiningNNE of the second is a third, 100 diam. X 2.6.Early Islamic-Samarran. An underlying Sasanianoccupation seemed possible at the single mound tothe SW, while the pair to the NE probably contin-ued somewhat later than the Samarran period.

671 - 900 ENE X 300, with continuous debris atplain level and occasional small mounds rising to 2m. The presence of dunes to the E may obscure astill further extension. Probably small Uruk andCassite occupations. Mainly Parthian.

672 - 140 diam. X 3. 400 m NNW is a second,90 diam. X 3. 300 m NNW of the second is athird, 200 NNW X 150 X 2.5. The second andthird mounds have traces of Cassite, primarilyAchaemenian, some Parthian. The first mound hasAchaemenian-Parthian also but is primarily Sa-sanian.

673 - 180 N X 100 X 0.3, although the limits ofsurface sherd concentration are not well defined.Trace of Uruk. Primarily Old Babylonian, a littleMiddle Babylonian.

674 - 150 NNW X 110 X 2.2. 300 m S is a sec-ond, 100 diam. X 2. There may be a small, under-lying Cassite settlement at the N mound. Both areSasanian-Early Islamic.

675 80 diam. X 2.8, with smaller, lowermounds 80 m NNW and 30 m SSE. Mainly Achae-menian, also a little Parthian-Sasanian.

676 - Debris at plain level within an area 400diam, a mound 110 diam. X 1 standing near thecenter. A second mound is 400 m SE, 100 diam. X1.2. Sasanian.

677 104 diam. X 0.2, partly obscured by wind-blown sand and heavy vegetation. 200 m SSW is asecond area of debris, with sparse sherds at plainlevel within an area perhaps 140 NW X 80.Slightly farther S is a third, extremely small clusterof debris. Local flooding from the main drainageoutlet of the Mussayib irrigation project madethis site inaccessible after 1968 visit. Early and/orMiddle Uruk.

678 - 750 WNW X 180 X 0.2. Early/Late Uruk.679 - 350 WNW X 200, nowhere more than very

slightly above plain level. A second, about 25 m indiam., is 700 m ESE. Parthian.

680 - Low hummocks with sparse pottery, withinan area 30 diam. Late Ubaid, Early Uruk.

681 - 250 NNW X 140, rising in small semi-detached mounds on ends to 2 m ht., but mostly

1 m or less. 26 cm sq. bricks observed. Sasanian-Early Islamic.

682 - 250 N X 200 X 4.5. Sasanian-Early/Mid-dle/Late Islamic.

683 270 NNW X 120 X 4. Sasanian-Early Is-lamic, a little Samarran.

684 100 NNW X 50 X 0.4. Sasanian.685 120 diam. X 1.2. 200 m S is a second, 70

NW X 50 X 0.6. Parthian.686 - Location is approximately that of "Imam

Nowaja" in Selby, Collingwood, and Bewsher1885. 30 diam. X 6, a steep conical mound largelycomposed of fallen bricks (30 cm sq.) and mortar;the remains of plastered walls are visible nearsummit. A buttressed wall extends NE, possiblyfunctioning as a kind of control facility across anancient canal leading in the direction of Zibliyat.Early Islamic pottery is not confined to the smallmound but occurs on the surrounding plain sur-face also.

687 - 120 diam. X 4, with shallow brick-robbingpits and many fragments of large bricks and mor-tar. A larger, lower mound of the same characteralso occurs 200 m N. Sasanian, some Early Islamic.

688 - 280 E X 80 X 0.8. Parthian.689 140 NNW X 110 X 1. Achaemenian-

Parthian.690 200 ENE X 120 X 1.6. Sasanian.691 350 NW X 120 X3.5. Trace of Uruk.

Achaemenian-Parthian.692 450 NW X 140 X 6.5. Adjacent NNE is a

second mound, 180 diam. X 2. Trace of Early Dy-nastic I. There is evidence of a small Larsa-OldBabylonian-Cassite settlement along the SW sideand at the SE end of the main mound. The majoroccupation is Achaemenian-Parthian.

693 At least 300 diam. X 5, bisected by an oldcanal course from NW. Full size may be obscuredby enveloping dunes. 400 m WNW is a second,180 diam. X 3.5. Parthian.

694 - 210 diam. X 0.8. Parthian.695 140 NW X 80 X 5. There is a small, low

outlying mound immediately N. Achaemedian-Parthian.

696 - 280 diam. X 2.8. Sasanian.697 - 240 diam. X 2.6, with a smaller, lower

mound adjacent NNW. Late Islamic.698 180 diam. X 2. Sasanian.699 - Irregularly outlined, but about 800 ENE X

350. E side rises abruptly to a fairly uniform ele-vation of 4.5 m with much broken brick and mor-tar, possibly indicating a walled settlement. W,probably across an old canal bed, mounds arelower and more dispersed, but debris is continu-ous. Sasanian.

258

oi.uchicago.edu

General Site Catalog

700 Zibliyat. The prominent ruined tower known bythis name lies at the N end of the ruins of a verysubstantial urban area. As outlined on the map,the larger ruins extend irregularly over morethan 2,000 N X 800. Most of this area is ele-vated 1-2 m above plain level, and there aremany prominent mounds rising 4-6 m. Muchfallen brick and mortar, as well as some largestanding walls in place, can be observed on thesurface of these mounds; except around the towerat the N end the brick is largely or exclusively 34cm sq. and tends to be uniformly reddish fired.This suggests that some of the major summits havevery large buildings at their core. S of the towerand near the W edge of the city is a specializedglassworking area that perhaps centers on a mound70 diam. X 4.5 where there is much melted glassand slag from kilns. What remains of the toweris apparently little more than its mud-brick core,with some interior division into rooms implied bysuggestions of arched roofing in exposed profiles.There are many 30 cm sq. bricks on the slopearound its base, however, that presumably havefallen from its original facing. This appears to in-dicate that the tower was not contemporary withthe major occupation of the site (cf. Layard 1897,p. 596). Parthian debris occurs sparsely but widely,perhaps most concentrated in the SW quadrant ofthe city but not predominating even there. Themajor occupation, surely associated with the 34cm sq. bricks that are profuse in all parts of theurban area, was Sasanian. Early Islamic debris isconfined to the N end of the site, and sherds ofthis date are relatively abundant around the ruinedtower there. SW of the tower and semidetachedfrom the apparent W limits of the main urbanarea is a very small settlement that continued intoMiddle and even Late Islamic times.

Since I wrote the material above, I have beeninformed by J. N. Postgate that a radiocarbon de-termination has recently been made on reeds takenfrom the Zibliyat tower. Designated as sampleBM-1416, the measurement was 1102 ± 43 B.P.(A.D. 848). Recalibrated for natural radiocarbonvariation according to the calibration tables ofClark (1975), this is equivalent to 1070 ± 70 B.P.or ca. A.D. 880 (Postgate, pers. comm.).

701 - 190 E X 80 X 1.5, bisected by a formercanal bed. A smaller mound is adjacent NW.Sasanian-Early Islamic.

702 - 250 NW X 140 X 4. Early/Middle/LateIslamic.

703 - 280 NNW X 160 X 1.3. Sasanian.704 - 350 N X 200 X 2.4. Sasanian, also a

small Recent occupation.

705 90 NW X 50 X 0.9. Achaemenian-Parthian.

706 - 100 NW X 60 X 2. Possibly a small, un-derlying Uruk site. Primarily Cassite, although anantecedent Larsa-Old Babylonian occupation isalso possible.

707 180 NE X 80 X 2.4. Early Islamic-Samar-ran.

708 150 diam. X 2. Sasanian-Early Islamic-Samarran.

709 1,000 WNW X 160 X 1.5. There areseveral small hummocks 200 m N. Sasanian-EarlyIslamic.

710 - 180 diam X 2.8, consisting of three sum-mits surrounding a central low area. 80 m WNWis a second, 80 diam. X 1.6. At plain level 150m N of main mound is a Parthian cemetery thatis being exposed by wind erosion. The mound isSasanian-Early Islamic-Samarran.

711 - 180 diam. X 2, but with debris at plainlevel extending over a much larger area. Probablya small Uruk occupation. Mainly Sasanian.

712 - 300 WNW X 90 X 1.5. Early/Middle/Late Islamic.

713 - 140 diam. X 2.8. Ur III-Larsa.714 A very small, apparently mixed lot of sur-

face debris, perhaps thrown up in the spoil banksof a much later canal that has itself now beenvirtually obliterated by wind erosion. Trace ofUruk. Larsa-Old Babylonian. Trace of Cassite.

715 -- Low, scattered hummocks and surface de-bris within an area 180 diam. Even in the absenceof any perceptible elevation of this site, it seemsto consist at least in part of a cemetery that isbeing exposed by wind erosion. Ur III-Larsa.

716 Tufil al-Ahwal. The number on base map ad-joins the S and E mounds of a group of fouroutlining a rough rectangle whose long axis isNW. While the SW and SE sides of this rectangleare oriented approximately as indicated, however,the other two sides are slightly divergent. Nor arethe mounds of equal size. The most prominent,a steep-sided cone rising 8-10 m, is on the E. TheS mound is only slightly smaller, but the W andN mounds are only half as high and have muchsmaller volumes. Presumably all were contem-poraneous in what appears to have been a singlephase of artificial construction for each of them,although no sherds could be found that weredefinitely associated with their construction oruse. While there is a small, sparse settlementaround the foot of the E mound, there can be nocertainty that it was coeval with the mound it-self. The settlement referred to was Sasanian, andthis may be taken provisionally as the date for the

259

oi.uchicago.edu

General Site Catalog

group of four mounds. Artificial mounds of thisgeneral type occur elsewhere in southern Iraq(e.g., sites 164 and 170 in the Warka survey), anda Parthian or Sasanian dating seems reasonablefor them.

717 250 NW X 120 X 1.9, rising to summitsof this ht. at both ends rather than in center.Sasanian-Early Islamic.

718 - 260 NW X 90 X 2. Continuous debris atplain level extends 140 m SE to a second mound,130 E X 60 X 1. Parthian.

719 - 360 WNW X 140 X 5. Sasanian-Early/Middle/Late Islamic.

720 - 300 diam. X 4. Trace of Uruk. Larsa-OldBabylonian-Cassite, some Parthian.

721 - A very small, low site bounded by canalbeds and perhaps consisting only of debris sec-ondarily thrown up in spoil banks. Cassite.

722 80 NW X 20 X 0.2. Flanked on both sidesby slight depressions suggesting old canal beds,perhaps indicating that the cultural debris wassecondarily thrown up in spoil banks. Very sparseadditional debris occurs across one of these beds,to the NE. On the mound between the beds, EarlyUruk and Ur III-Larsa. To the NE are a few sherdsof Jemdet Nasr or possibly Early Dynastic date.

723 90 diam X 0.3. Sasanian.724 200 NW X 80 X 1.3. Parthian.725 300 NW X 80 X 2. Sasanian-Early Islamic.726 Tell Abfi Sarifa. See account of site and small-

scale soundings there in 1968 (Adams 1970).Parthian-Sasanian-Early/Middle Islamic.

727 130 diam. X 3. 30 m E is a second, 120diam X 2. Debris at plain level, coeval with themounds, occupies a surrounding area of at least1,000 E X 500. While fairly dense in places, how-ever, the debris does not seem to reflect a con-tinuous settlement with the mounds as a nucleus.Parthian. Three coins could be provisionally iden-tified: A.D. 106, probable Parthian, probableSasanian.

728 - 120 diam. X 1.5. 200 m SW is a second,240 NW X 100 X 1. Sasanian, possibly also EarlyIslamic.

729 - 250 E X 120 X 3, although envelopingdunes may conceal a somewhat larger size.Sasanian-Early Islamic.

730 - 1,000 NW X 600, rising to 4.5 m near Nend. There are many other mounds within thecomplex of only slightly lesser ht. The site is en-tirely enclosed and partly covered by dunes, butfrom what can be seen it appears to represent asingle, nucleated urban center. Near the S endthere is an open area that may represent a largecourt, with several brick and mortar columns that

have fallen (and perhaps been partly carriedaway). Possibly these columns originally outlineda monumental gateway; there are many smallfragments of architectural ornamentation in stuccoin this vicinity. Most of the site is Achaemenian-Parthian, although settlement at the N end contin-ued into the Sasanian period on a small scale. Asmall hoard of corroded copper coins found nearthe S end included one that can probably be as-signed to Antiochus IV (175-63 B.C.), and twothat can only be identified to the extent that theyappear Parthian.

731 180 diam. X 0.2. Achaemenian-Parthian.732 100 diam., sparse pottery at plain level.

Sasanian.733 120 diam. X 0.2. Sasanian.734 90 NNW X 70, hummocks rising to 0.2.

100 m WNW are a few additional, scattered hum-mocks of debris. The larger area is Old Baby-lonian-Cassite, while the scattered hummocksWNW are Middle-Neo-Babylonian.

735 40 diam., hummocks rising to 0.2. Sasanian.736 An enclosing ring, in places a double ring,

of sherds at plain level, 40 diam. Possibly thesherds were inclusions in mud-brick walls thathave been destroyed by wind erosion, suggestingan isolated building like a khan (caravanserai) ormilitary post. Parthian.

737 80 diam. X 0.1, with a second, 70 diam. X0.1, 300 m E. Uruk.

738 110 diam. X 1.6. Uruk. Mainly Cassite-Middle Babylonian, perhaps a minor Achaemen-ian-Parthian occupation. Sasanian sherds werenot observed on the mound but are found dis-persed on the surrounding plain.

739 Perhaps 120 diam X 0.4, but limits of siteare difficult to define. Parthian.

740 250 diam. X 0.8. Widespread but sparsesherds indicate an Early/Middle Uruk occupation.Mainly Cassite-Middle Babylonian. Also someParthian-Sasanian.

741 - 140 diam. X 1, immediately SSE of 740.Parthian.

742 220 NE X 100 X 3, with small outliers NEand E. Immediately SSW is another, 160 diam. X2.5. Early Islamic-Samarran, possibly continuingsomewhat later.

743 - 100 NW X 60 X 1, largely surrounded bydunes. 200 m SW is a second, 170 diam. X 0.7.Immediately SSE of the second mound and 100 mSSW of the first is a third, 140 diam. X 0.3, withsparse pottery. The low mound with sparse potteryis Early Uruk. The other two are Sasanian.

744 - 150 diam., plain level. Another area ofdebris 300 m NW is perhaps 120 diam. Pottery

260

oi.uchicago.edu

General Site Catalog

is relatively sparse, particularly on the latter, sothat site limits are hard to define. Early Uruk.

745 8- 0 diam. X 0.6, with a few higher hum-mocks of debris. Uruk, Cassite, very limitedParthian.

746 280 NE X 50 X 0.2, either a narrow stripof settlement along a canal or merely debristhrown up in later canal spoil banks. Akkadian-Larsa.

747 120 diam. X 0.8. Middle/Neo-Babylonian.748 90 diam. X 0.4, with a few additional

hummocks of debris NW and N. Possible LateUruk. Jemdet Nasr. Possible, but at best verylimited, Ur III-Larsa. A few late graves that canbe only provisionally assigned to the long spanbetween Middle Babylonian and Parthian.

749 240 NW X 130 X 2.2. 150 m W is a secondarea of settlement, 200 NW X 80, that is mostlybelow 0.8 but that rises in a few hummocks to1.5. A small Uruk settlement probably underliesthe W part of the site. However, both moundsare mainly Cassite.

750 - 160 E X 130 X 1.8. 150 m ENE is a second,90 diam. X 0.8. Bricks 23 cm sq. Middle Islamic.

751 80 diam X 0.3. Sasanian.752 120 diam. X 0.8. Middle Islamic, coeval

with site 750.753 250 E X 170 X 2.2. 150 m NE is a second.

80 diam. X 2. Main mound primarily Achaemen-ian but continuing into Parthian. Smaller moundis mainly Seleucid-Parthian.

754 160 E X 80 X 1.9. Sasanian.755 140 diam. X 2. Sasanian.756 Main axis of the site follows the left bank

of the Shatt al-Nil for 400 m, with a perpendiculararm extending NNE from the center for 350 m.Mostly low, but rising to 2.2 m near center. Onopposite bank is a mound 220 NW X 120 X 1.4,while farther downstream is a third mound 80diam. X 2. early Islamic, at best limited Samar-ran.

757 - 220 diam. X 0.8. Samarran, possibly con-tinuing somewhat later.

758 - 110 diam. X 2. 200 m SSE is an additionalarea of debris at plain level, ill-defined but large.The debris at plain level is primarily Neo-Baby-lonian-Achaemenian, perhaps continuing intoSeleucid-Parthian. The mound may begin inAchaemenian, but it is primarily Seleucid-Parthian.

759 - Part of Tiulul Sutail. 250 E X 180 X 3.Parthian-Sasanian.

760 Part of Tilill Sutail. 700 E X 500 X 4.5, al-though only a central mound 100 m diam. reachesthis ht. Probably also Sasanian, but mainly EarlyIslamic-Samarran.

761 Two mounds 300 m apart NNE-SSW, each about130 diam. X 3. Parthian.

762 - 280 NW X 130 X 1.5, with debris at plainlevel extending farther NW. Sasanian.

763 160 diam. x 1. 250 m SW is a second, 110diam. X 0.8. Sasanian-Early Islamic.

764 380 NW X 240. Three semidetached sum-mits, the highest of 4.5 m to the SE, occur withinthis area, while the greater part of it consists onlyof low debris. The highest mound is mainlyParthian, the others mainly Sasanian. Three in-cantation bowls were recoved together on thelatter.

765 280 NE X 190 X 1.3. See sketch map. Forfuller description see Redman 1971. Early/MiddleUruk.

100 M

765

766 120 diam. X 2.6, with an area of equalsize at 0.5 elevation forming an apron that en-closes the entire NE half of the mound's peri-meter. 150 m SW is a second mound, 170 NNWX 100 X 1.6. Achaemenian-Parthian.

767 90 diam. X 2. Sasanian-Early Islamic.768 Part of Tuilfil Sutail. 300 diam. X 6. Other, smaller

mounds tail off SW. Widely scattered sherds hintat an underlying Uruk occupation. Sasanian.

769 - 150 N X 80 X 2. Early Islamic-Samarran,Late Islamic-Recent.

770 140 diam. X 2, with low debris tailing offNW for 200 m. Ur III-Larsa-Old Babylonian.

771 140 diam. X 2.5. Middle Babylonian.772 Low hummocks of debris within an area

perhaps 80 diam., but the limits are ill defined be-cause of surrounding dunes. Cassite, Sasanian.

773 120 NW X 60 X 1.5. Immediately NWis a second, 120 diam. X 0.8. To the S, across theold Shatt al-Nil canal from the second, is a third,130 NW X 40 X 0.8. Early Islamic.

774 70 diam. X 1. Sasanian-Early Islamic.775 170 NNE X 50 X 0.6. Early Islamic-

Samarran.

261

.:·······"""'"'···················..

oi.uchicago.edu

General Site Catalog

776 - 180 NNW X 120 X 0.8. Possibly Achae-menian, mainly Seleucid-Parthian.

777 - 130 NE X 70 X 1.3. A smaller mound lies100 m SW along an old levee, and other stillsmaller hummocks of debris continue farther inthis direction. Sasanian-Early Islamic.

778 - 140 diam. X 2.3. 90 m NE is a second, 90diam. X 0.1. Two much smaller mounds areevenly spaced SW. Probably a small, underlyingUruk site. Sasanian-Early Islamic, Late Islamic-Recent.

779 - 80 diam. X 0.5. 50 m NE is a second, 40diam. X 0.3. Mainly Sasanian, some Early Is-lamic.

780 - 200 NW X 100 X 2.4. 200 m NW is asecond, 100 diam. X 0.6, with a still smallermound immediately SW. Sasanian-Early Islamic.

781 - 700 WNW X 180 X 0.2, partly dune-cov-ered. Early/Middle Uruk, Jemdet Nasr, possiblyEarly Dynastic I. Limited Akkadian. MainlyLarsa-Old Babylonian.

782 - 220 ENE X 80 X 2. A small underlyingUruk settlement. Possibly Sasanian. Early Islamic.Mainly Samarran. Some Recent.

783 - 120 diam. X 0.3. Adjacent WSW is anarea of debris at plain level, 30 m in diam. TheWSW debris is Uruk and Larsa-Old Babylonian-Cassite. The low mound is Early Islamic.

784 120 diam. X 1.6. Late Islamic.785 - 220 ENE X 80 X 2. Late Islamic.786 Scattered low hummocks within an area

80 diam. Debris is sparse except on the hummocks.Middle Uruk doubtful, mainly Late Uruk. Traceof Jemdet Nasr.

787 - 140 diam. X 3, with lower mounds ex-tending almost continuously for 220 m farther tothe NNE. Achaemenian-Parthian.

788 180 NNE X 70 X 1.8. Surface obscuredby wind-laid sand. Sasanian, probably also EarlyIslamic.

789 - 160 diam. X 0.5. Trace of Uruk. Sasanian.790 250 NE X 220 X 0.3. Intensive collection

made within a circle of 5 m radius, center. Middle/Late Uruk, the former probably preponderant.There is also a little Sasanian-Early Islamic pot-tery that may not represent a primary occupa-tion.

791 - Main mound 170 diam. X 3.100 m ENE isa second, 200 X 70 X 2.6, forming a shallow c-shaped ridge open to the NE. 400 m WSW is athird, 80 diam. X 1.8. Sasanian.

792 - 260 diam. X 0.3. Intensive collection madein circle, 5 m radius, center. Several apparent pot-tery kilns were noted in the central part of thesite. Early Uruk.

793 - 60 diam., very sparse debris at plain level.Early Uruk.

794 90 WNW X 50 X 0.5. Adjacent SE is amound 70 NW X 30 X 0.2. Trace of Cassite. Pri-marily Middle Babylonian. There are Parthiangraves on both mounds that appear to trail off SE.

795 - Plain-level debris, a few low hummockswithin an area 50 diam. Cassite.

796 Part of Tilfil Sutail. Scattered mounds withinan area 400 diam., some rising to 3 m ht. Bricks24, 25, and 29 cm sq. Early Islamic, possiblycontinuing into the Samarran period.

797 280 NW X 100 X 2.3. Trace of Uruk.Achaemenian-Parthian.

798 600 E X 450, mostly low but rising to a3.5 m summit. Mainly Sasanian, less Parthian. TheEarly Islamic occupation was largely confined tothe most elevated, central portion of the site. Thereis also a much earlier settlement on the ESE endof site, lightly overlain by later debris. Its originalsize can only be guessed at 5 ha. Probably JedmetNasr, mainly Early Dynastic I, some later EarlyDynastic.

799 120 diam. X 3. 250 m E is a second,smaller. Sasanian.

800 80 diam. X 0.8. Sasanian.801 450 NW X 250 X 5, rising to this ht. only

near W end. Mound is Neo-Babylonian-Parthian.Sherds suggesting a small Early Dynastic I occupa-tion are found in the bed of the Shatt al-Nil im-mediately SW.

802 90 diam. X 1.4. Indication of an underly-ing Early/Middle Uruk occupation. Mostly Cas-site, probably continuing into Neo-Babylonian.

803 70 diam. X 0.8. A still smaller mound lies50 m NW. Probably a small, underlying Uruk set-tlement. Some Sasanian. Mainly Late Islamic-Recent.

804 130 diam., plain level. Sparse surface pot-tery necessitated an essentially exhaustive collec-tion of diagnostic types from the entire site in orderto establish an adequate basis for dating. EarlyUruk.

805 - 70 diam. Surface pottery is sparse every-where, and only relatively more plentiful on a fewlow hummocks apparently representing kilns. Thecollection essentially exhausted the diagnostictypes to be observed on the site, save that thenumber of beveled-rim bowl sherds could havebeen at least doubled by including progressivelysmaller sherds. Late Uruk.

806 - 240 E X 140 X 3.2. Parthian.807 - 70 diam. X 2.2. Ur III-Larsa.808 - 180 diam. X 2, but with debris tailing off

NNW. Samarran, also some Recent. (A sgraffiato

262

oi.uchicago.edu

General Site Catalog

sherd from this site is illustrated in Adams 1970,fig. 16h.)

809 300 E X 180 X 0.2. Sasanian.810 40 diam. X 0.3. S across old canal from

NNW is a second, perhaps 100 diam. X 0.1. LateIslamic-Recent.

811 - 170 E X 90 X 0.8, with lower, dispersedhummocks of debris extending at least 500 mfarther E. Early Islamic-Samarran.

812 120 diam. X 0.8. Parthian.813 90 diam. X 1. 200 m E is a second, 40 diam.

X 0.2. The main mound is Samarran-Middle Is-lamic, the smaller one Samarran only.

814 140 diam. X 1. Sasanian, limited Early Is-lamic.

815 130 WNW X 70 X 1. Ur III-Larsa.816 140 diam. X 0.8. Sasanian.817 200 NW X 120 X 0.8. Achaemenian-Par-

thian.818 260 NW X 140 X 2, surmounted by sev-

eral high dunes. Debris at plain level extends NWalmost to 817, at least 500 m, but is too sparse torepresent a continuous settlement for this distance.There is an Early/Middle Uruk occupation here,although it could not be localized. Mainly Larsa-Old Babylonian-Cassite.

819 90 diam. X 1. Early Islamic-Samarrantprobably continuing a little later.

*, ' :MILL}*.

BRICK WALLS

& PAVINGS C

LARGE BUILDI

820 Sparse, low hummocks within an area 120m in diam. Middle-Neo-Babylonian.

821 70 diam. X 0.6. Small underlying Uruk set-tlement. Sasanian-Early Islamic.

822 110 N X 60 X 1.4. Cassite-Middle Baby-lonian.

823 - 240 E X 120 X 2.5. Sasanian-Early Is-lamic.

824 120 diam. X 0.4. Early Uruk, Larsa-OldBabylonian, very limited Cassite.

825 110 diam. X 2. Adjoining SSW is a second,90 diam. X 1.8 Sasanian.

826 Tell al-Arsan. This impressive site can be dividedinto two parts. To the SE is a very well pre-served circular citadel 220 m in diam. Its outerwall, built entirely of baked brick 30 cm sq., canbe followed for virtually the entire circumference.In addition, there are outer structures at intervalsthat suggest protective towers, gates, and perhapseven a moat. Within the circumvallation are nu-merous wall footings, also of 29-30 cm sq. brick.In the center is a mound about 60 m in diam. thatis superimposed on the larger circular platformenclosed by the wall, reaching a maximum eleva-tion of about 6 m. On its summit are traces of apossible rectangular building. Immediately NW ofthe larger, circular part of the site is a lower areawith numerous, well-preserved wall footings of

(LATER?) BRICK KILNS.** 826100 m

263

i

..~···"·~C

iI

, , ,,. . ,' .' .. . . .. . .

oi.uchicago.edu

General Site Catalog

29-30 cm sq. brack, suggesting that large buildingscould be planned with very superficial trenching.There are also areas with thin walls at irregularangles suggesting small private houses. See sketchplan. The small mound within the circular wall isprimarily of Uruk date, although Cassite pottery ispresent in quantities around its foot. Sparse Cas-site debris occurs elsewhere in the circular citadelas well, but not in the lower settlement to the NW.The main occupation was Parthian, possibly con-tinuing briefly into the Sasanian period. One coincan be attributed to Osroes, ca. A.D. 110. Unfor-tunately, this fine site is now isolated for much ofthe year by marshes (and residual salt flats) cre-ated by the outflow from the Mussayib main drain-age canal.

827 70 diam. X 2.3. Pottery is sparse; much ofmound consists of brick-kiln slag and cinders ofuncertain date. Probably to be assigned to theCassite-Middle/Neo-Babylonian range.

828 - 220 diam. X 4. A second, 140 diam. X 2,is 100 m SSE. Larsa-Old Babylonian-Cassite.

829 - 120 WNW X 25 X 0.1. Early/MiddleUruk.

830 Obscured by dunes, but at least 450 NNEX 250. SSW end 5 m high, NNE end lower. Achae-menian-Parthian.

831 300 diam. X 2.8. Very extensive, althoughgenerally small and fairly shallow, pitting of thesurface of this mound occurred between my 1968and 1975 visits. One effect of this pitting is toobscure almost completely the scatter of surfacesherds needed for dating. A substantial EarlyUruk settlement. Mainly Ur III-Larsa. ReducedAchaemenian-Parthian occupation.

832 350 NE X 220 X 2.6. The pitting that al-most continuously covers site 831, only 200 m SE,does not (yet) extend to this mound. Early Uruk islargely limited to SE end. Major occupation Ur III-Larsa. NE end continues into the Old Babylonianperiod.

833 Two small mounds, each about 100 diam.X 2.5. One is immediately SW of site 832, thesecond 100 m farther S. A third, about 1,000 mdirectly S of the two, is 130 NNW X 40 X 0.6.The third, smallest mound is Ur III-Larsa. Thefirst, closest to site 832, is Larsa-Old Babylonian-Cassite. The other mound is Sasanian.

834 - 250 diam. X 6. Outlying mounds 30 mWNW (110 diam. X 0.8) and 100 m SW. Muchdebris also at plain level, probably extending atleast 350 m ESE, but this is partly obscured bydrifting sand and dunes. Mainly Larsa-Old Baby-lonian-Cassite. Limited Neo-Babylonian-Achae-

menian-Parthian. Sasanian, in addition to the earlymaterial, occurs only on WNW outlier.

835 - 170 diam. X 1.2. Early/Middle Uruk, Cas-site-Middle Babylonian.

836 Tell Dauran. 180 N X 100 X 6.5. A prominent,steep-sided, flattopped landmark. Achaemenian-Parthian.

837 120 diam., plain level. Sparse surface pot-tery made necessary a fairly exhaustive collectionfrom the entire site rather than the selection of asample area. Early Uruk. Numerous Sasanian-Early Islamic sherds were assumed not to repre-sent a real occupation but to be strays from nearbysite 838.

838 200 NE X 90 X 2.8. 40 m NW is a second,180 diam. X 2.5. Small outliers are adjacent NEand 400 m SW of the first. Two other smallmounds, each about 80 diam. X2.4, form an E-Wline about 150 m E of the first. At the secondmound there were small Uruk, Cassite occupa-tions. Mainly Middle/Neo-Babylonian. The re-mainder of the group is Sasanian, with a probablyreduced Early Islamic occupation also, and some ofthis later debris naturally occurs on the earlymound as well.

839 150 NE X 80 X 2.5. Sasanian-Early Is-lamic.

840 100 diam. X 0.8. Sasanian.841 160 diam. X 4.5. Neo-Babylonian-Achae-

menian, possibly 'continuing somewhat later.842 - 750 NE X 200 X 2.8. 200 m SW is a sec-

ond mound, 280 diam. X 2.3, the elevated part ofthe mound forming a c-shaped ring open to theSW. Sasanian.

843 - 80 diam. X 2, with lower and smallermounds trailing off NW and SE along old levee.Sasanian-Early Islamic.

844 60 diam., low. Ur III-Larsa.845 320 NW X 250 X 3. Early Uruk, Larsa-

Old Babylonian-Cassite, Sasanian.846 - 160 diam. X 0.8. 300 m NNW is a second,

120 diam. X 0.6. Sasanian.847 450 diam. X 4, dune-covered. Sasanian-

Early Islamic.848 Tell Abfi Dhuwari, 950 NW X 200 X 3.5. Ad-

ditional smaller mounds, the closest 150 diam.X 2.2, and plain-level debris extend E for a con-siderable distance. As at site 1013 nearby, there isextensive surface-level architecture at this site thatcould be easily mapped from a low-level air photo-graph. Bricks in the exposed (foundation?) coursesare mainly 21 and 23 cm sq., with a few brokenand reused 25 cm bricks used as wall interior fillrather than facing. Adjoining houses with rela-tively small rooms seem to follow the central axis

264

oi.uchicago.edu

General Site Catalog

of the site for its entire NW end, perhaps hintingat a street layout. Rooms and courts are seeminglylarger and walls thicker to the SE. Nothing un-equivocally Sasanian seen. Early Islamic appar-ently uniform in distribution over the entire site,and perceptibly more common than at Nippur.This suggests a later termination than at Nippur,possibly continuing into the Samarran period.

849 30 diam. X 0.8. Sasanian-Early Islamic.850 110 E X 40 X 0.8. Probable Sasanian.

Early Islamic.851 180 E X 100 X 2. Sasanian, possibly also

Early Islamic.852 - 240 E X 130 X 1.2. Probable Sasanian.

Mainly Early Islamic, some Samarran.853 180 diam. X 2.2, with a smaller mound

immediately E. Uruk, Sasanian-Early Islamic-Samarran.

854 300 WNW X 220 X 2. 50 m ESE is asecond, 180 diam. X 1.2. Probably there is asmall, underlying Uruk settlement here. The firstmound is Cassite-Middle/Neo-Babylonian, thesecond similar but apparently without Cassite.

855 - 220 NW X 120 X 1. Possible Sasanian.Early Islamic.

856 140 E. X 70 X 0.8. Probable Sasanian.Mainly Early Islamic, some Samarran.

857 - 150 diam. X1.5. Sasanian.858 Tell Abfi 'Allma. 450 E X 200 X 5. Trace of

Uruk. Cassite-Parthian.859 - 230 NNW X 170 X 2.3. 200 m NE is a

small, contemporary outlier, 60 diam. X 0.4.Bricks 22 cm sq. Sasanian-Early Islamic-Samar-ran-probably later.

860 140 diam. X 2, with a small conical moundat SSE end rising to 4.2 m. There is a small outlier100 m WSW. Sasanian.

861 260 NNW X 170 X 1. Larsa-Old Baby-lonian-Cassite-Middle Babylonian.

862 110 NE X 30 X 0.3. 200 m NE is a sec-ond, 110 NE X 60 X 0.4. Sasanian-Early Islamic.

863 - 200 NE X 140 X 2. 200 m farther NE isa second, 300 NE X 90 X 1. Immediately N ofsecond is a third, 90 diam. X 1. Sasanian-EarlyIslamic.

864 - 90 diam. X 0.8. Sasanian.865 - 160 NW X 100 X 2.2. Some Cassite,

mainly Neo-Babylonian-Achaemenian.866 - 130 diam. X 2. Rare Parthian. Mainly Sa-

sanian.867 - 90 diam. X 4.5. Neo-Babylonian-Achae-

menian-Parthian.868 - 150 diam. X 1.8, with low hummocks of

debris continuing SW along old levee. Sasanian-Early Islamic.

869 180 diam. X 4. Rare Larsa. Mainly OldBabylonian-Cassite. Rare Neo-Babylonian.

870 - 150 ENE X 120 X 2. Parthian.871 Ishan Abfi Basfir al-Sharqi. 250 diam. X 5.5.

Akkadian-Larsa, Achaemenian-Parthian, sparseSasanian.

872 - 120 diam. X 2.8, but with a wide, low stripof debris extending 350 m NE from main mound.Achaemenian-Parthian. Sasanian pottery is local-ized on NE end of low strip.

873 - 180 diam. X 5. Parthian-Sasanian-EarlyIslamic.

874 200 diam. X 4.5. Parthian-Sasanian-EarlyIslamic.

875 Ishan Abf Basfir al-Gharbi. 500 NW X 350 X6. Rare Cassite-Middle/Neo-Babylonian. MainlyAchaemenian-Parthian.

876 220 WNW X 120 X 2.5. Trace of Uruk.Sasanian.

877 Tell Abfi Dhaba'. From Inspectorate of Surveysrecords, file 54, register 1616. Recorded as Neo-Babylonian-Parthian.

878 Tell Abui Shejayr. From Inspectorate of Surveysrecords, file 10, register 1465. Recorded as Neo-Babylonian-Parthian, Islamic.

879 Tell Abf Hamis. From Inspectorate of Surveysrecords, file 10, register 1465. Recorded as Neo-Babylonian-Parthian, Islamic.

880 Tell Abi G6da. From Inspectorate of Surveysrecords, file 35, register 1587. Recorded as Neo-Babylonian-Parthian.

881 Tell al-Dubaysiya. From Inspectorate of Surveysrecords, file 29, register 1587. Recorded as Islamic.

882 Tell Bismaya. 680 E X 400 X 10. Seleucid-Parthian-Sasanian.

883 Tell Tuwaym. From Inspectorate of Surveys rec-ords, file 47, register 1587. Recorded as Parthian.

884 Tell Abi Kelb. From Inspectorate of Surveys rec-ords, file 40, register 1587. Recorded as Neo-Babylonian-Parthian.

885 Ishan Abu Shfira. From Inspectorate of Surveysrecords, file 30, register 1587. Recorded as Islamic.

886 Tell al-Jela'a. From Inspectorate of Surveys rec-ords, file 294, register 440. Recorded as Islamic.

887 Tell al-Baghal. From Inspectorate of Surveys rec-ords, file 291, register 440. No dating assessmentavailable.

888 Tell al-Haytaniya. From Inspectorate of Surveysrecords, file 286, register 440. Recorded as Islamic.

889 Tell Khesayn al-Kabir. From Inspectorate of Sur-veys records, file 289, register 440. Recorded asSasanian.

890 Tell Mayyid. 150 diam. X 2. Sasanian-EarlyIslamic.

265

oi.uchicago.edu

General Site Catalog

891 Tell Abfi Milleh. 160 diam. X 2.4. Seleucid-Parthian.

892 Tell Krinwis. 60 diam. X 2.8. Small hummocksof debris extended NW and SE from moundalong a broad, old levee. Late Islamic.

893 Tell Khesayn al-Seghir From Inspectorate of Sur-veys records, file 280, register 440. Recorded asIslamic.

894 Tell Abui Dhahab. From Inspectorate of Surveysrecords, file 46, register 1587. Recorded asParthian.

895 Tell Bukhera'. From Inspectorate of Surveys rec-ords, file 45, register 1587. No dating assessmentavailable.

896 Tell Abti Skhayr. From Inspectorate of Surveysrecords, file 44, register 1587. Recorded asParthian.

897 Tell al-Dhuba'i. From Inspectorate of Surveysrecords, file 42, register 1587. Recorded asParthian.

898 Tell al-Ikhaywan. From Inspectorate of Surveysrecords, file 284, register 1587. Recorded asAkkadian-Ur III, Parthian. (N.b. This datingshould be treated with great reserve. Early datingsalso were assigned in these records to sites 878,882, and 891, none of which was found to havebeen occupied earlier than the Neo-Babylonianperiod.)

899 Tell Mugharat al-Gharbi. From Inspectorate ofSurveys records, file 37, register 1587. Recordedas Neo-Babylonian-Parthian.

900 Tell Mugharat al-Sharqi. From Inspectorate ofSurveys records, file 49, register 1587. Recordedas Neo-Babylonian-Parthian, Islamic.

901 Tell Abti Khay, 750 NW X 350 X 9. Neo-Baby-lonian-Achaemenian-Parthian.

902 200 NW X 160 X 2. Parthian-Sasanian.903 100 E X 70 X 2.2, with a small outlier ad-

jacent to NW end. Sasanian-Early Islamic, someSamarran.

904 280 NNW X 160 X 2.3. Sasanian.905 300 NNW X 140 X 2.5. Sasanian-Early

Islamic-possible Samarran, Recent.906 200 diam. X 3, but with additional debris

at plain level extending NW for 300 m. Sasanian,Late Islamic. The latter occupation is limited tothe mound.

907 - 250 NE X 160 X 2.4. Immediately NNWis a second, 160 diam. X 2. Achaemenian-Parth-ian.

908 - 220 NW X 140 X 2, additional debris tail-ing off SE for 300 m. Sasanian.

909 - 50 diam. X 1.2, the character of the debrissuggesting an isolated building rather than a set-tlement. Sasanian.

910 - 140 diam. X 2. Sasanian-Early Islamic.911 Part of Tflfil Sutail. 400 N X 100, maximum

elevation 6 m at N end and with a low saddlein middle. Sherds also extend widely to E at plainlevel. Mainly Sasanian-Early Islamic, but with asmall Samarran occupation also.

912 130 diam. X 0.2. Debris may continue Sunder adjoining dunes, for 150 m S is a sparsecluster of sherds within an area 80 m in diam. Asimilar cluster, in this case surely detached, occurs200 m farther SSW. Early Uruk.

913 - 80 diam. X 2. 100 m W is a second, 140diam. X 0.8. A third, 200 m SSW and SSE of thefirst and second respectively, is 180 diam. X 1. Ad-ditional debris tails off farther SSE. Sasanian.

914 - 400 NNW X 300, mostly 1 m or less butrising to 2 m at SSE end. Achaemenian-Parthian-Sasanian.

915 90 diam. X 1.8. Sasanian-Early Islamic.916 400 NW X 80 X 2.8, with smaller, lower

clusters of debris extending farther NW. Sasanian-Early Islamic, trace of Samarran.

917 80 E X 50 X 1.2. Crust of white deliques-cent salts and associated soil conditions severelylimit pottery collection and identification. Proba-bly in the Neo-Babylonian-Achaemenian-Parth-ian range, but cannot be specified more closely.

918 110 NW X 50 X 2. A second, muchsmaller mound lies about 100 m WNW. Very sa-line; sparse pottery. Sasanian, possibly also EarlyIslamic.

919 - 130 WNW X 50 X 0.7. Very saline; sparsepottery. Sasanian, probably also Early Islamic.

920 Ishan Nabi Yunis. From Inspectorate of Surveysrecords, file 356, register 2792. Recorded asIslamic.

921 140 diam. X 0.3. Pottery sparse. Parthian.922 140 E X 110 X 2. Sasanian-Early Islamic,

trace of Samarran.923 2 mounds 80 diam. X 2.4, about 100 m

apart NW-SE. Sasanian-Early Islamic.924 Tell Abu Khorfan. 320 diam. X 3.5. Trace of

Early Dynastic I. Sasanian-Early Islamic.925 - 550 NW X 100, rising from 0.8 SE to 2.8

NW. Larger, low part of site Sasanian. High NWend includes small Early Islamic and Recent occu-pations.

926 - 350 NW X 180 X 1. Sasanian.927 - 200 NNW X 120 X 2.2. Sasanian.928 Ishan Sbayj. From Inspectorate of Surveys rec-

ords, file 334, register 2792. Recorded as Ubaid,Sasanian. (The Ubaid designation probably indi-cates the presence of clay sickles. In view of thepaucity of Ubaid sites, an Uruk designation ismuch more likely.)

266

oi.uchicago.edu

General Site Catalog

929 Ishan al-Jarna. From Inspectorate of Surveysrecords, file 335, register 2792. No provisionaldating recorded.

930 and 931 Through some confusion, three sitenames are identified with these two locations:Ishan al Nuwawis, Tuilul al-Wardiya, and Abf al-Milleh. From Inspectorate of Surveys records.All are in file 338, register 2792. Recorded asParthian.

932 120 diam. X 1.8. Cassite-Parthian.933 Main mound 320 N X 180 X 4. 300 m W

is a second, 100 diam. X 1. Trace of Uruk, possibleAkkadian ribbed ware on smaller mound. Bothare mainly Larsa-Old Babylonian.

934 140 diam. X 0.2. Sasanian.935 Sparse sherds at plain level within an area

90 m in diam. Early Uruk.936 80 diam., plain level. 100 m SE is a site 40

diam., plain level. SE site Early Uruk, NW siteRecent.

937 Sparse pottery at plain level within an area80 m in diam. Uruk.

938 100 diam. X 1.8, with low debris extend-ing 80 m NW, Sasanian-Early Islamic.

939 200 N X 120 X 0.6. Numerous hummocksconsist in the main of ancient kilns, some withfused pottery still in place. Kiln wasters of fusedclay sickles, as well as numerous complete speci-ments and hundreds of fragments, were particu-larly prominent. Early Uruk.

940 - 220 diam. X 0.3. 60 m E is a secondmound, 80 diam. X 0.1, constituting an appar-ently separate nucleation, although the debris wasrelatively sparse. The remains of at least one pot-tery kiln are visible near the center of the largermound. Main mound primarily Early/MiddleUruk. Late Uruk largely limited to SW quadrant.Intensive collection #1 was made in circular area5 m in radius near center of it. Collection #2 wasmade somewhat NE of apparent center of smaller,entirely Late Uruk mound. There is a little Sasa-nian-Early Islamic pottery on the smaller site,possibly not reflecting a real occupation.

941 220 ENE X 140 X 2.4. Mud-brick pavingor tomb roofing exposed by wind erosion. Sasa-nian.

942 - 220 WNW X 90 X 2.4. A smaller, lowertell lies 100 m NE. There are brick-covered tombson the main mound that have been partly exposedby wind erosion. The main mound is Parthian,possibly also some Sasanian. The smaller moundappears to be only Sasanian.

943 - Plain level, perhaps 180 diam., but pot-tery sparse, limits diffuse. Parthian.

944 - 250 diam., plain level. Immediately NNE

is a second large, low area of same date, only acentral nubbin reaching 1 m ht. More scatteredoutcrops of debris occur also to the S. Trace ofUruk. Akkadian-Larsa. The scattered debris tothe S is Parthian.

945 25 diam., plain level. Jemdet Nasr.946 An almost continuous strip of debris ex-

tending E-W for 1,200 m. Width varies from 80 to120 m. Most prominent elevation, at W end, is alow mound 120 E X 90 X 0.8. Smaller outcropsof sparse debris continue at intervals still fartherto W. Trace of Uruk. Parthian-Sasanian.

947 - 180 diam. X 3.3. 100 m SSW is a smallermound, 160 NNE X 30 X 0.6. Probably a smallUruk settlement, confined to the smaller mound.Both are mainly Larsa-Old Babylonian-Cassite-Middle Babylonian.

948 - 260 diam. X 2.8, surmounted by dunes.Sasanian.

949 260 NNE X 150 X 0.5. Late Islamic.950 - 120 NNW X 20 X 0.4, although debris at

plain level west of low mound may double itswidth. Possible Jemdet Nasr. Early Dynastic I.

951 - Sparse debris, scattered hummocks withinperhaps 100 diam. Possibly a small early settle-ment, of Uruk or Jemdet Nasr date. Trace of Ak-kadian. Ur III-Larsa. Recent.

952 80 NNE X 30 X 0.3. This site is believedto have been obliterated during large-scale land-leveling operations in 1975. Early Uruk, JemdetNasr.

953 Low hummocks and debris at plain levelwithin an area 110 N X 20. Possible Jemdet Nasr.Early Dynastic I.

954 - 160 N X 90 X 0.2, consisting mainly of ill-defined clusters of debris at surface level. Traces ofUruk, Cassite. Middle/Neo-Babylonian, mainlyParthian.

955 180 diam., boundaries somewhat indeter-minate. A series of scattered, low, brick-coveredmounds, the highest 2.6 m. The shape and contoursof the individual mounds suggest separate build-ings. Bricks 18-23 cm sq. Sasanian-Early Islamic.The small bricks suggest a somewhat longer con-tinuation of the site than can be confirmed fromthe ceramic evidence.

956 - 120 diam. X 0.4. Achaemenian-Parthian.957 - 180 diam. X 1.4. Sasanian-Early Islamic,

Recent. Probably associated with the earlier occu-pation is the broken-off base of a plainware jarcontaining traces of an incantation.

958 - Three small hillocks of debris. Parthian.959 - 120 diam. X 0.3, partly dune-covered.

Trace of Uruk and Cassite, mainly Middle Baby-lonian-Achaemenian, a few Parthian graves.

267

oi.uchicago.edu

General Site Catalog

960 260 diam. X 1.6. Trace of Uruk. Larsa-Cassite.

961 - 100 diam X0.3, bisected by old canal bedfrom NW. Trace of Uruk. Neo-Babylonian-Achae-menian-Parthian.

962 - 140 NW X 80 X 1, with debris continuingSE under dunes. Sasanian.

963 300 NW X 120 X 3. Sasanian-Early Is-lamic.

964 150 NW X 50 X 0.1. This site is believedto have been obliterated during large-scale land-leveling operations in 1975. Early/Middle Uruk.

965 - Size and ht. obscured by dunes but about150 diam., low. Sasanian.

966 140 diam. X 0. 7. Sasanian.967 120 NE X 90 X 3. Achaemenian-Parthian.968 Central mound 150 diam. X 1.2. Immedi-

ately NW is a second, 100 NW X 60 X 1. 150 mESE is a third, 120 NE X 90 X 1.6. MainlySasanian. Some Early Islamic on first and thirdmounds.

969 - 180 NE X 150 X 1, although overlyingdunes may obscure full size and ht. Sasanian-EarlyIslamic.

970 At least 160 diam. X 1, but overlyingdunes may obscure larger size. Primarily Ur III-Larsa; also Parthian.

971 140 N X 50 X 2.1. Cassite.972 130 N X 40 X 2.2. 40 m NNE is a second,

100 diam. X 1.1. A third, 40 diam. X 0.8, is im-mediately WSW of the second. Additional lowdebris extends NNE and SSW from the group.Mainly Cassite, but a few examples of Parthianand/or Sasanian glazed wares also were seen.

973 90 NNW X 20 X 0.4. Sasanian.974 160 diam. X 2, with a small conical mound

at SE end rising to 5 m. SE across a short gap inlevee is a second site, 150 ESE X 60, with a 4 mconical mound on its W end. These two conicalmounds may have been mud-brick forts protectingthe ends of a dam across the break in old canallevee between the two mounds, and hummocksin the bed of the break may be vestiges of the damitself. Some Parthian on the NW mound, but bothare mainly Late Islamic-Recent.

975 - 360 NNE X 160 X 0.2. 50 m ENE is anucleated area of debris at plain level, 80 m indiam. About 600 m SW is an area of similar size,also without perceptible elevation except for afew low hummocks. The main site and its ENEoutlier are Early Uruk, with at most a very limitedMiddle Uruk continuation. The intensive collec-tion was made within a circle of 5 m radius E ofthe center of the main site. The SW site is exclu-sively Late Uruk.

976 Low hummocks within an area roughly 60diam., although somewhat elongated SE. Early/Middle Uruk, Middle/Neo-Babylonian.

977 - Low hummocks, debris at plain level withinan area 80 diam. 30 m E across a later canal bedis a second, 120 N X 15-20, sparse pottery atplain level. The W part is Early Uruk, the E partLate Uruk-Jemdet Nasr.

978 150 N X 110 X 0.4. Sasanian.979 Few small hummocks, sparse debris at

plain level, perhaps 90 diam. Possible Late Uruk.Jemdet Nasr.

980 90 diam. X 0.1. Early/Middle Uruk.981 160 diam. X 0.2, although debris is sparse

and there is no perceptible elevation in the centralpart of site. 300 m NW is a second, 80 NW X50 X 0.1. Early Uruk.

982 - Low hummocks, sparse pottery, 30 diam.Early/Middle Uruk.

983 Tfiulul Werrish. An irregular chain of mounds ex-ceeding 1 km in length. Surrounded and some-what overlain by dunes, it may or may not con-stitute a single settlement. The main summit atthe N is 170 diam. X 4.2. The second, 100 mS, is 130 diam. X 2. Immediately SSW of thatis a third, 160 diam. X 2.5. Lower mounds con-tinue S and SSW under dunes. Main mound at N,Ur III-Larsa, Parthian, very limited Early Islamic.Remainder of site Sasanian, except that a smallmound well down the chain to the S, partly ob-scured by dunes, is Neo-Babylonian-Parthian.

984 180 diam. X 1.2, but girdled and partlycovered with dunes and hence possibly larger.Parthian-Sasanian.

985 Tell Abfi Jawan. Main mound 300 ENE X220 X 7, but lower mounds adjacent to ends in-crease length of complex to 500 m. Parthian-Sasanian.

986 180 diam. X 4.5. Sasanian-Early Islamic.987 100 NW X 70 X 1.4. A second, lower but

of about the same size, is adjacent ESE. Neo-Babylonian-Achaemenian-Parthian.

988 200 NE X 140 X 2, bisected by old canalbed from NW. 80 m SW is a second, 110 diam. X2.2. A third, much smaller, is 120 m W of first.Sasanian.

989 Adjacent settlements on opposite banks ofa former canal. To the N is a mound 350 WNW X80 X 2.2, with four closely spaced summits. Atthe W end of settlement S of canal is a detachedmound 120 diam. X 3. Farther E is a mound 250WNW X 90 X 2.6, with scattered debris andsmall mounds continuing farther ESE. The Wend of the mound to the N of the canal is Neo-Babylonian-Achaemenian-Parthian, while the E

268

oi.uchicago.edu

General Site Catalog

end is mainly Sasanian. The detached mound Sof canal is Cassite-Parthian, while the mound Eof it is Parthian-Sasanian.

990 200 NW X 170 X 1.5. A second, some-what smaller mound is immediately WNW.Sasanian.

991 - 110 diam. X 2. 200 m SSE is a second, 100N X 50 X 0.3. The main site is Neo-Babylonian-Achaemenian, possibly continuing into Parthian.The small mound is Recent.

992 - 70 WNW X 40 X 1.5. Sasanian, possiblyalso Early Islamic.

993 80 diam. X 0.2. Cassite.994 80 diam. X 0.2. Cassite.995 110 NNW X 40 X 0.6. Parthian.996 Tell al-Thien. 170 NW X 140 X 2.5.

Saline; limited collection. Possibly Akkadian.Mainly Ur III-Larsa.

997 100 NE X 70 X 0.8, SW end much re-duced in width. Ur III-Larsa.

998 Tell Tine. A small, high, conical mound, prob-ably the remains of a qal'a, is adjacent to theWSW edge of a mound 180 diam. X 2.5. Thelarger mound is Achaemenian-Parthian. The qal'aand its immediate surroundings are Late Islamic-Recent.

999 170 WNW X 140 X 4.2. Cassite-Parthian.1000 380 NW X 240 X 4. Saline surface, sparse

pottery. Larsa-Old Babylonian-Cassite.1001 Tell Drehem, ancient Puzrish- (or Sellush-) Dagan.

560 NW X 275 X 8.5, reaching this ht. onlyin a small eminence suggestive of a zigguratnear the SE end of site. Most of the area is lessthan 2 m in elevation. Surface very saline andspongy, sharply limiting surface collections, butextended search concentrating on spoil banksaround old excavations or pits produced an ade-quate sample. Entire collection was consistentwith an occupation limited to the Ur III-Larsaperiod.

1002 - 510 NW X 180 X 4.2, although only themiddle of three mounds in line constituting thissite reaches this full width. Trace of Uruk. Neo-Babylonian-Achaemenian-Parthian, very limitedLate Islamic-Recent.

1003 - 220 diam. X 2.2, bisected by an old canalbed from NW. Trace of Ur III-Larsa. Cassite-Parthian.

1004 Qal'a al-Ghanam. Square mud-brick enclosure,sides 60 m long and oriented 050° and 320°.Towers at four corners. 250 m WSW is a conicalmound 30 diam. X 2, probably a much less wellpreserved qal'a. Late Islamic-Recent.

1005 Tell al-'Arris. 150 diam. X 4. 50 m E is a second,180 E X 80 X 2. There is an imam immediately

N, still built of reed mats in 1968 but replacedby a brick building soon afterward. Trace ofUruk. Cassite-Parthian.

1006 - 140 NW X 80 X 1.5. Pottery sparse.Cassite.

1007 90 NW X 30 X 1.8. Old Babylonian-Cassite.

1008 100 diam. X 0.4. Akkadian-Larsa.1009 100 ENE X 40 X 0.4. Possibly Akkadian,

mainly Ur III-Larsa.1010 - 110 diam. X 1. 100 m SW is a second, 240

E X 110 X 0.7. Sasanian.1011 80 diam. X 0.2. 100 m SE is a second, also

small. Mainly Parthian, some Sasanian.1012 80 E X 50 X 0.4. Sasanian.1013 280 E X 220 X 5. A small outlier is ad-

jacent to SW end and other small, low moundsextend 200 m NE to right bank of Shatt al-Nil.Well-preserved architectural traces of what maybe a single building of 23 cm sq. brick were photo-graphed by kite, permitting much of the plan tobe traced. The building, of which the S end isomitted in the plan, is about 85 N X 40. Wall hts.are preserved to at least 0.5 m, although this maybe mainly foundation courses with walls them-selves of mud brick. Sasanian-Early Islamic, someSamarran.

50 M

1013

1014 - 240 NW X 150 X 4, with four separatesummits within this area. A small fifth mound is200 m NE. Probably Sasanian. Early Islamic, alittle Samarran.

1015 - 200 diam. X 6. Smaller mounds and sur-face debris extend 250 m S to a second, 220 E X80 X 4.5. 50 m ENE of first mound is a third, 200NW X 160 X 2.5. Other, smaller mounds lie 150m WNW and 300 m ESE. Many walls observedon summit of main mound, of 24-27 cm sq. brick.A complete blue-glazed jar was found with rimprotruding just above surface of main mound.Both the jar and its incipient exposure beneatheroding topmost level were exactly paralleled atsite 1213. (This drawing was previously publishedin Adams 1970, pl. 3, fig. 7d, under an earliersurvey field number, now superseded.) ProbablySasanian, mainly Early Islamic, some Samarran.

269

oi.uchicago.edu

General Site Catalog

1016 180 E X 100 X 1.4. There is also a smalloutlier 150 m W. Sasanian-Early Islamic.

1017 70 diam. X 5, consisting in large part ofthe ruins of a single building apparently con-structed of 21 cm brick. Low debris with sparsepottery extends 120 m WSW, and there is also asmall outcrop 100 m SSE. Early Islamic. The sizeof the bricks suggests that the solitary buildingmay be later than the abandonment of the under-lying mound, perhaps of Middle Islamic date, butno pottery was observed attesting to an occupa-tion of the site during that period.

1018 90 diam. X 0.5. Sasanian.1019 170 diam. X 0.8. Early Uruk, Ur III-Larsa.1020 250 diam. X 0.8. Intensive collection #1

from a circle of 5 m radius, N central, #2 fromsame, S central (about 100 m S of #1). 150 m Sis a second mound, 140 diam. X 0.1. Mainly EarlyUruk. A localized continuation into Middle orLate Uruk times is confined to the S part of themain mound.

1021 160 diam. X 0.4. Uruk, mainly Cassite.1022 70 diam. X 1.2. Probably Sasanian, mainly

Early Islamic.1023 - 120 diam. X 2.6. A second mound, 120

diam. X 2, is 80 m S with a low ribbon of debrisarcing first SE and then SW to connect them.Possibly Sasanian. Early Islamic, some Samarran.

1024 140 N X 90 X 2.4. Early Uruk, Ur III-Larsa.

1025 450 WNW X 180 X 0.8. Trace of Cassite.Middle Babylonian. Probably mainly Neo-Baby-lonian. Achaemenian.

1026 - 280 diam. X 4. Traces of walls on sum-mit, 26 cm sq. brick. Sasanian-Early Islamic, alittle Samarran.

1027 - 160 diam. X 2.2. 20 m SW across a branchor alternate bed of the Shatt al-Nil, the site con-tinues for an area 90 diam. X 0.2. Eroded mud-brick structures, possibly tombs, rise slightly abovethe surface of the main mound. Probably a small,underlying Uruk occupation. Probably someLarsa-Old Babylonian. Mainly Cassite.

1028 - 240 NNW X 170 X 2.2. Sasanian, someEarly Islamic.

1029 - 80 diam. X 0.7. Sasanian.1030 - 180 diam. X 0.4, on left bank of Shatt al-

Nil. Opposite is a mound 300 NW, from 40 to110 m in width and 2.5 m high. There is much26 cm sq. brick. Early Islamic.

1031 - 140 NW X 90 X 3. Lower debris tails offSE. Traces of Late Uruk, Early Dynastic. MainlyUr III-Larsa. The debris to the SE is Early Islamic.

1032 - 770 WNW X 380, consisting of large, lowhummocks and much debris on intervening plain

surface. Later canal branches cross the site at in-tervals, with debris also found on their low spoilbanks. Early/Middle Uruk settlement probablylimited to SE end of site, as was the Early DynasticI occupation. The Jemdet Nasr occupation seemsto have been much more extensive, but the dom-inant component in surface collections is Akka-dian-Larsa.

1033 90 diam. X 0.2. Sasanian.1034 180 WNW X 150 X 1.3. 100 m SSW is

a second, 160 diam. X 0.9. A third, 75 NNW X20 X 0.2, is about 400 m WSW of second. Appar-ently there is an underlying Early/Middle Uruksettlement at the second mound. The second andthird mounds are Cassite-Middle Babylonian,while the first mound may be Cassite only.

1035 110 diam. X 2.5. Bricks 23-27 cm sq. Onthe opposite, left bank of the Shatt al-Nil, 200 mabove this mound, is a second, 40 diam. X 1.Bricks on the latter are 21 cm sq. and pottery isvery sparse. The first mound is Early Islamic.The second is almost certainly later but cannotbe assigned a definite date from the very limitedceramic evidence.

1036 Four mounds forming an E-W line 350 X100 X 1.9. Immediately S is another, 120 diam. X2. There may be small Jemdat Nasr and Ur III-Larsa occupations on the NW end of site, or thesherds suggesting this may be merely strays fromadjacent site 1032. In any case, mainly Sasanian-Early Islamic.

1037 160 WNW X 120 X 2.2. Immediately Sis a second, 140 WNW X 90 X 1.6. Trace ofCassite. Neo-Babylonian-Achaemenian-Parthian.

1038 140 diam. X 1.2. Probably Sasanian. EarlyIslamic.

1039 Tell Hindi, 350 NW X 200 X 4.5, with smaller,lower mounds adjoining NE that give the complexthe shape of a T. Mostly or only Sasanian onoutlying parts of site, some Early Islamic on mainmound.

1040 60 diam. X 0.3. Sasanian.1041 90 diam. X 2. Sasanian.1042 220 WNW X 160 X 2.4. 300 m ENE is a

second, 180 WNW X 100 X 0.8. Probably Sasan-ian, particularly at the second mound, but bothare mainly Early Islamic.

1043 - 150 ENE X 80 X 0.8. At WSW end,slightly detached from site, is a brick pile 10 diam.X 1.5 (bricks 20-21 cm sq.). Possible Sasanian.Early Islamic. The brick pile may be the ruin of asolitary, somewhat later building.

1044 - 280 NE X 160 X 2, debris tailing off SW.Limited Uruk. Neo-Babylonian-Achaemenian-Parthian.

270

oi.uchicago.edu

General Site Catalog

1045 140 NW X 110 X 2.4. Probably Sasanian.Early Islamic.

1046 - 330 NE X 260 X 0.1, but with well-de-fined boundaries and dense debris in spite of lowht. Early Uruk.

1047 200 NNE X 90 X 1. 40 m S is a second,120 diam. X 0.6. A third is 150 m W of second,140 N X 90 X 0.4. First two Sasanian, thirdSasanian-Early Islamic.

1048 110 diam. X 1. Sasanian.1049 Two adjacent sites 250 NE X 70 X 0.2,

separated by a wide later canal. Possibly theywere originally joined, with the canal having beencut through to divide them. Probably a small set-tlement of Jemdet Nasr or Early Dynastic date.Mainly Akkadian-Larsa.

1050 80 diam. X 0.6. Trace of Uruk. Achaemen-ian-Parthian.

1051 - 170 NW X 130 X 2. Probably Sasanian.Early Islamic.

1052 90 diam. X 1. Sasanian-Early Islamic.1053 100 NW X 30, reaching 2.5 m ht. only in

a prominent brick pile (23 cm sq.) near SE end.Early Islamic-Samarran.

1054 - 100 N X 60 X 0.8. Perhaps a small Urukoccupation. Larsa-Old Babylonian-Cassite.

1055 250 NW X 160 X 1.5. Ur III-Larsa.1056 900 N X 250-300, decreasing to 100-150

near S end. Low except near center, where it risesto 2.4 m ht. Early Dynastic I, and possible JemdetNasr, are confined to central part of site. MainlyAkkadian-Larsa.

1057 300 NE X 250, with many low summitsrising to 1 m or less. Sasanian.

1058 420 NW X 100 X 3. Band of debris nar-rows and elevation declines toward the SE. SEend Sasanian only, NW end Sasanian-Early Is-lamic.

1059 230 NW X 160 X 0.3. Mainly JemdetNasr. A little Akkadian-Larsa.

1060 90 diam. X 0.8. Sasanian.1061 90 diam., sparse pottery. A few brick piles

(20-22 cm sq. brick) along the Shatt al-Nil riseto 2 m. Early Islamic. The small bricks may re-flect isolated, later buildings.

1062 - 180 NNE X 90 X 0.4. Sasanian.1063 - 220 N X 70, scattered low hummocks and

continuous debris. Probably Sasanian. Early Is-lamic.

1064 - 130 NW X 80 X 1.4. Trace of Uruk.Achaemenian-Parthian. One overfired or refiredbrick has a very faint stamp impression. Accord-ing to Douglas Kennedy, the few legible signs sug-gest that it is Neo-Babylonian.

1065 - 300 NNW X 180 X 1.6. 50 m WNW of

the NNW end of the site is a second, 100 WNWX 50 X 0.8. Several mud-brick pavings and out-lines of buildings, also many 34 cm sq. bakedbricks on main mound. Sasanian.

1066 - 150 N X 130 X 1.8. Sasanian. MainlyEarly Islamic.

1067 - 220 E X 150 X 2. Probably a small Urukoccupation. Mainly Neo-Babylonian-Achaemen-ian-Parthian.

1068 160 diam. X 2, with small hummocks tail-ing off farther S. Sasanian-Early Islamic.

1069 - 270 E X 140 X 1.9, with the E end lowerand slightly detached. Small Uruk occupation.Mainly Larsa-Old Babylonian. Less Cassite.

1070 - 250 diam. X 0.8, with remains of severalkilns forming higher hummocks. Small Uruk oc-cupation. Neo-Babylonian-Achaemenian-Parth-ian.

1071 - 460 WNW X 300 X 1.5, surmounted bya high ridge of dunes. Small Uruk occupation.Mainly Ur III-Old Babylonian.

1072 See sketch map showing loci of three in-tensive collections within circles of 5 m radius.Maximum elevation coincides with area of coneconcentration shown on map immediately SE ofcollection #1, and is about 40 cm. A modern ma-chine-dug canal trench was in the process of beingextended through the mound here in 1975, al-though still at a very superficial level. Cautiousprobing below banks of cones that appeared mostlikely to be in situ failed to disclose any thatdefinitely continued below the present surface.Collection #2 was made in the vicinity of an ap-parent Middle or Late Uruk pottery kiln. The mainmound, on separate components of which all threeintensive collections were made, is Early/Middle/Late Uruk. The ENE mound apparently is Early/Middle Uruk only. There is a very light sprinklingof later pottery, including Sasanian-Early Islamicand Recent, but hardly enough to represent sig-nificant occupations. Weathered shell collected onmain mound has been identified by Stephen F.Lintner as Gastropoda Prosobranchia Mesogas-tropoda Strombidae Strombus (Conomurex) de-corvus (RCding), "a marine species common to the'Arab' or 'Persian' Gulf" (pers. comm., 7 Septem-ber 1976). Also collected on the main mound werefive obsidian blade segments that have subjectedto neutron activation analysis by A. Colin Renfrewand John Dixon. Four of them are reported togive analyses typical for group 4C, correspondingto the obsidian source at Nemrut Dag on thenorthwest side of Lake Van. The fifth correspondsto group ig, still not precisely located but prob-ably some distance to the west or southwest of

27:

oi.uchicago.edu

General Site Catalog

Lake Van (Renfrew, pers. comm., 10 July 1978;cf. Renfrew and Dixon 1976).

1072

0. .

CONES. 100m

KILN ........ I1073 Imam Khudhr. 160 diam. X 2.2, surmounted by

a domed shrine 8 m square. Sasanian, Recent.1074 80 NNW X 40 X 0.4. Seleucid-Parthian.1075 Umm al-Tus. 120 diam. X 2.6. 500 m SE is

a second, 80 diam. X 0.6. 100 m SE of secondare the ruins of a qal'a. The second mound andthe nearby qal'a are Old Babylonian-Cassite, Re-cent. The first has little Cassite, is mainly Middle-Neo-Babylonian.

1076 220 NNE X 90 X 1.6. 100 m NW is asecond, 90 diam. X 1.6. Sasanian.

1077 100 diam. X 0.8. Middle/Neo-Babylonian.1078 230 NE X 160, rising to 2.4 m ht. only at

SW end. Sasanian-Early Islamic.1079 An irregular group of seven or eight

mounds falling in an area 900 NNE X 200, all1.8 m ht. or less. Mainly Sasanian. Some EarlyIslamic, Recent.

1080 - 280 diam. X 1.8. Middle/Neo-Babylonian,Sasanian-Early Islamic.

1081 220 NW X 140 X 2.2. Achaemenian-Parthian.

1082 Tell Abyadh. 180 diam. X 5. Small, low outlier120 m NE. Mainly Cassite-Achaemenian. LimitedSasanian.

1083 - 140 diam. X 3, with a smaller mound ad-jacent SE. Cassite-Achaemenian, limited Sasanian.

1084 Tell al-Howaylah. 200 diam. X 2. Sasanian-possible Early Islamic.

1085 - 170 diam. X 0.7, with a second, smallermound adjoining WNW. Sasanian-possible EarlyIslamic.

1086 - 200 NW X 90 X 1. Collecting conditionspoor as a result of adjacent village, modern debris,soil salinity. Early Dynastic II/III, Cassite, someParthian.

1087 Tell Kofyaya. 260 diam. X 2.4. Trace of EarlyDynastic I. Sasanian-Early Islamic.

1088 300 NW X 190 X 2.4. Sparse pottery.

..

272

''

Mainly Ur III-Larsa. Some Neo-Babylonian-Achaemenian.

1089 130 diam. X 3. Sparse pottery. Early Dy-nastic II/III-Akkadian-Ur III-Larsa, Recent.

1090 Tell Fakhar. 350 NE X 250 X2.8. Cassite-MiddleBabylonian; possibly also Neo-Babylonian.

1091 - 250 NW X 200 X 2. Cassite-Middle Baby-lonian, Parthian.

1092 70 diam. X 1.4. Sasanian-Early Islamic.1093 280 NW X 120 X 1. Limited collection.

Probably Middle/Neo-Babylonian. Mainly Sasan-ian-Early Islamic.

1094 80 diam. X 2.2, although this may includethe eroded ruins of a small qal'a or even a stabil-ized dune. Possible Middle/Neo-Babylonian. Sa-sanian, Recent.

1095 - 250 WNW X 140 X 2. Surface pottery ex-tremely sparse and decomposed. 30 m NNW is asecond mound, 110 WNW X 60 X 1.2. Mainmound probably Sasanian. NNW mound includedtrace of Uruk, possible Middle/Neo-Babylonian.Mainly Sasanian.

1096 Tell Umm al-Fugas. 350 NW X 250 X 3. 100m NW is a second (or continuation of the first),180 diam. X 1.8. Limited collection owing tohigh soil salinity; hence types could be recordedonly on a presence-absence basis. Middle/LateUruk. Mainly Jemdet Nasr. Limited continuationinto Early Dynastic I.

1097 Ishan al-'Arraj. 200 ENE X 100 X 2.2. Sasanian-Early Islamic.

1098 180 diam. X 1.5, although surface potteryis largely confined to central part of the elevatedarea. Saline soil; poor collecting conditions. Prob-ably Cassite-Middle Babylonian.

1099 - 170 NW X 120 X 2. These dimensionshave surely been reduced, however, by a majormodern canal cut through the S part of the site,and additional parts of the mound that have beenremoved for road fill. Old Babylonian-Cassite.Inspectorate of Surveys records also assign an"Ubaid" dating to this mound, which is likely tomean at most that one or more clay sickles wereobserved at an earlier time.

1100 - Low hummocks within an area probably180 NE X 100, but obscured by drifted sand.Some Uruk, mainly Jemdet Nasr, a little Ur III-Larsa.

1101 Part of Tulil Abu Dan. 400 diam. X 4.5. Parthian-Sasanian-Early Islamic.

1102 Part of Tiulul Abui Dan. 110 diam. X 4.5. Cassite-Parthian.

1103 Tell Dalmaj. 350 NNE X 280 X 8. Early/MiddleUruk, Larsa-Parthian. A small Sasanian occupa-tion is confined to the summit.

oi.uchicago.edu

General Site Catalog

1104 Rubahi Dalmaj. 140 diam. X 4.5. Sasanian-lim-ited Early Islamic.

1105 180 NNW X 150 X 0.5. Neo-Babylonian-Achaemenian-Parthian.

1106 120 diam. X 0.4. Sasanian-Early Islamic.1107 300 N X 180 X 4.2. Immediately W is a

second, 80 diam. X 2.2. 80 m farther W is a third,130 diam. X 1. Sasanian-Early Islamic.

1108 Tell al-Selu'. 280 diam. X 5. Trace of Uruk. Neo-Babylonian-Achaemenian-Parthian.

1109 - 190 diam. X 2. Limited Uruk, mainlyLarsa-Old Babylonian, less Cassite. A few Sasan-ian sherds were regarded as strays from site1110.

1110 220 diam., with four low summits rising to1 m. Sasanian.

1111 280 NNE X 100 X 2. Probably Sasanian.Early Islamic.

1112 150 diam. Much evidence of late brick-kilnactivity so that site may have been higher orig-inally, but two large hummocks still reach 0.8-1m ht. Possibly a small Uruk settlement. Ur III-Larsa.

1113 200 diam. X 2. Early Uruk. Mainly Cassite.A little Sasanian.

1114 300 NE X 140 X 4. Early/Middle Uruk,Cassite, Parthian.

1115 - Main mound (or perhaps a series of threeadjoining mounds) 450 ENE X 210 X 3. SE, de-tached, is another, 170 diam. X 2.5. Possible Uruksite. Mainly Akkadian-Larsa. A few Parthian slip-per-coffin fragments may reflect a cemetery. Thereis an isolated Sasanian building (32 cm sq. bricks)on top of the WSW end of the main mound.

1116 330 diam. X 7.5. Sasanian-Early Islamic.1117 320 ENE X 150 X 1. Sasanian-Early

Islamic.1118 - 80 E X 40 X 0.1. 30 m W is a second, 70

E X 40 X 0.1. But while these limits reflect theslightly elevated areas with densest debris, it mustbe noted that there is a very wide dispersal of re-latively sparse sherds. Intensive collections fromcircles of 5 m radius: #1, W end of E mound (se-lected as a possible kiln area with many sherds);#2, W end of W mound. Early Uruk, possibly alsoMiddle Uruk.

1119 - A small site completely drifted over withlow, stabilized dunes and other aeolian deposits.Hummocks have formed around tamarisk roots,and it was the unusual elevation of these hum-mocks that suggested the existence of a site here.On the basis of only sporadic exposures of surfacematerial, the site is probably about 120 diam X0.6. Trace of Uruk. Sasanian-Early Islamic.

1120 300 diam. X 5.5. Plain-level debris extendsat least 200 m in all directions from the outerslopes, and 60 m N there is a second mound of lessthan 1 ha area that is almost as high. Sasanian-Early Islamic.

1121 - 320 ENE X 180 X 3. Settlement debriscontinues WSW for 900 m along branch canallevee leading from Shatt al-Nil, thence upstreamalong the Shatt for 250 m farther. Traces can beseen of buried brick arches, apparently still intact,of 25 cm sq. brick. Perhaps they represent branchcanal headworks. Sasanian-Early Islamic.

1122 130 diam. X 2.2. Sasanian-Early Islamic.1123 120 NNE X 60 X 0.1. Achaemenian-

Parthian.1124 310 E X 240, the maximum width some-

what W of center. Maximum ht. 1 m, risingabruptly to this on W end and sloping awaygradually to N, E, and S. Much kiln activity con-centrated near W end, and concentrations ofwasters and cinders there may account for ht. In-tensive collection made in circle of 5 m radius nearW summit, supplemented by individual types ob-served elsewhere. Mainly Early Uruk. Middle/Late Uruk localized.

1125 140 diam. X 0.8. 30 m SSW is a second,120 diam. X 0.8. Sasanian-Early Islamic.

1126 - 160 diam. X 2.5. 80 m SW a second, 120diam. X 0.8, with ruin of a tower. Sasanian. Thesmaller mound is Recent.

1127 Tell Abfi Jawarir. 280 NE X 240 X 5. AdjacentNNW, perhaps across an old canal bed, is asecond, 180 NE X 80 X 2. Achaemenian-Parthian.

1128 Main mound 180 diam. X 1, but lowerdebris extends SSW in an irregular outline to givea maximum length of 380 m. Sasanian.

1129 - 200 NW X 120 X2.2. Early/Middle/LateUruk-Jemdet Nasr, Larsa-Old Babylonian-Cas-site.

1130 - 400 NW X 330, with three 4 m summitsand three lesser ones. Bricks 33 cm sq. Trace ofUruk. Sasanian, probably also Early Islamic.

1131 320 NW X 180 X 2.4. Middle/Late Uruk.Mainly Sasanian.

1132 - 140 N X 110 X 1.4. Sasanian-Early Is-lamic.

1133 - 180 diam. X 3. Sasanian-Early Islamic.1134 - 160 WNW X 70 X 2.5. A 1.5 m mound

on the W end may be the ruin of a small qal'a.Trace of Uruk. Seleucid-Parthian.

1135 Tell Rubahiyat al-Torra. See sketch map. Prin-cipal summit is 5 m high. Probably an underlying,small Uruk site. Primarily Parthian-Sasanian-Early Islamic.

273

oi.uchicago.edu

General Site Catalog1135

1136 130 E X 90 X 0.3. Parthian-Sasanian.1137 140 NW X 100 X 1.2, virtually all the

elevation probably due to a much later structureof which a few broken, reused bricks may be seenon the surface. Plain-level debris, the density ofsherds decreasing only gradually at the margins,describes a much larger site area, perhaps 230WNW X 120. After a short gap to the W thereis a second site, roughly 100 diam. and very low.Intensive collection #1 was made on the N centralportion of the E mound, #2 on the E central partof the W mound; both were taken in circles of 5 mradius. Mainly Early/Middle/Late Uruk. JemdetNasr-Early Dynastic I limited to a small area onthe adjacent ends of the E and W mounds. Verylimited Cassite.

1138 350 E X 130 X 2. Sasanian.1139 300 diam., divided by the Shatt al-Nil, ris-

ing to 3.5 m. Bricks 24 cm sq. particularly numer-ous on N side. Probably Sasanian. Early Islamic.

1140 - 120 diam. X 2, on left bank of Shatt al-Nil. 20 m NNE is a second mound, 50 diam. X 1.Small, high, conical mounds of baked brick frag-ments occur downstream and upstream of site onsame bank. Brick sizes include 23 and 27 cm sq.Sasanian-Early Islamic.

1141 200 diam. X 2.4, left bank of Shatt al-Nil.On opposite bank a second, 130 diam. X 2.4.Also on right bank, 100 m NW, a third, 100 diam.X 2.2. Sasanian-Early Islamic.

1142 - 140 E X 120 X 2. Adjacent ENE is asecond, 120 NE X 90 X 2. N of the second is athird mound, 120 diam. X 2. NE of the secondis an area of low debris, 70 diam. X 0.6. All thesemay or may not form parts of a single, contiguoussettlement; surrounding and overlying dunes madesurface examination inadequate. Sasanian-possi-ble Early Islamic.

1143 Site occurs in the middle of a belt of veryheavy dune deposits that overlie, and probablyare derived from, the levee of a major ancientwatercourse. The mound must certainly be 250-300 m diam. X 4-5 m ht., if not more, but effec-tive surface examination was limited to less thana tenth of this area. Akkadian-Larsa.

1144 - An irregular but extensive settlement alongthe right bank of the Shatt al-Nil. About 680 NWX 90-150 X 2.4, with a W extension for 220 malong a branch canal offtake. Sasanian-EarlyIslamic.

5 300n M

274

v.v "I,I 1145 320 ENE X 170 X 3. Seleucid-Parthian.1146 260 diam. X 6. A second mound, im-

mediately E, is 240 N X 180 X 2.2. Sasanian-Early Islamic.

1147 180 NW X 90 X 1.8. Sasanian-possibleEarly Islamic.

1148 Tell Abfi Idhin. 200 E X 170 X 4. Immediately Eis a second, 150 NW X 90 X 1.8, and two smallermounds tail off SW from W end. Sasanian-EarlyIslamic.

1149 - 180 NW X 150 X 1.6. A smaller, lowermound adjoins NE. Sasanian-Early Islamic.

1150 - 250 NW X 90 X 1. Sasanian-Early Is-lamic.

1151 - Four small mounds that may have beenbrick kilns, the largest 80 diam. X 2.5. Sasanian.

1152 Tell Hamayma. 400 NW X 240 X 6.5. Possiblya small, underlying Uruk settlement. Trace ofCassite. Neo-Babylonian-Achaemenian-limitedSeleucid-Parthian.

1153 - 120 diam. X 1.3, bisected by a canal leveefrom NW. Site consists primarily of several an-cient brick kilns. Seleucid-Parthian, although smallhummocks of Sasanian pottery are found along acanal levee from N that passes W of site.

1154 Perhaps 160 diam., plain level. Relativelysparse pottery, and definition of limits of site iscomplicated by very wide dispersal of debris. In-tensive collection within circle of 5 m radius,mound center. Possibly Early/Middle, mainly LateUruk.

1155 - 90 diam. X 2. 20 m E is a second, 90 diam.X 1. Sasanian-Early Islamic.

1156 100 diam. X 1.8. 200 m S along an oldcanal levee is a second, 80 diam. X 1.6. A third,smaller, is 150 m farther S. Neo-Babylonian-Achaemenian-Seleucid-Parthian.

1157 - 150 diam., plain level. Exposures of debrisare sparse and somewhat discontinuous. Larsa.Limited Parthian.

1158 - 180 diam. X 1.8. 50 m WNW is a second,70 diam. X 1.4. A ribbon of surface debris con-nects and surrounds the mounds and extends 300m farther E. Parthian.

1159 - At least 300 NW X 150, but sparse debrisis much more widely dispersed than this, andclear margins are lacking. Intensive collectionwithin circle of 5 m radius, NW central. EarlyUruk.

1160 - 120 NNW X 50 X 0.1. Parthian.1161 - 220 diam. X 2.4. Trace of Cassite.

Neo-Babylonian-Achaemenian-limited Seleucid-Parthian.

1162 - 550 NNW X 160 X 2.2. Parthian.1163 - 130 N X 70 X 0.1. Intensive collection

oi.uchicago.edu

General Site Catalog

made in circle of 5 m radius, center. Late Uruk.1164 130 NW X 70 X 0.1. 30 m SE is a second,

30 diam. X 0.1. Intensive collection made in circleof 5 m radius, main mound, NW central. Early/Middle Uruk. Late Uruk limited.

1165 - 230 diam. X 0.5. Intensive collection madein circle of 5 m radius, S central. Middle-mainlyLate Uruk.

1166 - 380 NW X 280 X 2. Early/Middle Uruk,Jemdet Nasr. Limited continuation in Early Dy-nastic I. Akkadian sherds sparse but fairly wide-spread, Ur III-Larsa extremely limited and local-ized. Old Babylonian-Cassite again widespreadand substantial.

1167 280 NE X 60-80 X 0.6. Sasanian.1168 35 diam. X 0.1. Jemdet Nasr.1169 180 NNE X 140 X 2.2. Possible Uruk and

Cassite settlements-or just strays from neighbor-ing site 1170. Mainly Seleucid-Parthian.

1170 130 diam. X 1.4. Early/Middle Uruk, OldBabylonian-Cassite.

1171 Jer'at Hayf. 750 NW X 120 X2.8, but cf. irregu-lar outline shown on base map. Additional smallmounds continue SE along an old canal levee butcould not be visited because of intensive cultiva-tion around them. Perhaps a small, underlyingCassite site. Mainly Neo-Babylonian-Achaemen-ian-Parthian.

1172 See sketch-map. It is not clear whether thisis a cluster of adjoining settlements or the partiallysubmerged remains of a single, nucleated one.Sherd density declines only slowly at the marginsof site area. That, together with low but fairlythick shrub cover of the plain here and hts. ofdebris nowhere exceeding 0.2, introduces an ele-ment of uncertainty and arbitrariness into anymap like the one shown. Intensive collections #1and #2 both were made in circles of 5 m radiusat points shown. Collection #3 included everydiagnostic sherd seen on a NW-SE traverse of theindicated mound. Early/Middle/Late Uruk, prob-ably with some reduction of occupied area in LateUruk.

1172

S01'

fN ..

300 M

1173 - 110 N X 70 X 0.8. Sasanian-Early Islamic.1174 180 NW X 110 X 2.8. Small Uruk settle-

ment is likely. Trace of Early Dynastic I. PrimarilyOld Babylonian-Cassite, with a smaller Sasanian-Early Islamic occupation on the lower slope andadjoining plain to the E.

1175 400 NE X 180 X 2.5. Air photographssuggest, however, that originally there were twomounds in a NE line, with the one at the NE endbuilt out over the former bed of a major water-course. There is Early Dynastic II/III debris alongthe NW slope of the SW mound, but both aremainly Old Babylonian-Cassite-Middle Babylo-nian. The main bed must have been abandonedby that time.

1176 160 diam. X 2.2. Sasanian-Early Islamic.1177 Discontinuous clusters of debris at plain

level within an area 140 diam. Old Babylonian-Cassite.

1178 - 190 N X 110 X 0.8. Failing light precludedan intensive sample collection. Types were re-corded on a presence-absence basis only. Uruk,Early Dynastic I/II/III-limited Akkadian.

1179 - 240 diam. X 3. Trace of Uruk. Larsa-OldBabylonian-Cassite.

1180 350 diam. X 3.2. Probably a small Uruksite. A few Larsa sherds may be strays from nearbysite 1188. Mainly Neo-Babylonian-Achaemenian-Parthian, limited Sasanian.

1181 - 300 NW X 200 X 2.2, with lower debrisand small, low mounds tailing off NW toward site1180. There are possible traces of an outer walland/or tower of baked brick at NW corner ofmound. NW part of the site seemed mainly Ak-kadian-Larsa, SE end predominantly Neo-Baby-lonian-Achaemenian.

1182 Tell al-Sema'. 140 diam. X 1, with a small,ruined tower forming a mound of debris that risesa further 1.5 m. Rare Cassite. Mainly Sasanian-Early Islamic. Also Recent.

1183 - 300 NNW X 160 X 2.8. Old Babylonian-Cassite, Sasanian-Early Islamic, Recent.

1184 - 180 NNW X 140 X 2.4. 30 m N is a sec-ond, 160 diam. X 2.2. A third, smaller and lower,is NNW of the second. A fourth, also small, is70 m WSW of the first. The first is Sasanian-possibly also Early Islamic. The others are Sasan-ian only.

1185 - 220 WNW X 140 X 2.4, right bank ofShatt al-Nil. 20 m WNW on same bank is a second,180 WNW X 130 X 2. A third, on left bank oppo-site first, is 300 WNW X 160 X 2 (but with onlythe central portion rising to substantial ht.). Bricks23, 27 cm sq. Trace of Uruk. Sasanian-Early Is-lamic-possibly Samarran.

275

oi.uchicago.edu

General Site Catalog

1186 Tell Abf 'Alayma. 160 diam. X 2.2. Sasanian-Early Islamic.

1187 - 180 WNW X 140 X 4.5, on right bank ofShatt al-Nil. Sasanian-Early Islamic.

1188 - An important town, recently and exten-sively looted, 1,050 NW X 630 X 2.5. A broadswale, though not descending to plain level andcontinuously covered with sherds and other debris,separates the NW and SE halves of the site. TheNW half, slightly higher and longer-lasting, hasnot been seriously disturbed. But the entire SE halfhas been almost continuously pitted. The SE endof the site remains somewhat ill defined owing toa continuous belt of dunes. Surface-level debrismay continue in that direction for 500 to 800 m.Early Dynastic II/III. Mainly Akkadian. Ur III-Larsa occupation is limited to the higher, NW partof the site.

1189 - 300 NE X 200 X 2.8. Sasanian-Early Is-lamic.

1190 100 diam. X 1. Sasanian.1191 170 E X 110 X 1.4. Sasanian-Early Is-

lamic.1192 The central mound is only 60 diam. X 1.4,

but surface debris tails off 150 m E and almost500 m W. Sasanian. There is an unusual pre-ponderance of storage jar sherds in the surfacedebris.

1193 Tell Mirza. 750 E X 600 X 6.5. Parts of thesite are only a little above plain level, but sur-face debris is dense and continuous. At plain level,moreover, debris also extends 200 m SE of site andfor varying but generally shorter distances in otherdirections. Bricks on site preponderantly 33-36 cmsq., but also 27-30 cm sq. 50 m E, across an old,major canal bed is a second mound that is surelya continuation of the same site, 120 diam. X 2.2,and from here indistinct small mounds and hum-mocks of debris tail off SSE along a levee. An iso-lated small tower, now in ruins, stands 150 m NWof this second mound. Parthian confined to thecanal levee passing between the two mounds andcontinuing SSE. Sasanian-possible (or at any rate,limited) Early Islamic.

1194 - 350 NW X 330 X 0.5. Intensive collectionmade in circle of 5 m radius, center. Possible traceof Ubaid II. Primarily Early/Middle Uruk, limitedLate Uruk. Also a very little Sasanian-Early Is-lamic.

1195 - 160 NW X 100 X 1.2, with small hum-mocks and plain-level debris extending 600 m SEalong old canal levee. 400 m NW is a second, 130NW X 80 X 0.2. Possibly a small Uruk site under-lies the second mound. Sasanian-possible EarlyIslamic.

1196 130 E X 70 X 0.2. Uruk. Mainly OldBabylonian-Cassite.

1197 140 E X 80 X 0.2. Possible Middle-mainlyLate Uruk. Jemdet Nasr-Early Dynastic I con-fined to a very small area at E end of site. SomeSasanian-Early Islamic.

1198 - 210 NNW X 160 X 0.3. Intensive collec-tion made in circle of 5 m radius, N central. EarlyUruk. A small Jemdet Nasr resettlement is confinedto the S end.

1199 190 diam. X 3. Early/Middle Uruk. Pri-marily Larsa-Old Babylonian.

1200 240 diam. X 0.8. Sasanian.1201 160 NNE X 120 X 1.4. Sasanian.1202 Perhaps 160 NW X 80, mostly at little

more than plain level. The remains of a single smallbuilding (28 cm sq. brick) rise to 1 m. Sasanian-Early Islamic.

1203 Part of the Tfilul Ruwayjah group (cf. site1226). 310 NW X 210 X 2.4. Sasanian-Early Is-lamic.

1204 140 diam. X 1.8. Sasanian-Early Islamic.1205 360 NW X 220. The summit of the main

mound, toward its SE end, is 1.5 m. Later settle-ment apparently was concentrated at the NW end,however, and a more localized summit there canbe described as perhaps 100 diam. X 2.2. Concen-trations of large (13-17 cm) clay cones with insetheads for pigment are found on the surface just NEof main mound summit. Intensive collections incircles of 5 m radius: #1, main mound summit, SE;#2, NW central. Mainly Early-Middle Uruk. Prob-ably Late Uruk. Jemdet Nasr-Early Dynastic Ilimited. The later, relatively small settlement onthe NW end is Old Babylonian-Cassite.

1206 - 180 diam. X 1.8. Old Babylonian-Cassite,limited Sasanian.

1207 - 150 NNW X 110 X 0.6. Trace of Uruk.Parthian-Sasanian.

1208 130 NNW X 100 X 0.8. Probably a small,underlying Uruk site. Old Babylonian-Cassite.

1209 Qubur Hathot. 280 NNW X 150 X 2.8, butthe low NNW end of site extends E to give it asomewhat triangular outline in plan. Trace ofUruk. Achaemenian-Parthian-Sasanian.

1210 - 130 diam. X 2. 150 m NNE, connected bya ribbon of plain-level debris, is a second mound220 diam. X 1.6. A third, immediately W of thefirst, is 140 diam. X 0.6. Probably a small, under-lying Uruk site. Trace of Ur III-Larsa. Sasanian-Early Islamic.

1211 - 160 NW X 120 X 2.5. Parthian, the sur-face debris seeming to include an unusually highproportion of large, pointed-base storage jars.

276

oi.uchicago.edu

General Site Catalog

1212 - 160 NW X 120 X 1. Sasanian-possibleEarly Islamic.

1213 An irregular strip of settlement along bothbanks of the Shatt al-Nil above an important con-centration of branch canal offtakes. The strip ex-ceeds 1.5 km in length and is generally from 100to 250 m in width. Parts of it are quite low, butseveral individual mounds exceed 4 m in ht. Theradiating branches that fan out here strongly implythat there was once a regulatory weir. No surfaceevidence could be found for it, though there areseveral suggestive indications of sluice gates. Many33 cm sq. bricks, particularly at downstream end,but 27 cm sq. are more common. Sasanian-EarlyIslamic. An intact jar was found buried in a thincasing of asphalt, rim at surface level, on mainright bank mound just above offtake fan. Brokenin antiquity, it was apparently set in asphalt to pro-vide storage. This implies complete wind deflationof the building housing this jar.

1214 400 diam., low and rolling; hts. not ex-ceeding 1.2 m. Sasanian-possibly also Early Is-lamic.

1215 300 NNW X 160 X 0.3. Probable JemdetNasr. Early Dynastic I.

1216 - 220 N X 150 X 3.5, with a low tail ofdebris extending E. 400 m SE is a second mound,220 diam. X 0.2. Intensive collections within cir-cles of 5 m radius on latter: #1, SE central;#2, NW central. Second mound: Possible MiddleUruk. Mainly Late Uruk. Jemdet Nasr much re-duced. Very sparse Sasanian-Early Islamic. Firstmound: Neo-Babylonian-Achaemenian is local-ized on E tail. Mainly Sasanian-Early Islamic.

1217 Tell Dowayhis. At least 300 diam. X 10-13,but size obscured by dunes. 250 m SSW, acrossa broad old channel from the WNW, is a smalloutlier 70 diam. X 1.2. Uruk, Early Dynastic II/III-Larsa.

1218 120 diam. X 1. Old Babylonian-Cassite.1219 240 NW X 200 X 1.7. Trace of Uruk. Cas-

site-Parthian.1220 Three adjoining mounds form a N-S line

300 X 160 X 3. Sasanian-possibly also Early Is-lamic.

1221 - 130 diam. X 2.2. Trace of Uruk. Old Baby-lonian-Cassite, Sasanian-Early Islamic.

1222 - 260 E X 140 X 2. Sasanian-Early Islamic,possibly continuing into Samarran.

1223 - 170 NE X 100 X 2. Sasanian-possibleEarly Islamic.

1224 - Three tells adjoining a NE canal levee, eachabout 70 diam. X 1. Sasanian-possible Early Is-lamic.

1225 150 diam. X 2. Sasanian-possible Early Is-lamic.

1226 Tell Ruwayjah. 220 diam. X 2.6. Trace of Cassite.Sasanian-possible Early Islamic.

1227 300 NE X 220 X 0.6. Sasanian-possibleEarly Islamic.

1228 - 80 diam. X 0.4. Sasanian-possible EarlyIslamic.

1229 - Eight semidetached mounds in an area 700NE X 450; see base map for locations. The largestand highest, in the SW and SE, rise to 4 and 4.5m respectively. Sasanian-possible Early Islamic.

1230 Part of Tell al-Laham; cf. site 1231. 160 diam. X2, with a ruined qal'a near W end rising an ad-ditional 2.4 m. Lower debris, mainly from brickkilns, extends the area of the site considerablyfarther to the NE. While the site falls within,or at least very close to, the circle of outwash debrisfrom site 1231, it is clearly detached from the rec-tangular town plan of the latter as seen on airphotographs. Moreover, the sherd evidence arguesthat it should be considered a separate and for themost part later settlement. A small Early/MiddleUruk settlement here is likely. Old Babylonian.Mainly Cassite. Middle/Neo-Babylonian-Achae-menian, Recent.

1231 Tell al-Laham. 500 diam. X 7. While this cir-cular form may best describe the present appear-ance of the site, however, the air photographs sug-gest a rectangular shape for the original town.Ur III-Larsa, Middle/Neo-Babylonian-Achaeme-nian-Parthian.

1232 120 diam. X 2.3. Seleucid-Parthian.1233 130 diam. X 2.2. Trace of Uruk. Old Baby-

lonian-Cassite.1234 200 NW X 120 X 0.5. Sasanian. There are

rare sherds of Old Babylonian-Cassite and Seleu-cid-Parthian, but these are assumed to have beensecondarily introduced from sites 1233 and 1232respectively.

1235 Tell Abfi Qubfir. 300 NW X 180 X 3. 90 mSSW is a second, 160 NW X 130 X 2. Cassite-Parthian.

1236 - 160 NW X 130 X 2. Rare Cassite, pri-marily Neo-Babylonian-Achaemenian.

1237 Tell Dlehim. 700 NE X 630 X 18, rising tothis ht. only in a steep, zigguratlike summit inthe WSW part of the site. The rest of the site ismuch lower, the most prominent mounds lyinggenerally to the E and ENE and not exceeding 1.5-2.5 m. Beyond the dimensions indicated, aerialphotographs indicate a possible further extensionof the site to the ESE, across a modern canal, andother, more ancient canals also appear to have

277

oi.uchicago.edu

General Site Catalog

been cut through the ruins as shown in the basemap. Differences in soil color and compactionmake it possible to trace an apparent street in theNE part of the site that is about 6 m in width andcan be followed for several hundred meters leadingin toward the apparent center of the city; it has atleast one intersection with a narrower street lead-ing laterally SSE. Poor drainage and high soil salin-ity make collecting conditions generally poor, andidentifiable ceramic types are correspondinglyscarce. The size of the site is such, however, that itwas possible in time to assemble a sufficient col-lection from all parts of the mound to attach rea-sonable confidence to a dating estimate. Quite pos-sibly, the major occupation of the site was in theEarly Uruk and Jemdet Nasr periods, for indi-cators of these periods are numerous and very ex-tensively distributed. The absence of beveled-rimbowls, on the other hand, may suggest a Late Urukhiatus. Later surface material, to be sure, is alsodistributed virtually as extensively as the ruinsthemselves and is naturally more abundant. Itsuggests an Akkadian-Ur III or possibly Akka-dian-Larsa range of occupation. A little Parthianpottery also occurs, concentrated around the highmound and perhaps representing no more than asmall, late fortification built on its summit.

1238 Tell Khazir. 180 NW X 150 X 3.5. The centralsummit is largely composed of fallen 21 cm sq.bricks. Sasanian-Early Islamic. Possibly there wasa single, later building on the summit after theabandonment of most of the site.

1239 200 NW X 140 X 1.8. A narrow ribbon oflow debris continues NW for perhaps 700 m, vir-tually to site 1238. Sasanian-Early Islamic.

1240 - 230 NNE X 160 X 2. Limited Akkadian-Larsa. Mainly Old Babylonian-Cassite.

1241 Tuiluil Abuf Gharkha. 250 diam. X 3. Sasanian-possible Early Islamic.

1242 200 diam. X 1.8. A strip of low debris,varying in width from 20 to 120 m, extends 500m NW from mound along an old canal levee. Sasa-nian-possibly Early Islamic.

1243 Qal'a Abi Gharkha. The ruined qal'a is 14 msquare, the largest of its four corner towers rising4.5 m on the SSW. Sparse pottery is found aroundit at plain level. Sasanian-Early Islamic, Recent.

1244 Tell al-Hraymsat. 240 NNE X 180 X 4.5. Profusesurface pottery. Akkadian-Larsa.

1245 Tell al-Hraymsat. 260 NW X 150 X 4.5. Neo-Babylonian-Achaemenian-Parthian.

1246 - Perhaps 150 diam. X 0.3, although partlyobscured by dunes. Sasanian-Early Islamic.

1247 - 220 NW X 120 X 0.2. Possible Uruk andJemdet Nasr. Limited Early Dynastic I. Mainly

Early Dynastic II/III. Very limited Akkadian-Larsa.

1248 250 diam. X 2.. Neo-Babylonian-Achae-menian-Parthian.

1249 - 240 NW X 160 X 3.5. Akkadian-Larsa,Cassite-Achaemenian. Possible Seleucid-Parthian.

1250 - 240 NW X 180 X 1.8. Old Babylonian-Cassite.

1251 - 220 NW X 140 X 3.6. Neo-Babylonian-Achaemenian-Parthian.

1252 140 diam. X 2.8, with a wide tail of lowerdebris extending more than 300 m NW. Seleucid-Parthian-Sasanian-possible Early Islamic.

1253 - 220 WNW X 160 X 2.4. Achaemenian-Parthian.

1254 170 diam. X 1.5. Seleucid-Parthian-Sasa-nian.

1255 - 130 diam. X 4.5, surrounded and partlycovered by dunes. Sasanian-Early Islamic.

1256 180 diam. X 2.6, completely surroundedby dunes. Larsa-Old Babylonian-Cassite.

1257 - 300 diam. X 6. At the N foot of the moundis an apparent cemetery area covering more than1 ha. It has been very recently and extensivelypitted. The cemetery is apparently Early DynasticII/III-Akkadian, the mound Akkadian-Larsa.

1258 - 220 diam. X 4.8, with debris tailing offSE for more than 300 m. Sasanian-Early Islamic.

1259 250 NNW X 140 X 1.2. Akkadian-Ur III-Larsa-Old Babylonian-Cassite. There is also alittle Sasanian-Early Islamic, but not enough toassure a primary settlement here.

1260 130 NNE X 60 X 0.8. Immediately SSE isa second of the same size. 50 m NW is a third, 40diam. X 1.8. Sasanian-Early Islamic.

1261 - 110 N X 70 X 0.2. Intensive collectionmade in circle of 5 m radius, N center. PossibleMiddle Uruk, mainly Late Uruk.

1262 400 NW X 280 X 4. Sasanian-Early Is-lamic.

1263 80 NNW X 70 X 2, tailing off SSE in lowdebris. 60 m ENE across an old canal junction is asecond, 180 WNW X 100 X 1.8. Parthian-Sasa-nian. Storage jar fragments are heavily representedin the surface materials at the first mound, brick-kiln refuse at the second.

1264 420 E X 160 X 2.5. Neo-Babylonian-Achaemenian-Parthian.

1265 250 diam. X 1.8. To the W is a secondarea of debris, 120 diam. X 0.3. In addition, a 50m wide band extends 200 m SE from the mainmound along an old canal levee. Sasanian-EarlyIslamic.

1266 240 ENE X 140 X 2. Sasanian-Early Is-lamic.

278

oi.uchicago.edu

General Site Catalog

1267 180 diam. X 1.8. Old Babylonian-Cassite.1268 250 NW X 150 X 2.6. Sasanian-Early Is-

lamic.1269 180 diam. X 2.2. Seleucid-Parthian.1270 320 NE X 130 X 2.2. Neo-Babylonian-

Achaemenian-Seleucid-Parthian.1271 - Perhaps 130 diam. X 3, although drifted

sand around base of mound may conceal low ex-tensions. Early/Middle Uruk, Early Dynastic II/III-very limited Akkadian.

1272 - 350 ENE X 170 X 5. Trace of Uruk, Neo-Babylonian-Achaemenian-Parthian.

1273 Tell Dhahiya. Continuous debris, multiple con-tiguous summits suggest a roughly square areaof settlement, 700 m sides oriented NW and NE.The main conical summit rises to 8.5 m. Sasanian-Early Islamic.

1274 300 diam. X 2.2, with a slight elongationto the W. Sasanian-Early Islamic.

1275 A square mud-brick enclosure, 90 m sidesoriented NW and NE, wall stubs still standing 3m high. Only Sasanian-Early Islamic pottery wasseen, although the condition of the walls arguesthat the enclosure itself is Recent.

1276 - 300 diam. X 4, tailing off 500 m SW. 600m farther SW is a small outlier, 200 NW X 140 X1.8. Sasanian-Early Islamic.

1277 - 180 diam. X 1. Sasanian-possible Early Is-lamic.

1278 300 WNW X 220 X 3.4. Very limitedUruk, Early Dynastic I. Primarily Akkadian-Cas-site. Sasanian storage jar sherds may reflect the useof the mound at that time only as a cemetery.

1279 160 diam. X 2.5. Larsa, Sasanian-Early Is-lamic.

1280 Eight mounds around a major canal junc-tion, within an area 700 NE X 450 and with hts.up to 3 m. But settlement was certainly not con-tinuous, as plain-level debris is sparse, and canalspoil banks also may exaggerate the apparent sizeof the site. Sasanian-Early Islamic.

1281 180 E X 90 X 1.6. Sasanian-Early Islamic.1282 190 diam. X 1.4. Neo-Babylonian-Achae-

menian-Parthian.1283 - 70 diam. 0.8, bisected by an old canal

from NW. Achaemenian-Parthian.1284 - About 250 diam., sparse and discontinuous

plain-level debris interspersed by eight low hum-mocks with denser sherd accumulations. Early/Middle Uruk. Primarily Cassite-Achaemenian.

1285 140 E X 70 X 2. Sasanian-Early Islamic.1286 - 400 NW X 140 X 3. Sasanian-Early Is-

lamic.1287 - 260 diam. X 4. 140 m NW is a second, 160

diam. X 2.2. Sasanian-Early Islamic.

1288 120 E X 90 X 2.8. Sasanian-Early Islamic.1289 220 NNW X 140 X 3.4. 40 m E is a second

mound, 180 diam. X 1.4. Sasanian-Early Islamic.The later period is seemingly less well representedon the second, lower mound.

1290 740 E X 160 (W end)-240 (E end) X 3.5.Akkadian-Larsa.

1291 - 210 N X 140 X 2.5. Trace of Uruk. Neo-Babylonian-Achaemenian.

1292 - 200 diam. X 1. Sasanian.1293 Tell Abui Dhaba'. 550 E X 280 X 4.2, but

both ground observation and air photographs sug-gest a pair of adjoining tells on an E-W line ratherthan a single elongated settlement. The site is ap-proached from the W by a straight, clearly de-fined bed that continues E beyond it. Several prob-ably lateral offtakes from this canal now take theform of curious low ridges capped with pottery.Probably this implies considerable wind deflationof the land surface in this vicinity, to the pointwhere sherds in abandoned canal branches pro-vided a protective cap over underlying sedimentsand ultimately left raised ribbons of debris insteadof incised channels. Sherds indicating a possibleMiddle/Late Uruk occupation are concentrated,although not completely localized, in the wide, lowsaddle between the two adjoining mounds. Theremainder of the site was more or less continu-ously occupied from Akkadian through Parthiantimes, with Larsa, Cassite, and Parthian the domi-nant surface components. Possibly, however, the Emound was in use primarily as a cemetery in theNeo-Babylonian, Achaemenian, and Parthian pe-riods.

1294 110 E X 80 X 3.2. Centered on the moundare traces of a rectangular enclosing wall 70 E X40, with irregularities in profile suggestive of cor-ner towers at least on the SW and SE. On the sum-mit of the mound, centered within this enclosure,is a similarly oriented square structure of 31-33cm sq. baked brick, 10 m on a side. The S half ofthe summit has eroded down to the foundationsof this structure, but it seems possible that it wasapproached on the W and perhaps N by stairwaysof baked brick. At the NW and NE corners of thisinner, square structure there are projecting tow-ers. Entrances through its 93-96 cm thick walls areplaced at the N end of the W side and in the middleof the N side; both are 90 cm wide. Early/MiddleUruk, Larsa-Old Babylonian are primary periodsassociated with habitation of the mound itself. Theoverlying structure is of Sasanian date and mayrepresent a fire temple.

1295 - 140 NNW X 90 X 2.2. Neo-Babylonian-Achaemenian-Parthian.

279

oi.uchicago.edu

General Site Catalog

1296 - 100 diam. X 1.8. Sasanian-possible EarlyIslamic.

1297 - Three small tells, probably separated bycanals, form an ENE line 220 X 120 X 1. Sasa-nian.

1298 320 N X 90 X 1. Sasanian-possibly alsoEarly Islamic.

1299 - 220 E X 140 X 1, bisected by an old (butprobably later than the site) canal from the N.Old Babylonian-Cassite.

1300 - 210 NNW X 90 X 0.8. A second, muchsmaller mound is 20 m WSW. Sasanian-possibleEarly Islamic.

1301 130 diam. X 1. Cassite-Achaemenian.1302 90 diam. X 0.8. Sasanian-possible Early

Islamic.1303 120 NW X 1.8. Trace of Uruk. Larsa.1304 Also known as Tell al-Hayyad (cf. site 1306).

240 diam. X 2.6. Probably a small, underlyingUruk site. Akkadian-Larsa.

1305 140 diam. X 1. 80 m W is a second, 70diam. X 0.1. A third, 50 m SW of first, is 80 diam.X 0.3. Neo-Babylonian-Achaemenian.

1306 Tell al-Hayyad. 1100 NE X 450, although dividedinto unequal components by the cutting of latercanals, as the air photographs and map show.The SE part of the mound is largest in area butreaches only 4.5 m ht. The central section is 7 mhigh, and the NE part, the smallest, is 5 m. Densepiles of a limited number of types of mass-pro-duced, utilitarian pottery abound all over the site.This, together with the problem of a later over-burden, led me to record types on a presence-absence basis rather than make several intensivecollections. Early Uruk types are most extensivelydistributed, around the entire perimeter and lowerslopes of the site. The large, central summit areaof the SW part of the site is overwhelmingly Jem-det Nasr, although numerous examples of Earlyand Late Uruk types also occur. There is also amuch more limited distribution here of Early Dy-nastic I pottery. All the same types occur on thecentral and NE portions of the mound, but the ad-ditional ht. there reflects smaller, late occupations:Neo-Babylonian-Achaemenian-Parthian. Parth-ian slipper coffins, but probably no real settle-ment, occur on the SW portion as well.

1307 - Perhaps 250 diam. X 0.8, although parti-ally obscured by surrounding dunes. Old Babylo-nian-Cassite.

1308 - 180 diam. X 2.8. Later canals cut throughthe mound have left three distinguishable summitsin an E-W line, the highest on the W. Primarily UrIII-Larsa, limited Old Babylonian.

1309 - 260 NW X 180 X 2.2. 500 m NW is anevident bifurcation in the ancient riverbed, onesinuous channel passing E of this site (and perhapsof Early Dynastic I date), the other W. Akkadian-Larsa.

1310 1,100 NW X 450 X 5. Adjacent SE is asecond, 250 NW X 190 X 3. Much smallermounds also occur on the opposite (right) bank ofancient watercourse passing through the W partof the main site. Many 21 cm sq. bricks. Sasanian-Early Islamic; the 21 cm bricks may suggest asomewhat longer continuation.

1311 280 diam. X 4. Limited Cassite. MostlyNeo-Babylonian-Achaemenian-Parthian. SomeSasanian graves.

1312 - 50 diam. X 0.2. Intensive collection madewithin circle of 5 m radius, N central part of site.Early/Middle Uruk, possibly also a small LateUruk occupation.

1313 - 250 NNW X 150 X 2.4. Sasanian-EarlyIslamic.

1314 300 diam. X 2. Sasanian. Early Islamicrare if present at all.

1315 190 NW X 150 X 0.4, the limits of the siteperhaps slightly reduced by surrounding dunes.Intensive collection made in circle of 5 m radius,center, supplemented by types observed elsewherethat are noted only on a presence-absence basis.Probable Middle Uruk, mainly Late Uruk. JemdetNasr occupation largely concentrated around akiln area in the S part of site.

1316 250 NNW X 180 X 3, with a low NNWextension. Small Uruk and Jemdet Nasr-EarlyDynastic I occupations. Mainly Sasanian-EarlyIslamic.

1317 - 180 diam. X 2. Possible Early Dynastic II/III. Primarily Larsa-Cassite. Relatively few Parth-ian sherds, and these may be only graves or straysfrom adjoining settlement at site 1318.

1318 Tell al-Hayyad al-'Alwi. 300 diam. X 7.Uruk, Larsa-Cassite. But surface materials areoverwhelmingly Neo-Babylonian-Achaemenian-Parthian.

1319 120 NW X 60 X 0.8. Late Islamic-Recent.1320 - Tell Jumayda. 250 WNW X 80 X 2.3.

Cassite-Middle Babylonian.1321 Tell Abui Wadiya. 180 NE X 140 X 1.2. Middle-

Neo-Babylonian. Minor Recent occupation.1322 Tell Abui Dukhun. From Inspectorate of Surveys

records, file 67, register 2126. Recorded as Cassite,Parthian.

1323 Farkh Abu Dukhun. From Inspectorate of Surveysrecords, file 68, register 2126. Recorded as OldBabylonian-Cassite.

1324 - 110 E X 40 X 0.3. Adjoining to the S is a

280

oi.uchicago.edu

General Site Catalog

second, 100 NW X 30 X 0.4. Cassite-MiddleBabylonian.

1325 Ishan Sayyid Ridha, 340 E X 260 X 4.2, Cassite-Middle/Neo-Babylonian.

1326 Mellahe (name taken from Samawa 3E sheet,"Rough Provisional Issue" of British 1/" mapseries, Survey of India Offices, 1916). Ruins ex-tend more than 500 m along the outer curveof a meander in a former watercourse. The onlyelevated part of the site, however, is a mound 250N X 180 X 3 (at both N and S ends) situated im-mediately NE of the apex of the meander. Herethere is a small reed-mat shrine, now in ruins.Trace of Cassite (probably only strays from nearbysite 1325), mainly Sasanian on the mound. Thelarger, lower site is primarily or exclusively Recent,and some Recent pottery naturally also occurs onthe mound. Cf. discussion of this settlement byLoftus (1856, pp. 141-43).

1327 Ishan Danghuz. 150 NE X 110 X 3. Late Islamic.1328 - 140 diam. X 3. Sasanian, possibly also

Early Islamic, Recent.1329 Tell al-Ahmar. 220 NW X 120 X 3.5. 40 m S

across an old canal levee is a second, 180 NW X150 X 2.8. Late Islamic.

1330 - 250 NW X 180, the summit of 3.5 m atSE end. NNE across an old canal bed is a second,120 diam. X 1. 200 m ESE of the second, and alsoon the left bank of the canal, is a third, 300 N X160 X 1. Late Islamic.

1331 100 NW X 40 X 0.2. Old Babylonian-Cassite.

1332 - A large ruined qal'a, the ruins of its fourcorner towers still rising to 4.5 m ht. The accom-panying settlement is 450 diam., although mostlyrepresented by debris at little more than plain level.Recent.

1333 Tell Tabiya. 220 diam. X 7. A ruined qal'ais on N slope of mound, an isolated tower NE.Larsa. Primarily Old Babylonian. Rare Cassite.Recent.

1334 170 ENE X 130 X 0.4. Sasanian, possiblyalso Early Islamic.

1335 - 180 E X 80 X 2. Sasanian.1336 Ishan Burhayniyah. From Inspectorate of Surveys

records, file 211, register 2295. Recorded asIslamic.

1337 Tell Khathale. A prominent qal'a is centeredon this mound, the former an irregular quad-rilateral 55-70 m on a side that rises to a maximumht. of 6.5 m at its NE corner. The underlyingmound is 240 diam. A small Early Uruk settlement.Primarily Early Dynastic II/III-Akkadian, limitedUr III-Larsa-Old Babylonian. A substantial Re-cent occupation.

1338 Sfiq al-Fawwar. Main mound 180 NW X 50 X1.2, littered with fragments of 21 cm sq. (andsmaller) bricks. Walls or wall footings observedin place, giving the impression that a consider-able part of this mound may cover a single largebuilding. To the SE, the site continues at plainlevel for several hundred m more, mixed with andinterrupted by canal spoil banks so that its fulldimensions are hard to determine. This lower partof the site clearly forms a wider band of settlementthan the width suggested by the main mound,however, and has denser accumulations of sherdssuggesting the ordinary accumulation of livingdebris rather than the sparser accumulation in andaround formal architecture. Late Islamic (espe-cially the main mound)-Recent (especially thelower settlement to the SE). Cf. Loftus 1856, pp.142-43 (quoted in Adams and Nissen 1972, p. 80)for a description of the ruins of this town about aquarter-century after its reported abandonment.

1339 - 220 NNE X 110 X 2. Larsa-Old Babylo-nian-Cassite.

1340 Tell Ghanime. 250 WNW X 110 X 2.6, althoughmost of the site is low except for a ruined buildingof baked brick at E end. Larsa-Achaemenian,Recent.

1341 - 80 diam. X 1.5. Vestigial remains of aruined qal'a. Old Babylonian-Cassite, Recent.

1342 200 N X 140 X 1.8. 300 m ENE is a sec-ond mound, 130 diam. X 0.6, with scattered, lowoutcrops of debris across intervening plain. Larsa-Old Babylonian, Recent.

1343 - 180 diam. X 2, with scattered debris atplain level NE and a smaller, lower summit SE.Sasanian.

1344 Muftul Sayyid. 60 diam. X 2, with a solitaryruined tower rising 3 m more. Sasanian, Recent.

1345 120 NW X 70 X 0.9. 40 m SW is a secondmound, 50 diam. X 1. The second mound ismainly Old Babylonian-Cassite, the first is mainlyRecent. Each has slight admixtures of the other'spottery, which are assumed not to represent pri-mary settlement.

1346 - A prominent ruined qal'a, oriented WNW-NNE and about 25 m square. It has three towers,lacking one only on WNW. 40 m S is a mound, 160N X 110 X 0.8. 60 m WSW of the latter is another,200 N X 80 X 2.2. The mounds are primarily OldBabylonian-Cassite, and some of the same potteryis found on the lower slopes of the decomposedwalls of the qal'a-perhaps having eroded out ofthe mud brick. There is also Recent pottery aroundthe qal'a.

1347 - 170 NNW X 70 X 2. 20 m SW is a secondmound, 80 NW X 30 X 0.2. The lower mound is

281

oi.uchicago.edu

General Site Catalog

primarily Larsa-Old Babylonian-Cassite, themain one Recent. A few early sherds also are foundon the main mound, however, primarily in pos-sible canal spoil banks along its lower sides.

1348 - 70 NW X 30 X 0.2, although the site mayextend for a moderate distance to the NE under adune now bordering it in this direction. Larsa-Old Babylonian.

1349 - 180 NW X 100 X 2.3. Larsa-Old Babylo-nian-Cassite. Possible Neo-Babylonian. Recent.

1350 90 diam. X 2. Adjoining NNE and E areareas of low debris of equal size to the mound.SE is a larger but less noticeable mound: 200 diam.,rising in one hummock to 1.3 but mostly 0.7 orless. Neo-Babylonian-Achaemenian, possible Se-leucid-Parthian. Recent.

1351 - 80 N X 30, plain-level debris. Larsa-OldBabylonian-Cassite.

1352 - 200 diam. X 5. Very saline; surface pot-tery limited. Old Babylonian-Cassite sherds werenoted in fields just N of site, but not on mound it-self. Their absence may be only a reflection of thepoor collecting conditions. In any case, mainly Sa-sanian-Early Islamic.

1353 260 NW X 160 X 3. An apparent street 3m in width can be traced running NE from nearthe summit of the mound for about 60 m, and thereis an area of ancient pottery kilns at the NE endof the site. Middle/Late Uruk-Jemdet Nasr.

1354 Tell Imam Sayyid Mohammed. 280 NW X 180 X3.5. On the summit of the mound is a small squareshrine with a conical blue dome and a nearbyattendant's hut. Surface pottery is extremelysparse. Possible Jemdet Nasr, although the fewsherds involved are more likely to have beensecondarily transported from nearby site 1353.Akkadian-Larsa, Recent.

1355 - 250 NW X 120 X 2.5. Poor collecting con-ditions. Middle/Late Uruk-Jemdet Nasr.

1356 Tell al-Dhubay'ah. 160 diam. X 3.5. Possibly aJemdet Nasr occupation. Mainly Sasanian, Recent.

1357 Also known as Tell al-Dhubay'ah. 110 diam. X4.5. Extremely poor collecting conditions; criteriafor dating somewhat doubtful. Late Uruk-JemdetNasr. Recent.

1358 Tell al-Hiz. 130 diam. X 6, with a ruined qal'aon the S end of the mound and perhaps theremains of a small hamlet farther N. Extremelypoor collecting conditions. Trace of Uruk. JemdetNasr-Early Dynastic I. Recent.

1359 - 160 diam. X 1.5. Old Babylonian-Cassite.1360 Ishan al Ahimar. 180 NW X 150 X 3. Middle/

Late Islamic.1361 - 160 diam. X 2.4. 100 m NNW across an

old canal levee is a second, 140 NW X 80 X 2.8.

450 m NW of the first is a third, 80 diam. X 1.Middle/Late Islamic.

1362 130 diam. X 2. Middle/Late Islamic.1363 350 WNW X 250 X 5.5. Middle/Late Is-

lamic.1364 120 diam. X 0.5. Late Islamic.1365 130 diam. X 0.4. Cassite.1366 360 NW, a strip of settlement along an old

canal bed that varies in width from 10 to 160 m.It is mostly low, but one hummock reaches 1.4m. Late Islamic.

1367 Strip of settlement on the right bank of aformer canal, 280 NW X 80-140; mostly low, buttwo summits reach 1.5 m ht. On NE or left bank,directly opposite, is only a very low and narrowstrip of settlement, but it continues SE for 300 mfarther and widens at that point to 120 m. LateIslamic.

1368 - A strip of ancient settlement extending for1,100 m along right bank of levee, varying in widthfrom 80 to 150 m and reaching a maximum ht. of1.5. A small left-bank settlement is directly oppo-site at approximately the midpoint. Late Islamic.

1369 300 diam. X 1.8. Late Islamic.1370 240 diam. X 1.4, with scattered mounds of

moderate size continuing to SE along left bank offormer canal. Late Islamic.

1371 Ruined qal'a, eroded into a shapeless lowmound, and small surrounding settlement. Recent.

1372 Tell Bayjat. 300 E X 90 X 2. 400 m W liesa second, 130 diam. X 0.8; immediately W of it athird, smaller but of the same ht. On the first,highest mound, trace of Uruk; Larsa-Old Baby-lonian-Cassite, Sasanian, Recent. The two smallermounds are Sasanian.

1373 100 diam. X 2.2. There is an equivalentarea of debris at plain level 250 m S. Cassite-Mid-dle Babylonian.

1374 120 diam. X 2.2. Sasanian-Early Islamic.1375 60 NW X 30 X 4. Early ceramics curiously

pulverized, heavily overlain by later debris, pre-cluding an intensive collection. Hence early typesare recorded on a presence-absence basis only. 20m NE is a low and somewhat discontinuous areaof debris, 100 NW X 100 X 40. The main moundis Middle/Late Uruk, Recent. The low area isSasanian.

1376 - 180 NNE X 120 X 1.4. Old Babylonian-Cassite.

1377 - 40 diam., debris at plain level. Seleucid-Parthian.

1378 Ishan Abu Sabkhaya. Five adjoining moundsgrouped in a shallow U open to the NW. Theyvary from 60 to 120 m in diam. and reach amaximum ht. of 3 m at NNE end. Sasanian.

282

oi.uchicago.edu

General Site Catalog

1379 Ishan al-Hutaimiya, but cf. discussion of nam-ing ambiguities in connection with site 1389.220 WNW X 140 X 2.2. Rare Cassite, mainlyMiddle/Neo-Babylonian, rare Achaemenian.

1380 - 150 diam. X 0.6. 300 m NNW of the firstis a second, 100 NW X 70 X 0.6. A third, 200 mNE of the first, is 60 diam. X 0.8. The third isParthian, the first and second are Sasanian.

1381 - 180 E X 120 X 0.5. Larsa-Old Babylo-nian, Sasanian.

1382 180 NW X 130 X 0.4. Sparse pottery. Sa-sanian.

1383 Ishan Bayt. 350 NW X 300 X 3. The substan-tially elevated portion of the mound is smallerthan these dimensions, which reflect the limits of abare, perceptibly raised mound surface set in alevel, grassy plain. To some degree the lateralextent may be exaggerated by slope wash, in otherwords, although progressively sparser sherds andother debris do occur over the entire raised surfaceand seem unlikely to have been moved very faroutward and downward by the same erosionalagency. Probably a small, underlying Uruk settle-ment. Extensive Early Dynastic II/III, less Ur III-Larsa. Trace of Cassite. Mainly Achaemenian-Parthian.

1384 - 170 diam. X 1.5, with a small, ruinedtower on NW end. Sasanian, Recent.

1385 Muftul Abui Duhn. Small, low settlement. Recent.1386 To Walter Andrae at the beginning of the cen-

tury, this was apparently "Abu Howasiduh"(Andrae 1903, map), perhaps now to be tran-scribed Abfi Khowaysidah. There were no localinformants in the vicinity at the time of our visitin 1975, and the continuing use of this name couldnot be confirmed. 180 diam. X 1. Perhaps a smallEarly/Middle Uruk settlement. Jedmet Nasr, pos-sible Early Dynastic I/II/III-Akkadian-Ur III-Larsa. Some Sasanian, Recent.

1387 Probably Abui Khowaysidah; cf. site 1386. 180diam. X 2.4. Late Islamic.

1388 100 E X 70 X 0.8. Late Islamic.1389 Umm al-Khezi. A brief excursus may be appro-

priate here on the inadequacy of names as refer-ence points even for important sites like thisone. We were able to obtain this name only afterasking a number of local informants whose flockswere grazing in the immediate area. In Walter An-drae's time (Andrae 1903), however, this was ap-parently known as Hetime, for a large site withthat name can be located in approximately thesame position on his map, "Umgebung von Faraune Abu Hatab." That name in turn is presuma-bly equivalent to site 1379 in this study, Ishan al-Hutaimiya, the identification given in the ¼" map

series and still current today. Umm al-Khezi, toAndrae (Umm al-Hizzi on his map) was a recentlyabandoned settlement that probably is equivalentto our site 1393. The obvious lesson is that onlygeographical coordinates, not names, constitute apermanent and unambiguous basis for site identi-fication.

670 E X 400 X 2.8, although some of the manyindividual summits making up this site may beslightly higher. It appears that the higher partsof the site are generally toward its peripheries, andthat the outer slopes were in many cases steep-sided. This suggests the possibility of a city wall.Larsa-Old Babylonian-Cassite-Middle Babylo-nian. Cassite (and to a lesser degree, earlier) sherdsare abundant everywhere, but later debris is gener-ally limited to the E portion of the site.

1390 - 180 NW X 140 X 1.9. Old Babylonian-Cassite-Middle Babylonian.

1391 100 N X 70 X 1.8. Late Islamic.1392 240 E X 190 X 2. 150 m SE is a second,

80 diam. X 0.8. There has been recent, shallowpitting for graves on the lower mound. Old Baby-lonian-Cassite-Middle Babylonian.

1393 - (Possibly Andrae's Umm al-Khezi, wherehe noted a ruined qal'a.) 70 diam. X 2.8. Immedi-ately E is an area of low debris, 180 NNE X 60.Some Sasanian, mainly Late Islamic.

1394 Again a confusion of names; cf. site 1389. Thisis probably the mound Loftus identified asDjemideh (1856, p. 143), a name now identifiedwith site 1320. To Andrae a half-century later, onthe other hand, it was Ishan al-Sahin. There wereno local informants in the vicinity from whom toobtain a modern name. 350 ENE X 220 X 2.5.Very extensively, deeply, and recently pitted. Im-mediately W is a second, very small mound. Pri-marily Late Uruk-Jemdet Nasr. A small OldBabylonian-Casste occupation is limited to theNE part of the site. The small mound to the Wis Sasanian.

1395 - 160 diam. X 3. Badly pitted, probably be-cause of its immediate proximity to site 1394. Sa-sanian, Late Islamic.

1396 140 E X 50 X 1.7. 40 m N is a secondmound, 50 diam. X 1.7. Late Islamic-Recent.

1397 - 130 diam. X 2.8. Two additional, verysmall, low mounds lie 120 m SE. The first moundis primarily Old Babylonian-Cassite. The othersare Sasanian, and there is a little Sasanian debrison the first mound also.

1398 - 130 NE X 40 X 0.7. 200 m SSW is a sec-ond, 100 diam. X 0.7. Sasanian.

1399 Tell Qal'a Badyar. 350 diam. X 2. Trace of Uruk.Old Babylonian-Cassite.

283

oi.uchicago.edu

General Site Catalog

1400 260 diam. X 2.2. Old Babylonian-Cassite.1401 160 NNW X 90 X 0.3. Cassite-Middle

Babylonian.1402 Tell al-Hibba. 250 N X 130 X 0.8. A ruined

qal'a on the summit rises 2 m more. There is anabandoned village immediately S of the mound.Pottery sparse. Sasanian-Early Islamic, Recent.

1403 80 diam. X 0.3. Cassite-Middle/Neo-Babylonian.

1404 - 450 NNW X 220 X 2. Old Babylonian-Cassite localized at SSE end. Mainly MiddleBabylonian-Parthian.

1405 - 340 NNW X 240 X 2.4. Trace of Uruk.Old Babylonian-Cassite.

1406 Ishan al-Howa. 280 NE X 140 X 2.5, with aruined qal'a on the SW end rising an additional3.5 m. There is a small mound 40 m SW. An-other, 60 m NE, is 180 diam. X 2.4, with a sayyid'sgrave marked by a blue flag. Sasanian-Early Is-lamic, Recent, the latter concentrated on the firstmound.

1407 Tell Abfi Thayla. 200 diam. X 2, with a smallermound to the S. A ruined qal'a forms a conicallandmark rising 5 m above W end. Sasanian,Recent.

1408 Tell Khayyal. 200 diam. X 1, with a ruined qal'arising an additional 2.5 m on SSE end. Sasanian-Early Islamic.

1409 - 550 NW X 180 X 0.8. The NW half ofthe site is Old Babylonian-Cassite, Sasanian-EarlyIslamic. The SE half was apparently occupied onlyduring the Sasanian-Early Islamic periods.

1410 - 120 diam. X 0.6, with a few higher hum-mocks, and with debris also tailing off 70 m NNW.Trace of Early Dynastic I. Old Babylonian-Cassite.Achaemenian-Parthian, limited Sasanian.

1411 130 N X 70 X 2. Old Babylonian-Cassite.1412 Perhaps 250 diam., mostly at little more

than plain level but with a small mound near Wedge rising 2 m. Limits are partly obscured bydunes. Old Babylonian-Cassite, Parthian.

1413 - 90 diam. X 0.4. Trace of Uruk. Very lim-ited Cassite. Mainly Middle/Neo-Babylonian.

1414 Tell Mu'azzam. 220 diam. X 0.8. A ruined qal'a11 m square, with four corner towers (the largestto the N), rises 4 m above NW end. Sparse pot-tery. Old Babylonian-Cassite, Recent.

1415 - 250 diam. X 2. Sasanian-Early Islamic.1416 - 100 NW X 60 X 0.8. A ruined qal'a at NW

end rises 4 m farther in ht. Owing to heavy ad-mixture of later debris, the early material on thismound was scattered, fragmentary, and perhapsnot wholly representative of the character of theunderlying levels. Hence the enumeration of thecollected sample, drawn from all parts of the

mound instead of from a delimited area, may notbe entirely comparable with other enumerations.Ubaid II, Early Uruk, possibly Middle Uruk,Sasanian, Recent.

1417 Tell Sukheri. About 700 NNW X 200, althoughpossibly a sprawling cluster of tells rather thana single settlement. Individual mounds are 2 m orless in ht. A prominent ruined qal'a, 18 m squareand oriented N, has four large corner towersrising 6 m above the SE end of site. Sasanian-Early Islamic. Limited Recent occupation.

1418 200 NNW X 140 X 0.9. Old Babylonian-Cassite.

1419 180 diam. X 2, although the limited partof the mound with any significant elevation con-sists mainly of brick-kiln debris. Neo-Babylonian-Achaemenian.

1420 300 NW X 150 X 2. Recently and exten-sively, if shallowly, pitted. Early Dynastic II/III-Akkadian, limited Ur III-Larsa.

1421 1,100 NW X 150-200 X 1.3 or less.Slightly sparser debris continues farther NW atlittle more than plain level, directly adjoining andperhaps underlying site 1311. Probable JemdetNasr. Early Dynastic I.

1422 - 240 NW X 120 X 1. A short distancefarther NW are two very small conical mounds,largely consisting of fallen bricks, that rise to aslightly greater ht. Sasanian-Early Islamic.

1423 300 NW X 250, one small mound rising to1.5 m but mostly debris at plain level. Early Dy-nastic II/III-Akkadian-Ur III-Larsa.

1424 - 500 NW X 180 (SE)-220 (NW) X 3.Sasanian-Early Islamic (similarity of surface ma-terials to 1432 noted).

1425 240 diam. X 2.5. Old Babylonian-Cassite.1426 750 WNW X 300 X 3, although W and S

portions of site may be partly obscured by a mas-sive dune belt that overlies (and probably is largelyderived from) levee deposits. Early Dynastic II/III.Mainly Akkadian. Some Ur III-Larsa.

1427 - 200 NW X 140 X 2.8. Sasanian-EarlyIslamic.

1428 140 diam. X 1.2, cut by a later canal. Uruk-probable Jemdet Nasr-Early Dynastic I. Cassite-Middle Babylonian. Limited Sasanian.

1429 - 180 diam. X 2.5, cut by a later canal.Middle/Neo-Babylonian. Limited Parthian pot-tery may only reflect use of the mound as a ceme-tery.

1430 - 220 NW X 180 X 2.5. An apparentlylater canal cuts through the SW portion of thesite. Probable Jemdet Nasr. Early Dynastic I.Cassite-Middle Babylonian. Limited Sasanian.

284

oi.uchicago.edu

General Site Catalog

1431 - 140 diam. X 0.8. The clear, raised tracesof an ancient canal levee that could be unambigu-ously followed from site 1314 through site 1430are lost a short distance above this site. RareCassite. Middle/Neo-Babylonian-Achaemenian-Parthian.

1432 - 500 NW X 350 X 4. 400 m E is a second,much smaller mound, 170 ENE X 100. The latteris mostly about 0.4 in ht. but rises to about twicethis amount in spoil banks around a few scattered,small robber pits. The main mound is Sasanian-Early Islamic. The small mound to the E is Early/Middle/Late Uruk. There are a few Sasanian-Early Islamic sherds on the small mound as well.

1433 230 E X 130 X 0.4. Sasanian.1434 250 diam. X 2.2. Uruk-probable Jemdet

Nasr-Early Dynastic I. Surface material is mainlyCassite-Achaemenian.

1435 - 280 NW X 170 X 3.5. A second, 250 mESE, is 240 NE X 140 (at SW end, tapering tonarrow ribbon of settlement at NE end) X 0.3.The main mound is Sasanian-Early Islamic, lim-ited Samarran. The lower mound is Sasanianonly.

1436 NW arm of site 650 X 220, E arm similarin length and width. Maximum ht. 3 m near theirjunction at an obtuse angle. Sasanian-EarlyIslamic.

1437 280 ENE X 220 X 2.8, reaching thiselevation only at the ENE end. An ancient canalfrom WSW passes directly through the site S ofsummit, apparently having been cut long afterthe abandonment. There has been shallow, recentpitting for graves. Probably a small Uruk settle-ment. Probable Jemdet Nasr and Early Dynastic Iconstitute the major occupation. The smallamount of Achaemenian-Parthian-Sasanian ma-terial may come only from graves.

1438 110 NNW X 30 X 0.6. Neo-Babylonian-Achaemenian-Parthian.

1439 Tulul Abu Fatas. 900 NE X 700 X 6. Neo-Baby-lonian-Achaemenian-Parthian. The massivenessof the site suggests a longer period of occupation,but only sherds typical of these periods were noted.

1440 - Possibly 220 N X 100 0.5, althoughdunes and low, wind-laid deposits obscure mostof it. Uruk, probable Jemdet Nasr, and EarlyDynastic I. Main periods of occupation Old Baby-lonian-Cassite.

1441 Tell Mugnas. 320 NNW X 180 X 3. Sasanian,possibly also Early Islamic.

1442 Tell Abu Qubur. 250 diam. X 2, with a ruinedtower rising an additional 4 m. Trace of Uruk.Ur III-Larsa-Old Babylonian-Cassite. Recent.

1443 Tell Barghasha. 160 N X 120 X 2.8. Possible

small Uruk settlement. Seleucid-Parthian, limitedSasanian.

1444 Tell Swayfi (recorded on British W" map and inInspectorate of Surveys records as Tell Suwaili).150 NNW X 90 X 1.2, surmounted by the ruinsof a small qal'a without corner towers. Sasanian-possible Early Islamic, Recent.

1445 - 250 NNW X 160 X 1.2. Traces of a smallancient canal from the NW adjoin the NE side ofmound, and 30 m away in the same direction isthe outer curve of a possible meander of a muchlarger ancient watercourse from the same direc-tion. Trace of Uruk. Larsa-Old Babylonian.

1446 - 200 diam. X 2. Ur III-Larsa, Seleucid-Parthian.

1447 - 160 NNW X 90 X 2. Achaemenian-Parthian.

1448 - 130 E X 70 X 1. Intensive collection #1made in circle of 5 m radius in center of site.Collection #2 was a supplementary, general col-lection made rapidly from entire mound underconditions of failing light. Middle/Late Uruk-Jemdet Nasr.

1449 250 NW X 180 X 3. Mainly Ur III-Larsa-Old Babylonian, with a reduced Neo-Babylonian-Achaemenian occupation.

1450 Perhaps 250 NE X 200 X 2.8, but sur-rounded and partly covered by dunes. ProbableJemdet Nasr. Early Dynastic I/II/III-Akkadian-Ur III-Larsa-Old Babylonian-Cassite.

1451 - 160 N X 90 X 0.4. Early Uruk, possibleJemdet Nasr. Early Dynastic I. Limited Seleucid-Parthian-Sasanian, these late periods perhapsrepresenting no more than the use of the site asa cemetery for nearby site 1452.

1452 Tell Mrowa'. 240 NNW X 150 X 5, with a ser-rated profile indicating a completely eroded qal'aon its summit. Ur III-Larsa, Seleucid-Parthian,Recent.

1453 120 E X 70 X 0.2. Sasanian, possibly alsoEarly Islamic.

1454 - 160 E X 100 X 1, with an old watercoursefrom the NW immediately adjacent E and then asecond mound, 160 NNW X 40 X 0.2. The firstmound may be the remains of a single building orbuilding complex, for the pottery is found onlyin ridges suggestive of its inclusion in mud-brickwalls, while the relatively more sterile courtyardsand rooms have recently been wind-deflated. Inaddition, baked planoconvex brick walls (or wallfootings) may outline a courtyard 45 X 24 m,oriented at 055°. The low E mound again con-sists of hummocks protected by a surface cap ofsherds while the more sterile areas apparentlyhave been wind-deflated, and at its NNW end

285

oi.uchicago.edu

General Site Catalog

there is much ancient pottery kiln debris. ProbableJemdet Nasr. Early Dynastic I. Considerably lesslate Early Dynastic-Akkadian. There is also muchRecent pottery on the W mound; perhaps the ob-served indications of architecture are only of thatdate.

1455 - 90 N X 40 X 0.4. Possible Neo-Babylon-ian, mainly Achaemenian-Parthian.

1456 140 NW X 80 X 1.8. Probably JemdetNasr. Early Dynastic I. Surface debris is mainlyNeo-Babylonian-Achaemenian-Parthian.

1457 - 180 NW X 130 X 0.4, although NW halfof site is only at plain level. Larsa-Old Babylonian,Recent.

1458 Perhaps 220 diam. X 1-1.5, but size ob-scured by enveloping dunes. Early Dynastic II/III-Akkadian. Primarily Ur III-Larsa. Limited OldBabylonian.

1459 This site is probably the one identified on theBritish ¼4" map as Marauwah, but several localinformants agreed that that name applied onlyto site 1452. No name could be elicited for thisone, but Andrae calls it Dibbin. Size is very diffi-cult to estimate, since the mound is completelybeset by large dunes. An estimate of 350-400 mdiam. is supported by surface evidence and by thearea of slight discoloration on air photographs.Probably at least 6-8 m in ht., but there was noadjacent "plain level" from which to measure.Probable Jemdet Nasr. Early Dynastic I. LaterEarly Dynastic not identified. Surface debris pri-marily Akkadian-Ur III-Larsa. One edge-stampedbrick bore a standard inscription of Amarsuenaof the type widely encountered in the Warka sur-vey (cf. Adams and Nissen 1972, p. 217).

1460 - 20 diam. X 1. Curving meander patternsof riverbed here seemingly pass under this smallmound and hence are antecedent to it. A secondmound of similar size occurs 40 m ENE, nearthe margins of the bed (which, of course, is asequence of beds). Immediately SSE of the latteris an area of sparse debris extending partly overthe bed, 140 diam., plain level. Early Uruk, prob-able Jemdet Nasr, and Early Dynastic I aremodestly represented at and near the secondmound. The first mound and the area of plain-level debris, both of which seem to postdate theexistence of a full-scale watercourse here, aremainly Old Babylonian-Cassite.

1461 - Site is apparently more than 400 diam.,although a very large dune covers much of its NEportion. The most prominent elevation (unlessunder the dune) is 2.2 m, immediately adjacentto the old watercourse along its SW margin. Mostextensive occupation apparently during Old Baby-

lonian-Cassite, continuing, particularly in the Whalf of site, into Middle/Neo-Babylonian-Achae-menian-Parthian.

1462 180 diam. X 2.2. Akkadian-Ur III-Larsa.1463 220 NW X 120, hummocks to 0.6 m, but

mostly plain level. Site may extend farther to Wthan these dimensions indicate, since dune coverin that direction is practically continuous. LimitedOld Babylonian-Cassite. Mainly Neo-Babylonian-Achaemenian-Seleucid-Parthian.

1464 (Identical with WS-001.) Tell Baydha. This sitewas recorded as "entirely surrounded and partlycovered by high dunes" at the time of our visit in1967 (Adams and Nissen 1972, p. 219). In 1975the immediate surroundings were dune-free, sothat repetition of the visit did not become ap-parent until after bearings were plotted. Probablybecause of this removal of covering deposits, Iwould now increase our original estimate of sizefrom 250 diam. x 6 to 350 diam. X 7. The originaldating appraisal was Ur III-Larsa-Old Babylo-nian-Cassite. In addition, recent looters havebrought to the surface many bricks with Seleucidstamps, possibly reflecting the use of the moundat that time only for tombs.

1465 - 240 WNW, along right bank of ancientcanal levee, X 120 X 0.8, with a 200 X 80 ex-tension SSW along branch canal. There are alsosmall settlements on opposite, left bank and fartherdownstream on right bank. Probably a small Uruksettlement. Sasanian-possible Early Islamic.

1466 - 80 diam. X 0.3. Possible Achaemenian,mainly Seleucid-Parthian.

1467 Perhaps 180 diam., but debris at plain levelis sparse so that limits are vaguely defined. Onlya small area of old kilns rises to 0.6 m. Neo-Baby-lonian-Achaemenian.

1468 - Dunes cover most of the lower portionsof this site, which apparently extends more than500 m along both banks of an ancient canal fromWNW. Several individual summits rise to 3 m ormore, and average width must be at least 150 m.Fairly numerous 21 cm sq. bricks. Sasanian-EarlyIslamic, possibly somewhat later.

1469 - Perhaps 240 diam., plain level, hummocksof 0.2 or less. A W extension of site may be cov-ered by dune belt in that direction. Neo-Baby-lonian-Achaemenian-Seleucid-Parthian.

1470 - 250 NNE X 170 X 0.8. Sasanian, possiblyalso Early Islamic.

1471 - 250 NW X 150 X 1.8. Probable small,underlying Uruk and Jemdet Nasr settlements.Mainly Neo-Babylonian-Achaemenian-Parthian.

1472 - Perhaps originally 220 NNW X 110 X 2,but dimensions may have been reduced by recent

286

oi.uchicago.edu

General Site Catalog

excavation of a deep drainage trench alongside.Sasanian-possibly also Early Islamic.

1473 Imam Idris. From Inspectorate of Surveys records:two sites are recorded, presumably adjoining. TellNabi Idris, file 63, register 2126; and Qubat(dome) Nabi Idris, file 64, register 2126. The tellis recorded as Neo-Babylonian, the shrine (pre-sumably on a second mound) as Cassite-Neo-Babylonian.

1474 Suiq Lemlum. This site was not visited but is lo-cated as shown in Inspectorate of Surveys records.Chesney (1868, p. 285) distinguishes between"old" and "new" Lemlum, and it is the latter thatis now designated Suq Lemlum. It was flourishingat the time of his visit in the 1830s, being describedas "a town containing a numerous populationdwelling in prettily constructed reed huts-whichare portable, and which had almost all been re-moved from their usual sites, on account of thefloods, when I first visited this place." On theother hand, a number of travelers' itinerariesmake it clear that it was entirely in ruins by theend of the century.

1475 Tfill Qal'a 'Alaywi. From Inspectorate of Surveysrecords, file 322, register 2622. Recorded as Is-lamic.

1476 Tulul (or Tlayl) Brayj. From Inspectorate of Sur-veys records, file 224, register 2622. Recorded asIsin-Larsa, Cassite.

1477 Qal'a Mohammed. From Inspectorate of Surveysrecords, file 323, register 2622. Recorded as Neo-Babylonian.

1478 Tuilul Rwayha. From Inspectorate of Surveys rec-ords, file 321, register 2622. Recorded as Islamic.

1479 Umm al-Dud. From Inspectorate of Surveys rec-ords, file 303, register 2622. Recorded as Sasanian.

1480 Tell Hib. From Inspectorate of Surveys records,file 313, register 2622. Recorded as Sasanian.

1481 Tell Nabi Ibrahim. From Inspectorate of Surveysrecords: two sites, presumably adjacent, are re-corded. Sharqi, file 311, register 2622; and Gharbi,file 312, register 2622. Both are recorded as Is-lamic.

1482 Ishan Mas'udiyah. From Inspectorate of Surveysrecords, file 325, register 2622. No provisionaldating recorded.

1483 Tell Hadadiyah. From Inspectorate of Surveys rec-ords, file 315, register 2622. Recorded as Islamic.

1484 - Probably 90 diam. X 2.5 originally, al-though partly stripped away during recent roadconstruction. Parthian.

1485 - 180 NE X 70 X 2, but occupational debrismerges with canal banks to the NW and hencemay cover a substantially larger area. Sasanian.

1486 Arth al-Nowayer. An isolated building 85 m

square. It lies immediately E of the Shatt al-Najmi,a former Euphrates branch, and adjoins a trackthat crosses the now-dry bed. The exposed lowerwalls and wall footings are of reused baked brick;the upper walls may have been of mud brick, forthey have entirely eroded away. There are no cor-ner towers. Sparse debris. Recent. Possibly a khanor police post adjoining a formerly importantEuphrates crossing. Reference to well sweeps insite name may suggest a terminal phase of use ofthe channel after the water level had fallen toolow to provide gravity flow into the existing net-work of canal offtakes.

1487 450 WNW X 180 X 3, although rising tothis ht. only in five small, distinguishable summitsnear WNW end. Sasanian, possible Early Islamic.

1488 300 diam. X 2.4. Parthian.1489 Tell Hsayn. 110 N X 40 X 3.5. Parthian.1490 - 70 diam. X 2.5, with a low outlier im-

mediately NNE. Seleucid-Parthian.1491 Tell Hujayl. 140 diam. X 2.5. Parthian.1492 - 90 NE X 40 X 1.2. Sparse surface pottery.

Sasanian.1493 Tell Shahal. 80 diam. X 5.5, its uppermost levels

recently and inexplicably bulldozed away. Im-mediately N is a second mound, 60 diam. X 2.5,and there is third, still smaller one 20 m S. Pot-tery is rare, especially on the main mound, andit may have been an artificial construction ratherthan a gradual accumulation of occupational de-bris. Sasanian.

1494 - 80 diam. X 0.7, with an approximatelyequal area of debris at plain level extending E.Sasanian, probably also Early Islamic.

1495 600 WNW X 300 X 2.5. Sasanian.1496 120 NW X 50 X 1. Much brick-kiln

debris. Sasanian, Recent.1497 80 diam. X 0.6. Sasanian.1498 Two adjoining 60 diam. X 2 mounds along

a NE axis, with a third, smaller one farther NE.Possible Neo-Babylonian-Achaemenian. MainlySeleucid-Parthian.

1499 - 220 NNE X 150, rising at the N end to4 m. 150 m W are the isolated remains of anancient brick kiln. 100 m SE is a second mound,200 diam. X 2. Seleucid-Parthian.

1500 - 600 ENE X 100-160, rising to 2.5 m ht. atENE end only. Site consists of occupational debrismixed with canal spoil banks. Immediately SE ofNE end, probably across an ancient canal, is asecond 120 diam. X 2. Parthian.

1501 Tell Hnayta. 220 diam. X 2.5. Very extensively,almost continuously covered with shallow pitsapparently directed at shallow graves. MainlyEarly Dynastic II/III, continuing on a smaller scale

287

oi.uchicago.edu

General Site Catalog

into Akkadian-Larsa times. The graves that seemto have been the focus of the looters' attention areSasanian.

1502 - 120 diam. X 2.8. Larsa-Old Babylonian,Sasanian.

1503 - 100 diam. X 4, not including a 2 m conicalpile of earth erected as a survey benchmark. Onthe summit are the remains of a qal'a 25 m square,the walls almost entirely eroded away. Ur III-Larsa, Achaemenian-Parthian, Recent.

1504 - 200 diam. X 2, with a 2.5 truncated coneof large, probably reused baked bricks suggestinga fairly recent shrine. NE across an old canal leveeis a second mound, 280 NW X 100 X 2.6.Sasanian-Early Islamic.

1505 450 NW X 130 X 2.5. Late Islamic, Re-cent.

1506 - 450 NNW X 100 X 2.2. Late Islamic,Recent.

1507 - Two very small but considerably elevated(2.5 m) mounds, one about 250 m ESE of theother. 300 m farther SSE is a third, 120 NW X80 X 1.8. The first two are Sasanian-Early Is-lamic, Late Islamic. The third is Late Islamic only.

1508 - 70 diam. X 2. Larsa-Old Babylonian. Pri-marily Cassite. Limited Sasanian, perhaps onlyreflecting use of the mound at that time as acemetery.

1509 - 420 NW X 220 X 2.4. Trace of Ur III-Larsa. Achaemenian-Parthian, possibly alsoSasanian.

1510 - 120 diam. X 1. Traces of Larsa-Old Baby-lonian-Cassite debris, possibly not representingprimary settlements. Mainly Sasanian.

1511 130 diam. X 0.8. Seleucid-Parthian.1512 100 diam. X 0.9. Seleucid-Parthian.1513 Tell Ibris. 130 E X 50 X 5, although these di-

mensions are somewhat conjectural owing to thedestruction accompanying recent canal-diggingaround the mound. Old Babylonian-Cassite, Re-cent.

1514 110 ENE X 30 X 1.2. 30 m SE is a secondmound, 60 diam. X 1. Sasanian, Recent.

1515 Tell al-Zelaga. 150 diam. X 4. 300 m SSW is asecond, 120 diam. X 1.5. Trace of Cassite, mainlyParthian, Recent at the main mound. Recent onlyat the smaller mound.

1516 - 160 diam. X 5, with a small, low moundadjoining ENE. Seleucid-Parthian.

1517 - 160 diam. X 3. Seleucid-Parthian.1518 Ishan al-Turrma. A low mound at least 250 diam.,

with an arm of debris extending SW around oracross a broad bay or channel. The most prom-inent part of the site is a conical summit 30 diam.X 9 that served as a survey benchmark. Sasanian.

1519 200 E X 160 X 3. Sasanian.1520 180 E X 120 X 4, reaching this ht. only

near W end. Sasanian.1521 Imam Abfi al-Fadil. The shrine, built of reused

brick, is about 6 X 6 m. It is in ruins, most of thedome having fallen in, but the rubble has beencleared from the interior, and green and whiteflags on the grave there suggest that it is stillperiodically in use. The underlying mound is 80NW X 60 X 3, with modern graves to the NWof the shrine. 180 m NW is a second mound, 180N X 120 X 3. 100 m NE, across a depressionsuggestive of an old watercourse, is a third mound,110 diam. X 1.5. Sasanian, possible Early Islamic.Recent pottery is plentiful only around the shrine.

1522 90 diam. X 1.4. Parthian.1523 200 diam. X 2.5, with an ancient E-W

canal passing through the N part of the site.Parthian-Sasanian.

1524 - 140 NW X 100 X 2.5. Sasanian.1525 Tfilul Tabbirat. 180 diam. X 4, with broad, an-

cient watercourse beds flowing around both sidesof the site from the N and NW and meeting 100 mSE. Sasanian, possible Early Islamic, Recent.

1526 - A ruined qal'a, its NE tower still standingto a ht. of 5 m, lies at the W end of a mound 130E X 80 X 1.2. Farther E, possibly across the bedof a former watercourse, is a mound 220 NNW X120 X 3.5. Possible Old Babylonian. Mainly Cas-site. Some Recent pottery also.

1527 Tfilfil Tabbirat. 190 NW X 120 X 5. PossibleParthian. Sasanian.

1528 150 WNW X 100 X 2.5. Old Babylonian,Recent.

1529 Tfilfil Jezzaz. This may be alternatively inter-preted, either as a half-dozen or more separatemounds strung out for 2 km along an E-W canallevee (as shown on base map) or as a more or lesscontinuous strip settlement for this distance. Sur-face pottery declines in density at plain-level in-tervals between mounds but is present throughout.Sasanian.

1530 Main mound 140 E X 110 X 2, but threeother, only slightly smaller mounds occur ESE,SE, and S, giving the group a NW axis of about900 m. Sasanian.

1531 - 70 NNW X 40 X 0.4. Recent.1532 Tflil Jezzaz. 240 NW X 140 X 4, seemingly

composed in very large part of slag from glassfurnaces. There are great numbers of irregularchunks of glass adhering to clay furnace lining,colorless or in green, brown, blue, and violethues. 400 m ESE is a second mound, 180 diam.X 2.5. Much glass slag occurs here also, but witha greater proportion of chunks of glass without

288

oi.uchicago.edu

General Site Catalog

furance-lining adhesions, as well as with greaternumbers of glass vessels of various kinds andwith more occupational debris. Sasanian-EarlyIslamic. It seems possible that we have here theremains of a specialized glass-producing unit anda nearby settlement where glass artifacts weremanufactured.

1553 Tfiluil Jezzaz. 280 diam. X 3, immediately E ofsite 1532 and with almost as much evidence ofglass production. Sasanian-Early Islamic.

1534 Tfilful Jezziz. Extends irregularly for about 1.5 kmENE along an ancient canal levee. Average widthappears to be about 230 m. Some elevation ofdebris is continuous, and many individual moundsrise to 2 or 3 m. Toward the ENE end is the mostprominent mound in the group, 50 diam. X 4,with the well-preserved bases of three glass kilnsforming its summit. Kiln debris and other evidenceof extensive glassmaking is very plentiful on allparts of the site. Sasanian, probably also EarlyIslamic.

1535 Four adjoining mounds along an ENEline, all less than 100 diam. X 1. Sasanian.

1536 - 140 E X 90 X 1.8. Sasanian-probableEarly Islamic.

1537 - Only a ruined tower 5 m in diam. survivesof a former qal'a. There is a small surroundingsettlement. Recent.

1538 Qal'a 17 m square, with towers at N andS corners, the latter still standing to a ht. of 6 m.To the NW is a mound 60 diam. X 0.6. Recent.

1539 - 80 diam. X 0.6. Sparse debris includes anancient brick kiln rising to 1.4 m. Sparse debrisalso occurs at intervals along an ancient ESE canallevee passing just S of site, and there is anotherapparent kiln 200 m S. Sasanian.

1540 240 NW X 140 X 0.1, low, irregularclusters and hummocks of fairly dense pottery anddebris. Cassite-Middle Babylonian.

1541 - 160 diam., mostly low but with the SWportion reaching a ht. of 1 m and one ENE hum-mock rising 2 m. Much brick-kiln debris. Sasanian,also a little Recent pottery.

1542 100 diam. X 3 in the NE, elsewhere 2 mor less. Brick-kiln debris. Parthian, possibly alsoSasanian.

1543 - 200 diam. X 2.5. Sasanian.1544 - 160 diam. X 0.8, bisected by an old canal

levee from NW. Sasanian.1545 - 140 NE X 80 X 0.2. Sasanian.1546 Ishin al-Soda. From Inspectorate of Surveys rec-

ords, file 317, register 2622. Recorded as Islamic.1547 Ishan Tammah. From Inspectorate of Surveys rec-

ords, file 316, register 2622. Recorded as Parthian-Sasanian.

1548 220 diam. x 0.5. Sasanian.1549 Tuilal al-Ajjaz. The E portion of this prominent

landmark and settlement is somewhat irregularbut may be described as 300 diam., with steeplysloping sides and with hillocks on its elevatedcentral portion rising to 6 m. To the W is a rela-tively lower area, apparently crossed by canalsfrom the N, then a W mound 600 N X 120 thatagain rises in hillocks to 6 m. Minor outlyingsettlements occur along an old canal levee lead-ing SSE, at distances of 300 and 1,100 m. Surfacetraces of major construction can be seen in manyplaces on the E mound. The most massive, on itsW end, appears to be a single monumental build-ing covering an area of at least 110 X 42, withwalls up to 3 m thick. Parthian-Sasanian.

1550 140 NE X 100 X 1.5. Salt encrustationssuggest mud-brick walls of a major ancient build-ing or buildings, covering much of the surface ofthe mound. Parthian.

1551 90 diam. X 0.5. Seleucid-Parthian.1552 240 NW X 180 X 1.2. Seleucid-Parthian.1553 A high, solitary, ruined tower on a mound

90 E X 20 X 0.7. Seleucid-Parthian, Recent.1554 - Ruined qal'a 16 m square, its largest (NW)

corner tower 5 m in diam. and still standing to aht. of 5 m. The underlying mound is 120 NE X80 X 1. Neo-Babylonian and/or Achaemenianprobable. Mainly Seleucid-Parthian, Recent.

1555 300 diam. X 2.4, sloping gently up to amaximum ht. at the N end. Site boundaries arewell defined, with suggestion of a continuouscontour break possibly arguing for an outer wall.Extensive if largely superficial plundering of shal-low graves-or graves made shallow by wind ero-sion of the mound summit. Ur III-Larsa. Thegraves on which the looting has been focused areSasanian-Early Islamic, but there is no suggestionof a contemporary settlement here to accompanythem.

1556 The most easily defined part of this site isa low mound at its E end 160 N X 120 X 0.6,but a ribbon of debris at plain level also extendsWNW for at least 800 m. Sasanian, possible EarlyIslamic.

1557 - 190 diam. X 3. Sasanian.1558 - 140 diam. X 1. Seleucid-Parthian.1559 - 240 NW X 100 X 2, bisected by an an-

cient canal bed from NW. Many glass kiln wasters.Sasanian-Early Islamic.

1560 - 500 NE X 150-250 X 1.5. Possible Parth-ian. Mainly Sasanian.

1561 - 80 diam. X 0.5. Sasanian, Recent.1562 - 120 NW X 60 X 0.3. Recent.1563 - 100 diam. X 1. Ur III-Larsa, Parthian.

289

oi.uchicago.edu

General Site Catalog

1564 120 NE X 50 X 0.3. Recent.1565 250 diam., no perceptible elevation, yet

with fairly dense and continuous surface pottery.Sasanian.

1566 - 170 WNW X 60 X 0.6. 200 m WNW is asecond, smaller mound. Sasanian, Recent.

1567 60 diam. X 0.4. Parthian-Sasanian.1568 120 diam. X 2, flattopped and steep-sided,

with the apparent remains of a defensive wallaround the top of the site. 140 m WNW are theruins of a large qal'a. Sasanian, probably alsoEarly Islamic. The wall on the summit of themound and the qal'a are Recent.

1569 - At least 200 diam. X 0.3, although dunesmay obscure further extensions W and N. 500 mESE is a second, 300 diam. X 0.8 but somewhatirregular in outline. Other low, irregular areas ofdebris are 200 m NW and NE of the second. Thefirst mound is Parthian-Sasanian. The remainingmounds are Sasanian, with small, scattered Recentsettlements as well.

1570 - 100 diam. X 1. 60 m S is a second, 160diam. X 0.8. Mainly Cassite-Middle Babylonian.There is a thin and limited Sasanian overlay.

1571 230 diam. X 3.8. Trace of Uruk. Cassite-Parthian. Small Recent occupation.

1572 160 diam. X 4.2, this ht. including a smallconical mound recently constructed as a surveybenchmark. Immediately W is the deep outletchannel of a substantial lake, and SW across thischannel is a second mound, 140 NNW X 60 X0.8. The shoreline of the lake at this point is about2 m above present bed level, although it has beenirregularly cut away by severe wind erosion. Thedifference in elevation declines progressively asone follows the shoreline NE and declines moreslowly as one follows it W. The lake bed is veryflat and almost featureless, although low hum-mocks of sand have formed around shrubs andsmall bushes. It is almost continuously coveredwith shells in areas where these aeolian depositsare not present. The mounds adjoining the outletare mainly Seleucid-Parthian, with small Sasanianand Recent occupations also. Probably the lake hasa longer and more complex history than this sug-gests, for it will be observed that archaeologicalsites are seemingly entirely absent in an area ofnearly 100 sq. km N, NW, and W of these mounds.On the other hand, relatively faint traces of an-cient canal levees can be followed through muchof this unsettled area. That seems to argue that thelake was a long-lived but not continuously main-tained geological feature, its bed periodically dryenough to afford irrigable tracts for agriculturewhile generally too subject to flooding to attract

settlement. The outlet channel also may have acomplex history. Its present course has the irregu-larity of a natural stream bed and yet follows ageneral course suggestive of a former canal. Per-haps it was an artificially maintained outlet at thetime of the major occupation of these mounds buthas been uncontrolled since Parthian or Sasaniantimes and hence modified in channel characteristicsby periodic natural overflows.

1573 - 200 NE X 40 X 0.2. Mainly Sasanian; alittle Recent pottery.

1574 - 20 diam. X 0.4. Cassite-Middle Babylo-nian.

1575 - 270 WNW X 180 X 0.6, with a lower ex-tension or immediately adjoining mound SW thatis 160 diam. Cassite-Middle Babylonian. SmallRecent occupation.

1576 30 diam. X 0.4. Seleucid-Parthian.1577 60 diam. X 2.6, with a well-defined chan-

nel below plain level curving around the N and Emargins of the mound. On its summit is a qal'a 20NW X 17, with a single corner tower on the S.Parthian-Sasanian, Recent.

1578 80 diam. X 1.2. Sasanian.1579 230 N X 180 X 2.2. Ur III-Larsa.1580 120 WNW X 30, low hummocks of debris

that include graves being exposed by wind ero-sion. A canal levee passing through this site can betraced ESE to where it overlies a much earlier,meandering watercourse discussed in connectionwith site 1590. Sasanian.

1581 160 diam. X 0.2, the S half at plain levelexcept for scattered hummocks. Sasanian, Recent.

1582 - 220 diam. X 0.3. Numerous small out-croppings of contemporary debris occur at inter-vals to ESE along old canal levee. Sasanian.

1583 130 E X 90 X 0.5. Old Babylonian-Cas-site-Middle Babylonian. Small Sasanian, Recentoccupations.

1584 Low hummocks of sparse debris within anarea 80 diam. Adjoins the S bank of an old canallevee from the WNW, with which it must be con-temporary. Cassite-Middle Babylonian.

1585 - 180 E X 70 X 0.5, with a shallow depres-sion suggesting that the site was bisected by anold canal from the NNE. Possible Achaemenian,mainly Seleucid-Parthian. Trace of Early Islamic,some Recent.

1586 - 70 diam. X 0.6. Sasanian, Recent.1587 - 120 NW X 70 X 2. Relatively recent, me-

andering watercourse channels flow around mostof the site and meet S of it. Seleucid-Parthian,Recent.

1588 - 200 NW X 110 X 2. Shallow looting of

290

oi.uchicago.edu

General Site Catalog

graves has resulted in a hummocky surface appear-ance. Parthian-Sasanian, Recent.

1589 - 20 diam., low hummocks similarly situ-ated to site 1584. Cassite-Middle Babylonian.

1590 A very small cluster of debris on the Sbank of an old levee, similar to site 1589. It hasa compact, well-defined, white-crusted bed 8 mwide, while the associated disturbance zone (pre-sumably spoil banks) is 17 m in total width. Thisbed is perpendicular to, and clearly overlies, amuch wider, meandering bed. The latter, consist-ing of integrading ridges and channels that canbe traced in wind-erosion patterns, bands markedby slight differences in the density of minute frag-ments of shell, and linear arrangements of lowshrubs, has a width of 75 m. The underlying water-course, presumably following an essentially "nat-ural" regime, could be followed only a shortdistance upstream from this point because of over-lying wind-laid deposits accumulating around lowshrubs. Downstream, however, its meanderingcourse could be followed to the NW foot of site1591, with bifurcating channels below there appar-ently continuing to sites 1600, 1601, and 1628. Onits right bank was a rather clear and uniformlycompacted, whitish band about 6 m wide that mayhave been a minor canal paralleling the generaldirection of the meandering watercourse, but thiswas soon lost beneath later wind-laid deposits. Onthe right bank of this 6 m canal, 120 m SW ofthe superimposed canal crossing, is a small pile of30 X 16 X 6 cm baked brick with sparse associ-ated pottery. The overlying canal from the WNW,following a straight and presumably "artificial"course, could be continuously traced past sites1584, 1589, 1590, 1592, and to within 300 m ofsite 1593, and more intermittently followed up-stream beyond site 1584 to site 1570. The smallcluster of debris here is Ur III-Larsa, and the samedating applies to all the settlements served by theunderlying, meandering watercourse. The over-lying, straight canal from the WNW, on the otherhand, serves a string of small sites that are all ofCassite-Middle Babylonian date.

1591 - Boundaries poorly defined, but at least 300WNW X 150 X 2. Bricks 24 X 17 X 8. MainlyUr III-Larsa, probably continuing on a small scaleinto the Old Babylonian period.

1592 A ribbon of settlement on the right bankof the same levee from the WNW as that identifiedin sites 1584, 1589, and 1590. 120 WNW, very lowand narrow. Trace of Uruk. Cassite-Middle Baby-lonian.

1593 - A roughly square qal'a, oriented N andmeasuring 27 m on a side. A well-defined meander-

ing channel below plain level curves around the Eside of the fort, but to the W and N the low moundunderlying it extends for 40 m. Some Sasanian. Themajor occupation seems to have been Recent.

1594 50 NNW X 20 X 1. There is a second, verysmall mound immediately NNW. Cassite-MiddleBabylonian. Minor Sasanian, Recent occupations.

1595 - Indistinct mounds, including much brick-kiln refuse, run NW along both sides of an ancientcanal levee for about 500 m. The strip of debris is180 wide at NW end, declining to 100 at SE. Par-thian-Sasanian.

1596 Two adjoining mounds on a NE-SW axis,each about 120 diam. X 2.2. Both are largelycomposed of brick-kiln slag and cinders. Parthian-Sasanian.

1597 Imam Mizhir. The name applies to a smallshrine of reed mats on one of the smaller, lowerNNE summits of the site. The site itself is 350diam., rising to 4.5 m in four large and manysmaller summits. Additional debris occurs at plainlevel for at least 250 m SE. Mainly Parthian, withan apparently thin Sasanian overlay. But the plain-level debris appears to be Sasanian, not Parthian.

1598 70 diam. X 0.6. 120 m WNW, across abroad, old canal levee, is a second, 40 diam. X0.3. Seleucid-Parthian, some Recent.

1599 90 E X 50 X 1. Mainly Cassite-MiddleBabylonian. Small Sasanian, Recent occupations.

1600 - A slightly curving ribbon of debris, 110 min length and in most places 10 m or less in width,on the outer, W bank of a meander loop of theformer watercourse identified in site 1590. Ur III-Larsa.

1601 15 diam. X 0.2, situated at the apex of ameander loop as in site 1600. Ur III-Larsa.

1602 - 120 diam. X 0.1. Trace of Uruk. Cassite-Middle-Neo-Babylonian. Trace of Parthian.

1603 180 NNW X 120 X 2. Cassite-MiddleBabylonian. Small Sasanian, Recent occupationson E.

1604 A scatter of sherds at plain level within anarea perhaps 100 diam., save that density onlydiminishes gradually toward the periphery so thatthe limits of the site are somewhat arbitrary. Pre-Ubaid and/or Ubaid I. See separately publishedaccount (Adams 1975a),

1605 - Two adjoining clusters of debris at plainlevel, that to the N 40 diam., that to the S 20 diam.Cassite-Middle Babylonian.

1606 Alwat Kred. Perhaps 300 or 350 diam. X 1.1,but surrounding canal spoil banks and limitedsurface pottery owing to high salinity make sizeestimate difficult. Low, salt-encrusted swellingsthat may have been outlying settlements occur at

291

oi.uchicago.edu

General Site Catalog

intervals to the NE and SW. Probably in the Se-leucid-Parthian-Sasanian range; the very limitedcollection makes greater specificity impossible.

1607 Tell Igdayr. 250 diam. X 1.8. Seleucid-Parthian.1608 Ishan Abfi al-Dinayn. 180 diam. X 4.5, with a

small outlying mound 250 m ESE. Very saline;collection sparse. Seleucid-Parthian.

1609 Imam Nabi Sulayman. Large (39 X 32) green-domed shrine of baked brick, surrounded bymodern cemetery, standing on the N shoulderof a mound 240 NNW X 80 X 2.4. Seleucid-Par-thian.

1610 300 NE X 200, although with appreciableelevation (to 1 m) only in SW. Saline encrustationssuggestively outline very substantial constructionto the NE, with walls up to 2 m thick and anapparent courtyard 60 m square. There may be asingle monumental building 140 NE X 110. Cas-site-Middle Babylonian.

1611 110 diam. X 0.6. Sasanian.1612 160 NE X 130 X 0.6. Well-made 29 X 29

X 7 bricks. Trace of Uruk. Mostly Cassite-MiddleBabylonian surface pottery, although the brickssuggest a continuation into Neo-Babylonian times.Sparse Sasanian pottery may all be from graves.

1613 140 NNE X 50 X 2. Mainly Cassite-Mid-dle Babylonian. Possible Parthian. Minor Sasanianoccupation.

1614 220 N X 120 X 1.2, although rising to 2.2in a conical earth mound at the N end built as asurvey benchmark. Sasanian.

1615 80 diam. X 0.8, the margins of the moundblending imperceptibly with the plain under con-ditions of high salinity. Owing to poor collectingconditions an intensive collection within a delim-ited area was impractical, and types were tabu-lated only on a presence-absence basis. One ormore pottery kilns can be seen in place on the Eside of the mound. Middle/Late Uruk. JemdetNasr and Early Dynastic I are limited to the E andSE portions of the mound.

1616 60 diam. X 0.8, with lower debris extend-ing S and SW for 100 m. A depression that is prob-ably a former watercourse bed curves around theN and E portions of the site. Parthian-Sasanian.

1617 - 110 NE X 40 X 1, although this estimatemay be inflated by adjacent canal levees. Saline;limited collection. Sasanian.

1618 - 120 diam. X 1.5. Very saline; poor collect-ing conditions. But pottery at plain level extends20 m W and up to 180 m E and NE of the mound.Trace of Uruk. The plain-level debris is primarilyUr III-Larsa, while the mound is Sasanian.

1619 - 180 diam. X 0.7. Sasanian.1620 - 280 WNW X 120 X 2.4, either two dis-

tinct settlements or one bisected by an old canalbed. The higher end of the site, to the ESE, ismainly Parthian, some Sasanian. The WNW endmay be only Sasanian.

1621 - 160 diam. X 0.8. Much slag and cinders,suggesting an ancient brick kiln. Sasanian.

1622 220 ENE X 160 X 0.6, with a few small,slightly higher hummocks. Sasanian.

1623 180 diam. X 1.2, bisected by an old canalbed from WNW. Scattered, low clusters of debrisalso occur SSW. Sasanian.

1624 - 180 N X 100 X 0.8. Some Parthian, mainlySasanian.

1625 Tell al-Mfishaur. 200 NW X 130 X 3. 120 mNNW is a second, 140 NW X 80 X 1.8, tailingoff farther NW through fields that have beenabandoned fairly recently for lack of adequatewater. Equidistant to the SW is a third smallmound, so that the three form roughly an equi-lateral triangle. Saline, hence limited collection.Possible Neo-Babylonian. Mainly Seleucid-Par-thian.

1626 150 diam. X 2, with a smaller, lowermound immediately E. Parthian.

1627 A strip settlement along an old canal levee,extending about 800 m ESE. There are three prin-cipal nuclei of settlement, where the site expandsin width to 100-150 m and rises to 1.5 m ht.Sasanian.

1628 210 diam. X 2. There has been old, exten-sive looting of shallow graves, the pits now largelyobliterated by wind-laid overlay. Collecting condi-tions are poor due to the latter. Parthian slipper-coffin fragments and Sasanian storage jars are themain visible surface components, but the absenceof other associated ceramic types argues that thesereflect only the late cemetery rather than a settle-ment. Hence the early component is likely to be theprimary one, difficult as it was to establish: Ur III-Larsa.

1629 250 diam. X 2, topped by a conical moundof apparently recent construction for cartographicsurvey. An irregular but continuous series of muchsmaller, lower mounds begins 150 m S and contin-ues S for 600 m, apparently defining an old canallevee. Sasanian.

1630 -- A sparse scatter of surface pottery within anarea 25 m in diam. Both Ur III-Larsa and Cassite-Middle Babylonian occupations can be definitelyidentified in spite of the limited amount of surfacematerial.

1631 - 800 NW, irregular widths up to 120 m. Twosummits at SE end are 2.8 high, that at NW is 2.0m. Achaemenian possible. Parthian-Sasanian.

1632 -- An agglomeration of low, irregular hum-

292

oi.uchicago.edu

General Site Catalog

mocks that can only be roughly described as 180diam. X 0.8. Similar, but smaller and more scat-tered, hummocks occur at intervals for about 1km S along old canal levee. Sasanian.

1633 160 diam. X 1. Trace of Uruk. Sasanian.1634 90 NW X 50 X 0.8. 300 m SSW is a sec-

ond, 50 NW X 30 X 0.7. Parthian.1635 Imim Sayyid Sh6ka. A newly rebuilt but re-

portedly very old shrine, now of baked brickand with the traditional blue dome. 300 m SWare the remains of a village that must have beenabandoned fairly recently, but an older settlementor one in the immediate vicinity of the shrine isnot in evidence. The local topography is obscured,however, by dunes surmounting an old canal leveefrom the NE.

'1636 Haush Umm al-Shah'ir. Two adjoining mud-brickqal'as; cf. plan and brief descriptive note in Finsterand Schmidt (1976, pp. 167-68). The underly-ing site merges with ancient levees and hence isdifficult to define, but a broad ribbon of debrisextends 250 m W, and there are also substantialoutcroppings to the N. Most of the site is lessthan 0.5 m in elevation, particularly to the E,but under the forts themselves there may be 1.5m of cultural accumulation. Sasanian, Recent.

1637 - 40 diam. X 1. Some Parthian, probablymainly Sasanian.

1638 - 80 diam. X 0.8, with graves and scattereddebris at plain level extending about 50 m NW andSE. Seleucid-Parthian. There are numerous bakedbricks bearing triangular Seleucid stamp impres-sions.

1639 - 90 NE X 30 X 0.3. Cassite-Middle Baby-lonian.

Other Relevant Sites in the Region

Listed below are sites initially cataloged as part of theWarka survey (Adams and Nissen 1972) that were visitedfor the purpose of making new, intensive sherd collec-tions for dating. Also listed and briefly described are a fewsites that were originally visited in connection with themuch earlier Akkad survey (Adams 1972) and that werenot revisited as a part of this study.

118 - Description as given in The Uruk Country-side, but dating assessment given therein supple-mented by intensive collection within a circle of 5m radius on the center of the main mound. EarlyUruk.

125 - Description as given in The Uruk Country-side, but dating assessment given therein supple-mented by intensive collections: #1, circle of 5 mradius about 100 m N of SE end of mound; #2

similar circle, NW central part of same mound;types observed only elsewhere on the mound arerecorded in the same listing on a presence-absencebasis. Probable Middle Uruk. Maximal extent ofsettlement in Late Uruk. Jemdet Nasr. Substan-tially reduced occupation in Early Dynastic I.

126 - Description as given in The Uruk Country-side, but dating assessment given therein supple-mented by intensive collection within a circle of5 m radius, NE central mound. Traces of a verylimited occupation in the Ubaid II/III periods werelocalized in a small part of the NW quadrant of themound. Primary occupation Late Uruk. JemdetNasr probable but uncertain. Early Dynastic I oc-cupation extremely limited.

128 Description as given in The Uruk Country-side, but dating assessment given therein supple-mented by intensive collections within two circlesof 5 m radius: #1, center N mound; 2, center Smound. Probably Middle Uruk. Maximum extentof settlement in Late Uruk. Jemdet Nasr occupa-tion probable but uncertain, Early Dynastic I occu-pation extremely limited. See separately publisheddiscussion of a group of nearly identical, bifaciallyworked flint implements found here during restudy(Adams 1975b).

245 Tell Mismar. (See Schmidt 1978.) Physical de-scription remains as given originally, but furtherdetails on dating are now available. Two intensivecollections were made for this purpose on the Nmound in the group, both within circles of 5 mradius. Collection #1 was made near the SE end,#2 on the N end near a substantial outcropping ofclay cones with heads idented for pigment. Un-systematic observations on the S mounds in thegroup indicate a primarily or exclusively JemdetNasr occupation. Ubaid-Early/Middle/LateUruk. Major occupation Jemdet Nasr, althoughthe character of the cones identified the publicbuilding on the surface of the N mound as proba-bly coeval with Eanna X in the Uruk excavations(J. Schmidt, pers. comm.), Early Uruk. Much re-duced occupation in Early Dynastic I. Sasanianburials confined to NW part of N mound, wherethey have been disturbed by recent shallow pitting.A kiln on the SW end of the site probably is alsoSasanian, although early kiln wasters on N moundargue for local ceramic production also in lateprehistoric times. There are also a few Recentburials.

A221 - The central mound in the group shown iskidney-shaped, roughly 250 NW X 100 X 4. Thedeep main drainage canal of the Mussayib projecthas been cut through the E end of the site, provid-ing an exposed section of occupational debris 5-6

293

oi.uchicago.edu

General Site Catalog

m below plain level (Harris and Adams 1957). Thecentral mound is Uruk, Akkadian-Larsa, Par-thian-Sasanian. The surrounding mounds in thegroup are Parthian-Sasanian only.

A259 -.. 120 NW X 50 X 1. SE half includes a pos-sible small Uruk settlement, mainly Larsa-OldBabylonian. NW half Middle Islamic.

A261 - 150 NNW X 70 X 3, bisected almost toplain level by an old canal cut that is now used asa roadbed. To the W is a very large, low settle-ment, 700 WNW X 150, and there are also smallmounds to the S. Probably a small, underlyingUruk site at the first mound, although it is mainlyLarsa-Old Babylonian-Cassite. The large, low set-tlement is Early/Middle-possibly Late Islamic,while the small mounds to the S are Sasanian.

A264 - 220 NNW X 120 X 2. 20 m SW is a sec-ond, 80 diam. X 1.5. The second, smaller moundis Early/Middle Uruk, Akkadian, Old Babylonian.The larger one may begin in the Old Babylonianperiod but is mainly Cassite, Sasanian.

A265 100 diam. X 3. Neo-Babylonian-Achae-menian.

A266 Tfuluil Abfi Adham. An irregular mound groupof about eight major summits within an area,1,300 ENE X 500. Neo-Babylonian-Achaeme-nian-Parthian.

A273 100 diam. X 2. Adjacent SE is a second, 75diam X 1.5. Cassite.

A274 - Multiple small summits form a dispersedgroup tending NNE for more than 1 km. Neo-Babylonian-Achaemenian-Parthian.

A275 Tell Abu Salabikh. A carefully detailed con-tour map of the site is given in Postgate andMoorey (1976, fig. 1). The two connected eastmounds rise to a height of about 5 m above plainlevel and cover approximately 28 ha. To the west,across a depression that may be an ancient riverchannel, is a 7 ha mound about 1 m lower inheight. South of the latter is a so-called Urukmound of perhaps 12 ha, 0.5 m lower still. Apresence-absence tabulation is given in table 6 fora surface collection made in 1957 by the authorand Dr. Vaughn E. Crawford.

Probable Ubaid. Early/Middle/Late Uruk-Jemdet Nasr-Early Dynastic I/II/III. It seemslikely that there was a progressively increasing sizeof settlement until the late Early Dynastic period.Some later (Proto-Imperial or Sargonic) potterywas found in a drain, leading the excavators tospeculate that there may have been a terminal Ak-kadian or even later building phase that is appar-ently unrepresented in the surface pottery now tobe found on the site (Postgate and Moorey 1976,p. 157). On the other hand, a small number oflater sherds (two large-spouted bowls, a channel-rim bowl, and a sherd with horizontal ribs, all ofAkkadian or slightly later date, were included inthe 1957 collection. While there may have beena shrine or some other commemorative structureerected here, this seems to imply an at most verylimited post-Early Dynastic occupation.

294

oi.uchicago.edu

Appendix

The Southern Marginsof Sumer

Archaeological Survey of the Areaof Eridu and Ur

Henry T Wright

INTRODUCTIONMethods of WorkGeographical Notes

THE SEQUENCE OF CERAMIC ASSEMBLAGES

Eridu and Hajji MuhammadEarly UbaidLate UbaidUrukJemdet Nasr and Early DynasticAkkadianUr III and Early LarsaLate Larsa and Old BabylonianCassite and Post-CassiteNeo-Babylonian, Achaemenid, and Later

THE DEVELOPMENT OF SETTLEMENT IN THE AREA OF UR AND ERIDU

Early Villages and Towns of Southern Sumer (Ubaidand Before)The First States (Uruk to Early Dynastic)The First Empires (Akkadian to Post-Cassite)The Death of a Region (Neo-Babylonian and Later)Concluding Remarks

SITE CATALOG

295

oi.uchicago.edu

Figures

1 Base Map of the Southern Sumer Survey Area, 19662 Ceramics of the Eridu, Hajji Muhamad, and Ubaid

Periods3 Ceramics of the Uruk Period4 Ceramics of the Jemdet Nasr and Early Dynastic

Periods5 Ceramics of the Jemdet Nasr and Early Dynastic

Periods6 Ceramics of the Jemdet Nasr and Early Dynastic

Periods7 Ceramics of the Akkadian Period8 Ceramics of the Ur III and Early Larsa Period9 Ceramics of the Ur III and Early Larsa Period

10 Ceramics of the Late Larsa and Old BabylonianPeriod

11 Ceramics of the Late Larsa and Old BabylonianPeriod

12 Ceramics of the Cassite and Post-Cassite Periods

13 Ceramics of the Cassite and Post-Cassite Periods14 Ceramics of the Neo-Babylonian and Later Periods15 Ceramics of the Neo-Babylonian and Later Periods16 Ceramics of the Neo-Babylonian and Later Periods17 Settlements of the Late Ubaid Period and the Dis-

tribution of Clay Sickles Found on Later Sites18 Settlements of the Late Uruk and Jemdet Nasr Periods19 Settlements of the Early Dynastic I Period20 Settlements of the Ur III and Early Larsa Period21 Settlements of the Late Larsa and Old Babylonian

Period22 Settlements of the Cassite Period23 Settlements of the Post-Cassite Period24 Settlements of the Neo-Babylonian and Later Periods25 Patterns of Settlement and Population in Southern

Sumer26 Selected Settlement Maps

296

oi.uchicago.edu

Introduction

The archaeological survey of southern Sumer was plannedearly in 1965 as one step in the survey of lower Iraq, aproject conceived and largely carried out between 1958and 1975 by Robert McC. Adams of the University ofChicago's Oriental Institute. By 1965 only the Diyalaplains and the region of ancient Akkad on the northernalluvium of the Tigris and Euphrates had been surveyed,and the reader may well query why we chose to jump tothe extreme south for the next survey. We made thischoice because we wanted a better understanding of set-tlement systems of the Ubaid period. It was in the souththat the only complete Ubaid stratigraphic sequence,that at Eridu, was known. Furthermore, a variety of re-ports indicated that the Eridu depression, partially sep-arated from the main alluvium by a sandstone ridge,might have a preserved early land surface on which bothlarge and small Ubaid sites were exposed without manymeters of obscuring later silt: (1) maps indicated that theplain around Eridu was several meters lower than thealluvium near Ur; (2) sterile sand at Eridu itself and at thenearby Ubaid site of Rejibah X was reportedly at plainlevel or even higher; and (3) a brief survey by Safar hadlocated several small Ubaid sites not far from Eridu. Withgood survey evidence we could relate the patterns of tech-nological specialization and social differentiation-im-plied by existing evidence from excavations-to patternsof population growth and land use as indicated by surveyevidence. Understanding of the Ubaid would have pro-vided a firm foundation for work on the later urban de-velopments so well evidenced to the north around Warka.However, as is often the case with initial surveys, we werewrong. Geological changes had obscured most of theearly land surfaces, and our data proved more relevant tolater historical periods than to earlier ones.

Support for the fieldwork was provided in part by theOriental Institute in the form of a new Land Rover and asmall budget for travel and labor, and in part by a gradu-ate fellowship from the National Science Foundation thatpaid the daily expenses of my wife, Fran Wright, and my-self. We arrived in Baghdad on 10 October 1965 and spent

the greater part of the next three months there, studyingmaterials from Eridu and related sites. The facilities of thenew Iraq Museum were made available to us by Dr.Faisal al-Wailly, then director-general of antiquities, thelate Sayyid Fuad Safar, then director of excavations, andthe late Dr. Faraj Basmachi, then director of the museum.Through most of our stay we resided in the BritishSchool of Archaeology through the kindness of Dr. DavidOates and Mr. Jeffrey Orchard. I profited then andthroughout the subsequent years from the advice of Dr.Joan Oates regarding the problems of prehistoric archae-ology in Iraq.

On 20 January 1966 we arrived at the excavation houseat the ancient city of Ur, 15 kilometers southwest of themodern city of Nasiriyah. For the next three months wecrisscrossed the areas around Ur, always accompanied byeither Hajji Hosseyn ibn Aboud or Sayyid Mahsen ibnNais, the site guards of Ur. The few official problems thatarose were handled by the directorate's representative inNasiriya, Sayyid Wa'il al-Ruba'i, and by Capt. Sabih al-Khafajeh of the national police. Much interest in our workwas expressed by Shaykh Muhammad ibn Manshet of theAl-Ghizzi. We were also visited and helped by the en-gineers of Energoinvest, Inc., of Yugoslavia and of theSwiss Engineering Consultants. During the actual ar-chaeological survey work, Mr. Nicholas Vester andMr. Julian Wootten helped us for several days. Thisreport has profited from close reading by Robert McC.Adams, John R. Alden, Curtis E. Larsen, and MatthewW. Stolper. Those failings that remain are entirely theauthor's.

METHODS OF WORK

The survey was conducted under difficult conditions:we had neither accurate maps nor air photographs nor so-phisticated mapping equipment. The base map for thesurvey was built up by triangulation from known pointson the Baghdad-to-Basra Railway, angles being measuredwith a Brunton compass. To our relief, this base map

297

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

checked well with the air photographs we were able toexamine after completing the survey.

Each day we visited from two to eight sites, the numberdepending on their size and their proximity to Ur. Oursweeps by Land Rover-on parallel lines from 0.5 to 1kilometer apart, topography permitting-were such thatwe should miss no site more than 0.5 meter high. Some ofthe recorded sites were lower, and only on the southwest(where dunes were encroaching and travel was difficult)could we have missed many. We mapped each site withcompass shots and pace measurement, noting any evi-dences of buildings, burials, kilns, or other features as themap was prepared. We noted the distribution of bakedbricks on the assumption that their abundance indicatedmore prosperous settlements or parts of settlements. Othermembers of the survey group walked back and forthacross the site at roughly 5-meter intervals collecting allbases, rims, and other diagnostic sherds. If the site waslarge or topographically complex, it was divided into sev-eral units, each separately collected. On very large sites,particularly to the southwest, time limitations forced usto compromise even this simple sampling approach. Thecollections resulting from this general collecting procedurewere then examined in order to assess the last major occu-pation of the site, and the site was searched again spe-cifically for diagnostic indicators of earlier, deeply buriedoccupations. Any such early artifacts were individuallylocated on the site map. With this procedure we surveyedthe area from the limits of modern cultivation on thenortheast, typically congruent with the northeast flank ofthe ancient Euphrates levee running through Ur, to theedge of the sand dunes and salt flats on the southwest,typically close to the southwest of the flank of ancientEuphrates levee swinging south of Eridu. On the north-west, we recorded sites around the railway station and vil-lage of Bat-ha, and on the southeast we recorded sitesaround the railway station and ancient mound of Tellal-Lahm. In this area of about 1,010 square kilometers werecorded 192 sites, many of which were in fact groups ofmounds. We also made trips along the modern Euphrateslevee to the north of the survey area and deep into thedesert to the fortress of al-Qusair (Finster and Schmidt1976), but in no case was such survey complete. Sites ofinterest seen on these trips will be noted in passing in thetext rather than included in the site catalog.

Each evening and each rainy day, we washed and num-bered the sherd samples and checked them against suchpublications as were available to us in the field. Fortu-nately, there are many small sites with short occupations,and before long recurrent associations of ceramic typesbecame apparent. The Ubaid and earlier phases were fa-miliar because of our study of the Eridu ceramics in theIraq Museum. Some ceramics of the Uruk, Jemdet Nasr,Early Dynastic, Agade, and Neo-Babylonian-Achaeme-nian periods at Ur had been published, and it was possible

to recognize these. Cassite and Ur III ceramics are rela-tively uniform throughout alluvial Mesopotamia, and wecould identify them with the help of photographs of Nip-pur ceramics given to us by Adams. In the field, our prob-lem was the early second millennium B.C.: the Isin-Larsaand Old Babylonian periods. Our tentative dating of twoassemblages to these times was confirmed by Fuad Safarduring a visit to Ur; comparable assemblages have onlyrecently been published. As the fieldwork progressed andour maps grew, consideration of our tentative datings ledto one conclusion: the preservation of land surfaces of thefourth millennium B.C. and earlier was poor because amajor southern channel of the Euphrates flowed throughthe Eridu depression during the second and first millenniaB.C., covering earlier surfaces almost everywhere. Afteruseful discussions with Fuad Safar, Robert and LindaBraidwood, Bruce Howe, and Halet Cambel-all ofwhom visited us at Ur for several days-and after muchthought, I decided to put aside work on the Ubaid periodand focus on the settlement system of the Early DynasticI period. The further survey work and small excavationthat became the basis of my doctoral dissertation (Wright1969) do not concern us here.

Upon returning to Baghdad in late May 1966, we be-gan the analysis in earnest. We were kindly given resi-dence in the home of Dr. Theresa Howard Carter.Through the prolonged efforts of Sayyid Mahmud al-Amin, we were able to examine air photographs of theEridu area. Dr. Basmachi offered us a large room for alaboratory. We prepared a final base map (fig. 1) and be-gan detailed recording of the sherds. We devised a tax-onomy based on clay body, clay color, and form andtabulated the sherds from each sample. The sherds werethen packed in a sturdy wooden case and stored in theIraq Museum, only the tabulations being sent to Chicagofor further analysis, a procedure that (for better or worse)prevents any revision of our typology. Final pottery typedescriptions, tables, illustrations, and period maps wereprepared in Chicago during the autumn of 1966. But thedemands of completing the dissertation and subsequentlypreparing for a new job intervened, and the survey reportwas set aside until time became available.

I had intended to use this time to undertake a numberof studies this present report does not contain. First, I hadhoped to undertake a statistical study of the ceramic as-sembleges using one of the newly developed seriation orordering techniques (e.g., Hole and Shaw 1967; Drennan1976). However, since the approximate relative dating ofthese ceramic assemblages is not in doubt, such a studywould serve primarily as an illustration or test of theavailable methods. This I leave to others. Second, I hadhoped to undertake a detailed study of the cuneiformsources, both the few pieces we found during our surveyand those published from Ur and other sites that wererelevant to the geography of the area. This might be very

298

oi.uchicago.edu

/

/"I

//

/

Fig. 1. Base map of the southern Sumer survey area, 1966.

299

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

useful, but I lack the full range of skills necessary for thistask, and the number of fixed points-settlements whoseancient names are known-is limited. Such a study mightbe more successful if it included a broader area. Third, Ihad intended to undertake a formal locational analysis ofsome of the settlement patterns. This plan foundered onseveral points. In the first place, the average length of thedefined ceramic periods is about three hundred years; it isrisky to assume that the sites on any map were all occu-pied at the same time. In the second place, the settlementpatterns are conditioned by the linearity of rivers andcanals, and it would be necessary to adapt or develop aspecial locational theory to undertake such an analysis. Inthe third place, although our survey area is naturallybounded on the south and east, it is arbitrarily cut off onthe west and north. Until very recently data on these areashas been unavailable.

The present report is primarily a descriptive document,presented in such a way that the data may be reworkedand integrated with that from other areas of alluvialMesopotamia. Broad interpretation is kept to that mini-mum necessary if the reader is to appreciate the specialfeatures of southern Sumer that condition any use towhich the data might be put. One should not forget thatthis survey used elementary methods and that improvedresurveys must be done in the future, using refined ceramicchronologies to solve the problems raised by this and otherstudies of ancient Mesopotamian settlement patterns.

GEOGRAPHICAL NOTES

Some areas of the Mesopotamian alluvium, for exam-ple, parts of the Hilla and Marsh areas, are today similarto what they were in past millennia. Today, however, noriver flows by Ur or Eridu, and what were once irrigatedlevees and freshwater marshes are shifting sand dunes andsalt-crusted mud flats. Thus, even if we wished to do so,we could not extrapolate directly from the modern en-vironment to past conditions in the area. Nevertheless,something can and must be said regarding geology andpossible past geography if we are to understand the an-cient settlement patterns.

Mesopotamia is on a tectonic boundary where the Ara-bian plate pushes against and slides under the Eurasianplate. Compression has folded the former sea bed betweenthe two plates, raising the Zagros-Taurus mountain arc.Sediments eroded from these mountains have been de-posited in the Tigris-Euphrates valley and beneath theGulf, isostatically forcing the earth's crust farther down-ward and compressing the earlier sediments in the result-ing basin. The area of interest to this discussion is on thesouth edge of this basin, where the layers of sediment risegently upward to the south, lying directly on the rela-tively stable parts of the Arabian shield.

The bedrock geology of the immediate area of Ur and

Eridu is obscured by the present alluvium. In general, thegeological strata dip gently down to the northeast. Theedge of the alluvium southwest of Eridu is an outcrop ofcherty limestone. Northeast of Eridu and southwest of Uris another outcrop, this one of gypsiferous Dibdibba sand-stone. This exposure rises and broadens to the southeast,where it conformably overlies a series of earlier strata;these in turn rest unconformably on the aforementionedcherty limestones of the Dammam formation. Thus theplain of Eridu is a cul-de-sac surrounded on three sidesby bedrock outcrops. The gypsiferous sandstone ridge,locally known as the "Hazim," that separates the Eridualluvium from the main Euphrates alluvium has beendated to the Pliocene by al-Naquib (1967). There are sitesof the late fifth millennium B.C. associated with its presentconfiguration. On the other hand, desert varnish is nota-ble on pebbles only above an elevation of about 40 metersabove sea level. In this area, pebbles with desert varnishoccur on land surfaces with Middle Paleolithic tools dat-ing to between 40,000 and 100,000 years ago. Thereforethe flanks of the ridge must have been exposed morerecently than the upper portion.

Another feature that may be very old is the line oflarge dunes more than 50 meters high that lies along thesouth edge of the alluvium, covering its border with thecherty limestones. A site of the early fifth millennium B.C.exists on these dunes; thus much of their bulk must haveaccumulated in earlier times. As far as I know, there areno studies of the minerology of the dune constituents.Whether their source was the desert to the south, the al-luvium itself exposed by the rivers downcutting in re-sponse to lower sea levels in times of glacial maxima, orboth is not yet known.

In short, it seems that the Hazim and the dunes hadassumed more or less their present form as sea level ap-proached its present height about 4000 B.C. (Larsen 1975),flooding the incised channels of the Tigris and the Eu-phrates and much of the Mesopotamian basin, thus cre-ating the present-day Gulf. Concomitant natural leveeformation would have begun along the many braidedchannels of the river. As soon as the levee on the south-ern channel of the Euphrates had reached the Ur area,the northwestern open end of the Eridu plain would havebeen obstructed and would have begun to receive flood-season discharge from the Euphrates. In years of majordischarge through more permanent breaches in the levee, alake or marsh may have been formed, with waters findingan outlet through a breach in the Hazim. This occurredat a point between sites EP 131-33 and EP 134-35, about14 kilometers east-northeast of Eridu. Since EP 134 has anumber of Late Ubaid sherds, it seems likely that thisoutlet had been scoured to more or less its present formby 4000 B.C. Since this date, we can propose that theEridu depression has been subjected to drastic cycles ofenvironmental change as a function of local changes in the

300

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

Euphrates channels. If water moved down the southern-most Euphartes channel into the Eridu depression, a fresh-water lake would form southeast of Eridu and drainthrough the break in the Hazim. Salt would be removedfrom the basin, and ground moisture would increasealong the river channel. Poplar, willow, and tamariskwould thrive on the riverbanks, and on the levees abrushy cover of Prosopis would survive in uncultivatedareas. The lake itself would probably support a marshyvegetation, predominantly reeds such as Phragmites,which now surround the Hor al-Hammar. When thesouthernmost Euphrates channel ceased to flow, the lakewould become a salty mud flat, with only a few salt-lov-ing, drought-tolerant plants surviving on the former leveesand channel bottoms. On the dunes only the Halyxylon,a small tree, would flourish, protected from woodcuttersby kilometers of barren alluvial desert. Eventually more

northerly Euphrates channels would become obstructed,and water would again shift into the southern channel,initiating another cycle.

This natural cycle in the Eridu basin has been deducedfrom a minimum of assumptions and must be verified bygeological surface mapping and leveling, the study of sedi-ment cores, and other techniques. It is possible that suchstudy would reveal that this proposed cycle never ran itscourse, since after 4500 B.C. human communities weresufficiently numerous that their efforts to control theflow of water would affect the distribution of vegetationand deposition of silt layers. Before this interaction ofnatural and cultural factors can be sketched, somethingmust be said regarding the archaeological chronologyupon which our understandings of community and chan-nel shifts are based.

The Sequence of Ceramic AssemblagesThis section presents the bases for dating surveyed sites.

It covers ceramic developments in this area from about5000 B.C. to at least 400 B.C. It is based on a preliminarystudy of apparently briefly occupied sites in southernSumer, sorted from the universe of 192 sites located dur-ing 1966.

A presentation like this is necessary because the avail-able published material is biased in favor of unbrokenand unweathered vessels from graves, while surface col-lections generally yield weathered sherds from domesticareas. A study based on the manipulation of surface col-lection statistics is bound to have errors, but until morestratified sherd samples are carefully reported in print,this approach will have to suffice.

In the pottery descriptions I attempt to use a simpleterminology. A "period" is defined on the basis of a ser-ies of surface collections in which a number of types occurtogether. Within periods, "complexes" are defined byseries of surface collections in which either a few typescluster together or a type is unusually common. A com-plex may represent a portion of a period, or it may repre-sent a ceramic group with special functions. When moreand better samples are available, this rather artificial di-vision will be unnecessary.

The terms for parts and attributes of vessels, and theircombinations into named types, are tentative. Each ce-ramic type is designated by a code beginning with theletters RJ for rim of jar, RB for rim of bowl, B for base,and 0 for other. Following the letters is a number indi-cating the period of most common occurrence. The num-ber 1 represents the earliest Eridu and Hajji Muhammadperiods, and 10 represents the latest Neo-Babylonian and

later periods. Next follows a hyphen and a number seri-ally assigned for each variant. Thus, RB8-2 is the seconddescribed bowl rim type of the Late Larsa-Old Babylo-nian period, a type that also happens to occur during thepreceding Early Larsa period. Each type is also given aninformal descriptive name. In the sections for each periodthere is first a general statement about the ceramics of theperiod, then a list of type descriptions. Each descriptionincludes the type number, the name, a figure reference,and any other observations. The site catalog gives thecounts of types from a series of collections. This allowsthe reader to examine patterns of the co-occurrence oftypes in "periods" and "complexes." Were this study tobe redone, I might well present site-by-site corpora asAdams and Nissen did in The Uruk Countryside (1972);thus presenting maximal information on ceramic varia-tions and allowing complete reevaluation when ade-quately described stratified ceramics are available for thearea. However, since the sherds were recorded typologi-cally, they must be presented in this way.

THE ERIDU AND HAJJI MUHAMMADPERIODS

Ceramics of these early periods are not well representedin our surface collections, and, since Lloyd and Safar(1948) and Oates (1960) have presented an extensivedescription of these materials, great detail is not neces-sary here.

Eridu period ceramics have a temper of medium-sizedsand particles, a yellow to buff body, and a similarlycolored slip. The most common form among the exca-

301

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

vated rims from the Eridu Temple Sounding is a small tomedium-sized hemispherical bowl. Their exteriors char-acteristically have a motif of parallel oblique paintedlines. In the earlier layers of this period, the bowl interiorsare usually decorated with pendant triangles; in laterlayers of this period (XVI to XIV) grids and checker-boards become common. The only Eridu sherd recoveredin the survey is an overfired example with the latercheckerboard interior (fig. 2a), although other sherds areknown from Ur and Tell Ubaid (Oates 1960, n. 18).

Hajji Muhammad period ceramics have a similar tem-per and slip, but usually they are harder and have a buffto green color. The paint is matte or glossy. There ismuch variation in vessel shape in the excavated sam-ples, but only fragments of the distinctive carinated bowlshave been recovered in the survey samples. These mediumto large shallow bowls have a broad outflared rim. Theexterior painting is usually bold zigzag bands; the riminteriors usually have an oblique crosshatch (fig. 2b),and base interiors have grids (fig. 2c) or arrangements oftriangles (fig. 2d). As will be discussed shortly, a simplervariant of this bowl continues into Early Ubaid times andis more common in the survey collections than are theseclassic forms.

THE EARLY UBAID PERIOD

Surface sherds of this period are found mixed with LateUbaid pieces, but they can be recognized by comparisonwith sherds from the Eridu Temple Sounding layers Xand IX. These ceramics have a fine sand temper or novisible temper and vary in color from buff to green. Themost frequent surface indicator of Early Ubaid occupa-tions is the carinated bowl with simple black bands(RB2-1).

RJ2-1: Interior-ledge-rim jar (fig. 2h). In this periodthese have a solidly painted rim and a complex shouldermotif of curved or crosshatched lines.

RB2-1: Carinated bowl with simple black Bands (fig.2e). These are thinner and less sharply carinated thanearlier Hajji Muhammad bowls, and the paint is matte,never glossy. The base interiors often have a rosette motifwith lines radiating from a central dot (fig. 2f). Un-painted variants occur.

RB2-2: Medium-sized deep bowl with interior tri-angle motif (fig. 2g).

RG2-3: Flat-lip basin with interior or rim-top tri-angle motif. This is similar to RB2-2 but heavier, larger,and with a flattened rim. Plain and simple banded vari-ants are common. All these variants may also occurduring the preceding Hajji Muhammad phase.

RB2-4: Bell-shaped cup. Many variants of these finevessels, both plain and painted, are known from Eridu(Lloyd and Safar 1948, pl. III).

THE LATE AND TERMINAL UBAID PERIOD

The ceramics of this period have a paste and colorsimilar to those of the Early Ubaid, but there is moregreen color, warping, and surface sludging from the over-firing of salt-impregnated clays. Many vessels exhibit thesurface scoring characteristic of wheel finishing.

Sites with dominant surface assemblages of this periodare recognizable from the fragments of painted sherds,particularly of incurved bowl rims (RB3-7 to 10) andledge rim bowl rims (RB3-11 and 12) and quantities ofclay sickle fragments (03-1), though the latter also con-tinue into later periods. A study of statistics on the ceram-ics from Eridu by G. L. Barnes (pers. comm.) shows thatat least two complexes are separable. The Late Ubaidcomplex is known from Eridu Temple Sounding, levelsVII and VI, and Hut Sounding, level V. It has high-neckedjars with completely painted (RJ3-3) or banded (RJ3-2)necks; plain forms are rare. Among the medium-sizedbowls, the ledge rim (RB3-11, 12) is more common thanthe incurved form, and both frequently have lip and in-teror motifs such as oblique lines, curved lines, or tri-angles. The Terminal Ubaid complex (cf. Wright et al.1975, p. 140) is known from the Eridu Hut Sounding,level IV. The high-necked jars have unpainted collars(RJ3-1). Among the medium-sized bowls, the incurvedform seems to be more common than the ledge-rim form,and both have solid painting or bands rather than geo-metric motifs. More than 35 percent of the incurved formshave impressed combing on the interior (RB3-10). Simplebowl forms are frequently unpainted. A deep basin withexterior ridges also seems to be typical (RB3-15, 16).

RJ3-1: High-neck jar, unpainted collar. These rangefrom medium to large. The bodies of these vessels areprobably hand-built and finished, but the rims may havebeen wheel-finished.

RJ3-2: High-necked jar, painted bands (fig. 2i). Theillustrated example is relatively large.

RJ3-3: High-neck jar, completely painted collar. Thisform is rare, occurring in both Early Ubaid and the be-ginning of Late Ubaid at Eridu.

RB3-1: Hemispherical bowl, unpainted. This formhas a simple rounded lip and is usually relatively shallow.None of the variants are useful chronological indicators.These and other bowls are apparently wheel-finished.

RB3-2: Hemispherical bowl, painted band.RB3-3: Hemispherical bowl, painted interior decora-

tions. Doubtless future work will reveal chronologicallydistinctive variants.

RB3-4: Small deep bowl, unpainted. Complete ex-amples from the cemeteries at Ur (Woolley 1956, pls. 17-19) and Eridu (Lloyd and Safar 1948, pl. III) indicatethat this form typically had a ring base.

RB3-5: Small deep bowl, painted bands.RB3-6: Small deep bowl, painted motifs (fig. 2/j). On

302

oi.uchicago.edu

I ]

4

A

ILLiTIHU"

._ -

II

r.IB

Em. - =

E AMm.

C

L 12D

E

F

H

(K

N

M* * .I- -~- 4v

Fig. 2. Ceramics of the Eridu, Hajji Muhammad, and Ubaid periods (2/5). (N.b.: B and P indicate Munsell color readings forclay body and paint respectively, when available. D indicates diameter when relevant.)

Eridu period bowl, EP-104.Hajji Muhammad period carinated bowl, B: 2.5Y 4/2, P:10YR 2/0, EP-104.Hajji Muhammad period bowl interior motif, EP-3.Hajji Muhammad period bowl interior motif, EP-104.RB2-1, carinated bowl, B: 5Y 6/3, P: 10YR 3/1, EP-29.RB2-1, bowl interior motif, B: 5Y 7/2, P: 5Y 3/1, EP-38.RB2-2, deep bowl, B: 5Y 4/1, P: 10YR 3/1, EP-141.RJ2-1, interior-ledge-rim jar, B: SY 7/4, P: 10YR 2/1,EP-141.

i. RJ3-2, high-necked jar, B: 5Y 5/4, P: 2.5Y 2/0, EP-141.j. RB3-6, small deep bowl, B: 2.5Y 5/4, P: 10YR 2/1, EP-104.

k. RB3-10, incurved bowl with flat lip, combed interior, B:SY 7/4, P: 10YR 3/1, EP-104.

1. RB3-12, ledge-rim bowl, B: 5Y 5/4, P: 7.5YR 2/0, EP-29.m. RB3-13, deep flat-lip basin, B: 5Y 5/4, EP-98.n. 03-1, clay sickle, B: 5Y 5/4, EP-98.o. 03-2, clay muller, B: SY 6/4, EP-98.

303

a.b.

C.

d.e.f.g.h.

'1A~l ýW;F MW AWM W

A A A

~I~·~L-';rC\·=L-~.~)---~L~·~__C~LC

I

G7.

o

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

the exterior are horizontal bands, between which areplaced variously oriented triangles, zigzags, grids, andother rare motifs. No sample is large enough to allowstudy of the chronological and spatial variations in thesemotifs.

RB3-7: Incurved bowls with flat lip, unpainted.RB3-8: Incurved bowls with flat lip, interior and ex-

terior painted bands.RB3-9: Incurved bowls with flat lip, painted motifs.

These are usually curved or zigzag lines below the interiorband.

RB3-10: Incurved bowl with flat lip, combed in-terior (fig. 2k).

RB3-11: Ledge-rim bowl, completely painted rim.RB3-12: Ledge-rim bowl, painted motifs (fig. 2/),

triangles, zigzags, and oblique lines are all found on theledge rims.

RB3-13: Flat-lip basins, unpainted.RB3-14: Flat-lip basin, painted bands. These are rare

and may be a final development of the similar formsknown from the Early Ubaid period (RB2-3).

RB3-15: Deep flat-lip basin with exterior ridge, un-painted (fig. 2m).

RB3-16: Deep flat-lip basin with exterior ridge,painted. The few painted examples have two horizontalbands enclosing curved lines.

03-1: Clay sickles (fig. 2n). The maximum use ofthese items in this area seems to be during these periods.However, the typical smaller form continues later, and sothe presence of small sickle fragments in a collection isconsidered only as suggestive of Ubaid cultivation. Notethat many stray clay sickle finds are close to Ubaid sitesbut far from Uruk sites and therefore are more likely tobe Ubaid. A number of our examples exhibit retouch onthe blades.

03-2: Clay muller (fig. 20). These are often termed"bent clay nails" in the literature. Our examples, like theexcavated examples from Ras al-'Amiyah (Stronach 1961,p. 107), often show signs of use as a pestle. These occurboth before and after the Late and Terminal Ubaidperiods.

03-3: Chert hoes. Bifacially flaked implements similarto certain varieties of late Acheulian bifaces.

URUK PERIOD

Except where noted, ceramics of this period survivingon the surface have a distinctive temper of fine roundedsand and coarse crushed rock, either calcite or quartz.This grit constitutes 30 percent of the paste. The bodyis generally dark green or gray.

Sites with dominant surface assemblages of this periodare marked by simple straight-rim jars (RJ4-1), beveled-rim bowl rims (RB4-1), large strap handles (04-1, 04-2),and large clay sickles. Even when weathering has removed

most distinctive ceramic items, a scatter of rough, fire-cracked chert flakes and limestone rubble suggests Urukoccupation, for Uruk ceramics were often associated withsuch debris.

RJ4-1: Simple straight-rim jars (fig. 3a-c): These aredistinguished by simple horizontal combing on the bodyexterior and by frequent attachment of 04-1 and 04-2handles.

RJ4-2: Thick-rim jars (fig. 3d-e): Distinguished bythe triangular section of the low outcurved rim.

RJ5-4: Flared expanded-rim jars (fig. 4j, k): These arerare.

RB4-1: Beveled-rim bowl (fig. 3f, g): Distinguishedby predominant gray brown color, straw temper, andfinger marks resulting from manufacture in a press mold.

RB5-2: Conical-bowl rim (fig. 6a-c).04-1: Twisted strap handle (Fig. 3h): Distinguished

by flattened section and relatively large size. These arelater Uruk.

04-2: Plain strap handle (fig. 3i): Distinguished byflattened section, double grooving, and relatively largesize.

04-3: Twisted lug handle (fig. 3j): Distinguished byfiner twistings and horizontal placement on the vessel.

04-4: Clay cones (fig. 6j-l): Rare. One example has ahollow head.

05-2: Spouts (fig. 5g-j). Similar to later forms.04-5: Droop spout: These are rare and perhaps late

in this period. Distinguished by long conical spout turneddownward at a marked angle. Often there is a slightlythickened rim on the spout.

04-6: Stamp-decorated sherds: These are rare andperhaps early. Distinguished by either individual cres-centic impressions or rocker impressions on the upperbody of a small jar.

04-7: Combed sherds: These are common. Distin-guished by complex overlapping spirals and curves, oftenpartially smoothed over. Simple combing continue intolater periods.

04-8: Large clay sickles: Very large forms withhandles thicker than 1.9 cm and blades wider than 5.5 cmmay be more typically Uruk in the Eridu area.

JEMDET NASR AND EARLY DYNASTICPERIOD

Except where noted, ceramics of this period have in-clusions of fine sand and very finely crushed rock, gen-erally calcite. The latter is generally observable onlyunder a microscope. The paste is often buff to reddish,but greenish overfired examples occur. Jars, large bowls,and the upper portions of conical bowls are often coveredwith a cream slip or wash.

Within this period, three complexes can be isolated.One is distinguished by rarity of decorated types, the

304

oi.uchicago.edu

BA

DC

E

F G

H

J

Fig. 3. Ceramics of the Uruk period (2/5).

a. RJ4-1, straight-rim jar, B: 2.5Y 6/4, EP-171.b. RJ4-1, straight-rim jar, B: 5Y 5/4, EP-171.c. RJ4-1, straight-rim jar, B: 5Y 6/4, EP-171.d. RJ4-2, thick-rim jar, B: 5Y 6/3, EP-3.e. RJ4-2, thick-rim jar, B: 5Y 7/3, EP-3./. RB4-1, beveled-rim bowl, B: 5Y 6/3, EP-3.

g.h.i.

1.k.

RB 4-1, beveled-rim bowl, B: 2.5Y 7/4, EP-3.04-1, twisted strap handle, B: 5Y 5/2, EP-29.04-2, plain strap handle, B: 5Y 6/4, EP-29.04-3, twisted lug handle, B: 5Y 5/4, EP-171.04-2, plain strap handle. B: 5Y 7/4, EP-171.

305

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

occurrence of small flat and twisted strap handles, andthe high frequency of rim alterations on conical bowls,most distinctively a slightly beveled-rim form (RB5-4).This is probably a Jemdet Nasr complex. The secondcomplex is distinguished by the frequent occurrence ofhatched strips, reserved slip, and incised and excised jarshoulders (05-5, 05-6, 05-7), solid-footed goblets (B5-5),double-rim dishes (05-1), and flat-rim jars (RJ5-5). Thisis an Early Dynastic I complex. The third complex isdistinguished by low frequencies of ledge-rim jars (RJS-1), so common in the previous two complexes; high fre-quencies of the simple flared-rim jars (RJ5-3), a scarcityof decorated sherds except for hatched strip (05-5). Thisis probably an Early Dynastic III complex. It is un-fortunate that the Jemdet Nasr and Early Dynastic IIIcomplexes are indicated primarily by the absence ofthings, for it is difficult to establish their presence unlessthey form the dominant complex on a site.

Sites with dominant assemblages of this period aremarked by quantities of conical bowl fragments, cream-slip sherds, and denticulate sickle blades.

RJ4-1: Simple straight-rim jars (fig. 3a, c): This typeis rare in surface collections from sites of this period,apparently because the potters were using a low-firedbody that disintegrates quickly when exposed. Thosethat occur are little different from their Uruk predecessors,though loop handles are replaced by simple oval lugs at-tached to the rim. None of these lugs have been foundin surface collections, but they are frequent in excavatedsamples (Wright 1969, p. 68, fig. 18e).

RJ5-1: Ledge-rim jars (fig. 4a-c): There is a possi-bility that sharp neck-shoulder and shoulder-body junc-tures are more typical of the Jemdet Nasr complex. Ves-sels of this type frequently have shoulder decorations (seebelow).

RJ5-2: Band-rim jars (fig. 4d-g): These may be morefrequent in Jemdet Nasr times.

RJ5-3: Simple round-rim jars (fig. 4h, i): Distin-guished by a coarse, often dark green body. Probablymore common in Early Dynastic III.

RJ5-4: Flared expanded-rim jars (fig. 4j, k): Distin-guished by light body color; high, gently outcurved neck;and slight, often concave bevel.

RJ5-5: Flat-rim jar (fig. 5a): May be more frequentin Late Early Dynastic I. Distinguished by hard, verysandy paste, and gray to light green color.

RB5-1: Large bowl (fig. 5b): These are common in ex-cavated Early Dynastic I material but occur in JemdetNasr times as well. Often with flattened rim and hatchedstrip below rim. Sometimes lightly chaff-tempered.

RBS-2: Simple conical bowl (fig. 6a-c, g): The sig-nificance of the considerable variation in rim thicknessand angle is not apparent.

RB5-3: Braced conical-bowl rim (fig. 6d): Probablylimited to Jemdet Nasr and Early Dynastic I. Distin-

guished by the thickened band on the exterior below therim.

RB5-4: Beveled conical-bowl rim (fig. 6e): Probablylimited to Jemdet Nasr.

RB5-5: Thickened conical-bowl rim (fig. 6/).RB5-6: Simple stone-bowl rim (fig. 5c): Very com-

mon in Early Dynastic I sites. Made of gypsum or locallimestone, both coarse, porous materials.

B5-1: Pinched ring base (fig. 5d): Distinguished byfinger marks around ring. Often scraped on interior.

B5-2: Flat base (fig. 5e): Distinguished by slight un-eveness on base. Often scraped on interior.

B5-3: Wide-mouth conical-bowl base (fig. 6i): Basestring cut.

B5-4: Narrow-mouth conical-bowl base (fig. 6h):Base string cut.

B5-5: Solid-footed goblet base (fig. 6j): Base stringcut. Frequently asymmetrical. These are typically EarlyDynastic I.

04-1: Twisted strap handle: Jemdet Nasr and EarlyDynastic I examples of this type are small and round insection. Some were attached to ledge-rim jars (RJ5-1)rather than simple straight-rim jars (RJ4-1).

04-2: Plain strap handle (fig. 3k): Jemdet Nasr andEarly Dynastic I examples of this type are small, oftenwith only one groove.

04-3: Twisted lug handle: Perhaps present in JemdetNasr.

04-4: Clay cones (fig. 6j-l): These occur no later thanJemdet Nasr. They vary greatly in size. Some show marksof the potter's knife across the head. Some are chippedacross the head and some are chipped on the side. Thismay reflect functional differences.

05-1: Double-rim dish (fig. 5/): Distinguished by highoutside rim, low inside rim, sometimes slightly outcurvedwith a small gap through it, and by pitting inside theinner rim, perhaps resulting from gentle pounding.

05-2: Spouts (fig. 5g-j): These vary from short tolong, straight to slightly curved, and are generally slightlyconical. Long forms are rare in Early Dynastic III.

05-3: Punctate decorated sherds: Distinguished byan irregular line of slits, either vertical, oblique, or hori-zontal, on jar shoulders.

05-4: Hatch-strip decorated sherd (fig. 6m): Stripvaries from wide to narrow, hatching varies from wideto narrow and vertical to oblique. Varieties of this havebeen called "cable ornament."

05-5: Reserved-slip decorated sherd (fig. 6n): In ourcollections this occurs on Early Dynastic I sites.

05-6: Incised sherd with excised triangles (fig. 60):This combination of shoulder decorations is apparentlyEarly Dynastic I throughout the alluvium. The incisingis generally oblique crosshatched within triangles or bandsradiating out from the neck, or within concentric circlesincised on the jar neck.

306

oi.uchicago.edu

A B

C D

E

F

G

2~..*ht,

H

* -.. *AJ~f.~

J K

Fig. 4. Ceramics of the Jemdet Nasr and Early Dynastic period (2/5).

a. RJS-1, ledge-rim jar, B: 5Y 7/3, EP-29. g. RJ5-2, band-rim jar, B: 5Y 7/4, EP-127.b. RJS-1, ledge-rim jar, B: 2.5Y 7/2, EP-47. h. RJ5-3, round-rim jar, B: 5Y 7/2, EP-30.c. RJ5-1, ledge-rim jar, B: 10YR 7/2, EP-47. i. RJ5-3, round-rim jar, B: 2.5Y 6/2, EP-30.d. RJS-2, band-rim jar, B: 2.5Y 7/4, EP-127. j. RJ5-4, flared expanded-rim jar, B: 5Y 6/2, EP-29.e. RJS-2, band-rim jar, B: 10YR 6/3, EP-4. k. RJ5-4, flared expanded-rim jar. B: 5Y 7/3, EP-29.I. RJS-2, band-rim jar, B: 2.5Y 3/0, EP-5.

307

c:::

oi.uchicago.edu

A

- iYn T

lei

B

~t t

D

C

E

at'.

P:~- r 45

A." ~

.1

a

F

JM

Fig. 5. Ceramics of the Jemdet Nasr and Early Dynastic period (2/5).

a. RJS-5, flat-rim jar, B: 5Y 7/3, EP-47. f. O5-1, double-rim dish, EP-79.b. RB5-1, large bowl, B: 5Y 6/4, EP-142. g. 05-2, spout, B: 10YR 6/3. EP-47.c. RB5-6, stone bowl, calcite, EP-156. h. 05-2, spout, B: 5Y 7/3, EP-47.d. B5-1, pinched ring base, B: 10YR 8/3, EP-30. i. 05-2, spout, B: SYR 6/4, EP-29.e. B5-2, flat base, B: 5Y 6/4, EP-22. j. 05-2, spout, B: 2.5Y 7/4 EP-3.

308

G

I

I• • °

I

oi.uchicago.edu

FB C

G H I

~fr

J

M

*I.

LK

N 0

Fig. 6. Ceramics of the Jemdet Nasr and Early Dynastic period (2/5).

a. RB5-2, conical-bowl rim, B: 2.5Y 2/0, D: 32, EP-60. g. RB5-2 on B5-3, conical bowl, B: 7.5YR 6/6, EP-30.

b. RB5-2, conical-bowl rim, B: 5Y 5/4, D: 16, EP-60. h. B5-4, narrow-mouth goblet base, B: 7.5YR 6/5, EP-30.

c. RB5-2, conical-bowl rim, B: 5Y 5/4, D: 14, EP-60. i. B5-5, solid-footed goblet base, B: 7.5YR 5/6, EP-82.

d. RB5-3, braced conical-bowl rim, B: 10YR 5/4, D: 48, j. 04-4, clay cones, B: 5Y 6/3, EP-5.EP-60. k. 04-4, clay cones, B: 5Y 7/4, EP-3.

e. RB5-4, beveled conical-bowl rim, B: 2.5Y 3/0, D:warped, 1. 04-4, clay cones, B: 2.5Y 7/4, EP-127.EP-60. m. 05-4, hatched strip decoration, B: 5Y 7/4, EP-79.

f. RB5-5, thickened conical-bowl rim, B: 10YR 5/3, D: 38, n. 05-5, reserved-slip decoration, B: 2.5Y 6/2, EP-29.

EP-60. o. 05-6, incised-excised decoration, B: 2.5Y 7/4, EP-79.

309

EA

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

05-7: Denticulate sickle blade: Fine denticulationsmay be more common in the Jemdet Nasr, and coarsedenticulations may be more common in Early Dynastic I.

AKKADIAN PERIOD

Single-component Akkadian sites are very rare in thesurvey area, so a complete complex cannot be defined.Certain distinctive types isolated from mixed collectionsare presented.

Ceramics of this complex are tempered with eithersmall quantities of fine sand or small quantities of vegetalmatter. They vary from buff to light tan.

Sites with dominant surface assemblages of this periodare marked by quantities of ridged sherds (cf. RJ6-2 andRJ6-3).

RJ6-1: Narrow ledge-rim jars (fig. 7a, b): Distin-guished by a gently outcurved neck and slight ledge, oftenconcave on top. This is probably a development fromRJ5-4 and may well occur earlier.

RJ6-2: Ridged thickened-rim jar (fig. 7c): Distin-guished by wide mouth and slightly outcurved rim, tri-angular in section.

RJ6-3: Ridged rounded band-rim jar (fig. 7d): Dis-tinguished by rounded rim folded over to leave a smallcrease under the rim (see illustration).

RB5-2: Conical-bowl rim (fig. 6a-c).RB6-1: Fine carinated vertical-rim bowl (fig. 11b):

Distinguished by relative thinness and by concavity ofthe vertical rim.

B5-4: Narrow mouth conical-bowl base (fig. 6b).06-1: Fine comb-decorated sherds (fig. 7e, /): Distin-

guished by straight combed rows alternating with wavycombed rows, often between ridges. These are commonduring this period but certainly occur both earlier andlater.

UR III AND EARLY LARSA PERIOD

Ceramics of this period are generally lightly chaff tem-pered and buff to yellowish in color.

Two complexes are recognized within the period. Inone, the High band-rim jar (RJ7-1) is frequent. This maybe an Ur III complex. In the other the single groovedband-rim jar (RJ7-3) and the rolled-rim jar (RJ8-1) arefrequent. This may be an Early Larsa complex. Sites withdominant assemblages of this period are generally markedby the occurrence of double grooved band-rim jars (RJ7-2).

RJ7-1: High band-rim jars (fig. Sa, b): Distinguishedby high, gently outcurved band rim.

RJ7-2: Double groove band-rim jar (fig. Sc, d).RJ7-3: Single groove band-rim jar (fig. 8e, f).RJ7-4: Grooved square-rim jar (fig. 8g): A large,

wide-mouth vessel with deep wide grooved and flattened

rim, square in section. Perhaps a development from RJ6-2.RJ7-5: Flattened rolled-over rim: In these complexes

this type includes vessels transitional between RJ7-4 andRJ8-1.

RJ8-1: Rolled-rim jar (fig. 10a, b).RB6-1: Fine carinated vertical-rim bowl (fig. 11b).RB8-1: Band-rim bowl (fig. 10g): Rare.RB8-2: Multiple-groove simple-rim bowl (fig. 10o).RB9-2: Shallow conical bowl (fig. 12h).B7-1: Disk base (fig. 9a): Frequently has a concavity

in interior.B7-2: Large button base (fig. 9b): Distinguished by

relatively large, round base separated from rounded vesselbottom by a sharp crease, and by rounded concavity inbottom of base.

B7-3: Small conoidal base (fig. 9d): Distinguished byan incurved base coming to point, slight rounded nipple,or slight flat area.

B7-4: Small button base (fig. 9e).B7-5: Ring base (fig. 9c): Distinguished by relatively

narrow jar body above the base. "Makers' marks" aresometimes visible on the base.

B7-6: Medium-high base (fig. 9f): Distinguished byrelatively large base, narrow stem, and rounded bottomof jar body above.

B8-2: Small cylindrical base (fig. 11g, h).06-1: Fine comb-decorated sherds (fig. 7e, f).07-1: Pot stand (fig. 9g): This type is distinguished by

band rims and top and bottom rims of equal diameter.08-1: Hatched-groove-decorated sherds (fig. 11k):

Rare.09-1: Pot stand (fig. 13g): Rare.

LATE LARSA AND OLD BABYLONIANPERIOD

Ceramics of this period are generally straw temperedand buff to light green in color. The shape category ofsmall bowls shows a tremendous amount of variationthat is not well understood, even though we have morecollections of this period than any others.

There are two complexes within the period. One hasthe hatch-rim jar (RJ8-2), the shallow groove-rim bowl(RB8-5), and the small cylindrical base (B8-2). The otherlacks these items. Whether these two complexes representstyle change through time within the period or someother factor is not known.

Sites with dominant surface assemblages of this periodare indicated by the occurrence of hatched-groove decor-ated sherds (08-1).

RJ8-1: Rolled-rim jar (fig. 10a, b): Distinguished bya gradual thickening of the slightly curved neck and arolling over of the lip resulting in a slight crease belowthe rim. Often hatched-groove decoration (fig. 11k) oc-curs on the vessels with this type of rim.

310

oi.uchicago.edu

A

C

D

Fig. 7. Ceramics of the Akkadian period (2/5).

a. RJ6-1, narrow ledge-rim jar, B: 10YR 7/4, EP-17.b. RJ6-1, narrow ledge-rim jar, B: 10YR 6/4, EP-17.c. RJ6-2, ridged thickened-rim jar, B: 10YR 6/4, EP-12.

d. RJ6-3, ridged rounded-band rim jar, B: 10YR 8/5, EP-30.e. 06-1, comb decoration, B: 10YR 6/3, EP-12.f. 06-1, comb decoration, B: 5Y 6/4, EP-12.

311

8

(r I

I st --- --Wt wla oI

· · , cr- r -- , -r· · · r u·. '.C? . · ·. i :i· -rCJ r

- · · · · 5;: · r r~,~ r · :···- · · · · · · , · · +, .· · ··. . ·r · · ·r r.. r'

oi.uchicago.edu

* . ... ..,

-~

C

A B

E

* ** . -

F

G

Fig. 8. Ceramics of the Ur III Early Larsa period (2/5).

a. RJ7-1, high band-rim jar, B: 5Y 6/3, EP-60.b. RJ7-1, high band-rim jar, B: 5Y 7/4, EP-30.c. RJ7-2, double groove band-rim jar, B: 10YR 6/4, EP-60.d. RJ7-2, double groove band-rim jar, B: 5Y 5/4, EP-84.

e. RJ7-3, single groove band-rim jar, B: 5Y 6/4, EP-60.f. RJ7-3, single groove band-rim jar, B: 5Y 6/4, EP-60.g. RJ7-4, grooved square-rim jar, B: 5Y 7/3, EP-84.

312

D

,--- - -

··.. · I- ·: :

oi.uchicago.edu

A 01C

E

F

G

Fig. 9. Ceramics of the Ur III and Early Larsa period (2/5).

a. B7-1, disk base, B: 5Y 6/4, EP-84.b. B7-2, large button base, B: 10YR 6/4, EP-29.c. B7-5, ring base, B: 7.5YR 6/6, EP-111.d. B7-3, small conoidal base, B: 5Y 6/3, EP-84.

e. B7-4, small button base, B: 10YR 6/5, EP-60./. B7-6, medium-high base, B: 5Y 6/2, EP-111.g. 07-1, pot stand, B: 5Y 7/3, EP-16.

313

D

............ j• m A w ..... UI- -r A M-M.&MI6-.Cylý1.tFL

CA B

ZWO--

16 :8

oi.uchicago.edu

A

-. ~ ____

4QO~

C

D E F

G

H

Fig. 10. Ceramics of the Late Larsa and Old Babylonian period (2/5).RJ8-1, rolled-rim jar, B: 2.5Y 6/6, EP-34.RJ8-1, rolled-rim jar, B: 2.5Y 6/5, EP-34.RJ8-2, hatched-groove rim jar, B: 5Y 6/4, EP-2.RJ8-3, thin simple jar rim, B: 10YR 8/4, EP-16.RJ8-3, thin simple jar rim, B: 2.5Y 8/4, EP-72.RJ8-3, thin simple jar rim, B: 10YR 7/4, EP-72.

g. RB8-1, band-rim bowl, B: 5Y 6/4, EP-4.h. RB8-2, multiple-groove simple-rim bowl, B: 2.5Y 7/4,

EP-2.i. RB8-3, irregular-groove simple-rim bowl, B: 5Y 6/4, EP-

133E.

314

a.b.C.

d.e.f.

__

4

B

oi.uchicago.edu

Law& A. -~3-ZX

B C

D E F

H I

Fig. 11. Ceramics of the Late Larsa and Old Babylonian period (

a. RB8-4, deep multiple-groove-rim bowl, B: 2.5Y 5/4, EP-1.b. RB6-1, fine carinated vertical-rim bowl, B: 5Y 5/3, EP-34.c. RB8-5, shallow single-groove-rim bowl, B: 10YR 7/5, EP-

150.d. B8-1, small high base, B: 10YR 5/4, EP-4.e. B8-1, small high base, B: 2.5Y 3/0, EP-4.

J K

/. B8-1, small high base, B: 10YR 6/5, EP-72.g. B8-2, small cylindrical base, B: 5Y 6/4, EP-34.h. B8-2, small cylindrical base, B: 10YR 6/4, EP-30.i. B8-3, incurved flat base, B: 5Y 6/6, EP-34.j. B8-3, incurved flat base, B: 2.5Y 6/4, EP-84.k. 08-1, hatched-groove decoration, B: 10YR 7/4, EP-34.

315

,~t

G

- Ir •

oi.uchicago.edu

A B

D

F

G

"~7" * *....-'~

H

Fig. 12. Ceramics of the Cassite and Post-Cassite periods (2/5).

a. RJ9-1, heavy band-rim jar, B: 5Y 6/4, EP-36.b. RJ9-1, heavy band-rim jar, B: 5Y 5/3, EP-89.c. RJ9-2, incurved rolled-rim jar, B: 5Y 5/3, EP-189.

d. RJ9-2, incurved rolled-rim jar, B: 5Y 5/4, EP-189.

e. RJ9-3, outcurved rolled-over-rim jar, B: 5Y 6/4, EP-189.

f. RJ9-3, outcurved rolled-over-rim jar, B: 5Y 7/4, EP-189.

g. RB9-1, heavy vertical-rim carinated bowl, B: 5Y 5/3, EP-

h.i.

16.RB9-2, shallow conical bowl, B: 5Y 6/4, EP-36.

RB9-3, shallow carinated bowl, B: 2.5Y 8/4, EP-36.

316

- . . ... C : . .../ .: .- ;- ...' . ".

E

-- --- -- -- r~L

.i

""

::4

oi.uchicago.edu

A

B

C

E

Fig. 13. Ceramics of the Cassite and Post-Cassite periods (2/5).

a. RB9-4, heavy ridged-rim bowl, B: 5Y 8/3, EP-189.b. RB9-5, heavy round beveled-rim bowl, B: 5Y 5/4, EP-186.c. B9-1, solid-footed chalice base, B: 5Y 5/4, EP-89.d. B9-1, solid-footed chalice base, B: 5Y 7/2, EP-6.

317

D

F

G

e. B9-2, rough cylindrical base, B: 5Y 5/4, EP-89.f. B9-3, small disk base, B: 5Y 7/3, EP-189.g. 09-1, pot stand, B: 5Y 5/3, EP-89.

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

RJ8-2: Hatched-groove-rim jar (fig. 10c): Distin-guished by a band rim with either one or two groovescreating two or three ridges respectively. The lower ormiddle ridge is hatched.

RJ 8-3: thin simple jar rims (fig. 10d-f).RB6-1: Fine carinated vertical-rim bowl (fig. 11b).RB8-1: Band-rim bowl (fig. 10g): Distinguished by a

rounded band rim on a large, shallow bowl. Often thereis an incised wavy line as shown in the illustration.

RB8-2: Multiple-groove simple-rim bowl (fig. 10h):Distinguished by a yellowish body, small hemisphericalshape, and up to eight closely spaced grooves on the ex-terior of the rim.

RB8-3: Irregular-groove simple-rim bowl (fig. 10i):Distinguished by buff to green body, thick walls, andseries of thickenings resembling the RB8-2 groovings.Often there is an incised wavy line as shown in the illus-tration.

RB8-4: Deep multiple-groove-rim bowl (fig. 11a):Distinguished by near-straight sides. Some grooving wasdone with a stylus held at various angles while the potwas still turning on the wheel, but some seems to havebeen done with a notched template held against the rotat-ing vessel. The significance of this difference is unknown.

RB8-5: Shallow single-groove-rim bowl (fig. 11c):Distinguished by simple rim with single exterior groove.

B7-1: Disk base (fig. 9a).B7-3: Small conoid base (fig. 9d).B8-1: Small high base (fig. 11d-f): Distinguished by

small cylindrical base with a high stem and roundedbody above. In section, this body is often shaped likethe letter U. The body is often yellowish.

B8-2: Small cylindrical base (fig. 11g, h): Distin-guished by thick base and marked but regular interiorthickenings.

B8-3: Incurved flat base (fig. 11i, j): Distinguishedby a cylindrical or slightly constricted body with arounded bottom that is formed into a small, flat sold base.A similar Neo-Babylonian and Achaemenian type is gen-erally smaller and yellower.

08-1: Hatched-groove decorated sherds (fig. 11k):Distinguished as two to four parallel grooves around a jarneck with hatches across several of the resulting grooves.Only more complex varieties are definitely of this period.

08-2: Sherds with heavy black lines painted in bitu-minous paint. This also recurs in the Neo-Babylonianand later periods.

CASSITE AND POST-CASSITE PERIOD

Ceramics of this period are thick and have heavy vegetaltemper. They vary from light brown to dark green.

There are two complexes within this period. One hasheavy band-rim jars (RJ9-1) and shallow carinated bowls(RB9-3). This is probably a Cassite complex. The other

has incurved rolled-rim jars (RJ9-2), outcurved rolled-rim jars (RJ9-3), heavy ridged-rim bowls (RB9-4), heavyrounded bevel-rim bowls (RB9-5), and a small disk base(B9-3). This is probably a Post-Cassite complex. Oftenthese two complexes occur on the same site.

Sites with dominant surface assemblages of this periodare marked by the occurrence of solid-footed chalicebases (B9-1).

RJ7-5: Flattened rolled-over-rim jars: In this period,this category includes small fragments of RJ9-1 on whichthe distinctive neck characteristics are no longer pre-served.

RJ8-1: Rolled-rim jar (fig. lOa, b): In this period, thiscategory includes small fragments of RJ9-2 on which theneck characteristics are no longer visible.

RJ8-3: Thin simple jar rims (fig. 10d-f).RJ9-1: Heavy band-rim jars (fig. 12a, b). Distin-

guished by a neck that slopes below the rim to a markedcrease, then curves out and down to form the body ofthe vessel.

RJ9-2: Incurred rolled-rim jars (fig. 12c, d): Distin-guished by the lack of a neck, the thickened, slightly rolledrim is formed directly from the incurved upper body ofthe vessel.

RJ9-3: Outcurved rolled-over jar rim (fig. 12e, f):Distinguished by outcurved neck and lip thickened androlled over to the point where there is a marked con-cavity below it.

RB8-1: Band-rim bowl (fig. 10g).RB8-3: Irregular-groove simple-rim bowl (fig. 10i).RB8-5: Shallow single-groove rim bowl (fig. 11c).RB9-1: Heavy vertical-rim carinated bowl (fig. 12g).

Distinguished by thickness of body and by slightly rolled-over rim.

RB9-2: Shallow Conical Bowl (fig. 12h). Both largeand small forms occur.

RB9-3: Shallow carinated bowl (fig. 12i): Distin-guished by marked carination halfway between the rimand the base, generally small.

RB9-4: Heavy ridged-rim bowl (fig. 13a): Distin-guished by a thick rim and very thick ridge below therim. Probably a development of RB9-1.

RB9-5: Heavy round bevel-rim bowl (fig. 13b): Dis-tinguished by a slight lip at the lower edge of the rim.

B7-1: Disk base (fig. 9a).B7-2: Large button base (fig. 9b).B7-3: Small conoid base (fig. 9d).B7-4: Small button base (fig. 9e).B9-1: Solid-footed chalice base (fig. 13c, d): Note the

heavy ridges on the interior.B9-2: Rough cylindrical base (fig. 13e): Distinguished

by large size and irregular finger markings on lower body.B9-3: Small disk base (fig. 13/): Distinguished by its

small size. Some examples are string cut.09-1: Pot stand (fig. 13g): Distinguished by unequal

318

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

diameter of the rims and by the rounded contour of therims in section.

NEO-BABYLONIAN, ACHAEMENIAN, ANDLATER PERIOD

Ceramics of this period have light to medium vegetaltempering in the larger-sized vessels and a sandy body inthe smaller vessels. Colors range from yellowish togreenish. Glaze is frequent on certain types, though it isoften weathered to a white encrustation.

Sites with dominant assemblages of this period aremarked by Round-rim jar rims (RJ10-1).

Sites on which a majority of round-rim jars (RJ10-1)lack ridges seem to have few model horses (010-3),stamped sherds (010-5), and eggshell-ware sherds. Thismay represent a Neo-Babylonian complex, since siteswith this assemblage seem to be on older positions onthe canal systems. Unfortunately there are few sites witha clear dominance of this complex.

RJ10-1: round-rim jar rim (fig. 14a, b): Distinguishedby vertical or slightly concave thin-walled neck andthickened rim. Often the rounded section of the rimthickening, as shown in the illustration, is slightly flattenedon the top and exterior, creating a squared section. Varie-ties of this type have plain necks, necks with a singlesharp ridge, necks with multiple sharp rdges, and neckswith hatched ridges. Glaze occurs on some examples.

RJ10-2: Thick-rim jar (fig. 14c): Distinguished by aslight rim thickening and low concave neck, changingsmoothly into a convex body, frequently grooved. Perhapsthis developed from RJ9-3.

RJ10-3: Club-rim jar (fig. 14d): Distinguished by avertical or slightly concave thin-walled neck with a smallrounded rim separated from the neck by a relatively deepdownward groove.

RJ10-4: Large thick-rim jar (fig. 14e): Distinguishedby a large, relatively wide mouth and the immediatesloping out from the rim to a rounded body. Grooving orridging is frequent.

RJ10-5: Concave band-rim jar (fig. 14f): Distin-guished by a large, relatively wide mouth, thickened,flattened rim, and deep upward groove on the body iso-lating a wide concave band.

RB10-1: Simple ledge-rim bowl (fig. 14g): Distin-guished by thin walls. On some sites, glaze is preservedon many of these rims.

RB10-2: Narrow ledge-rim carinated bowl (fig. 14h):Same.

RB10-3: Wide ledge-rim carinated bowl (fig. 15a):Same.

RB10-4: Inverted-rim carinated bowl (fig. 15b): Same.RB10-5: Grooved trapezoidal-rim bowl (fig. 15c):

The illustrated example is relatively small and relativelythick walled for this type.

B10-1: Grooved flat base (fig. 15d): Distinguished byscoring on the bottom and grooving on lower body.Frequently glazed.

B10-2: Thin flat base (fig. 15f): Distinguished bythinness, sandy texture of the paste, and slightly convexirregular remnant of clay in the center of the interior.

B10-3: Hole base (fig. 15e): Distinguished by largesize of vessel, the hole, and the wide thickened rim aroundthe hole.

B10-4: Small incurved flat base (fig. 16a, b): Distin-guished by small size and yellow, sometimes sandy pasteclay body.

010-1: Simple loop handle (fig. 16c).010-2: Pinched lug handle (fig. 16d).010-3: Model horse (fig. 16e, f): Pinched mane and

long modeled tail frequent. Rider is rare.010-4: Stamped-node decorated sherds (fig. 16g):

Distinguished by fingerprint in the node on the insideand stamp outside, either plain as illustrated, or with avariety of designs.

010-5: Medallion-stamp decorated sherds (fig. 16i):The wheellike design is most common.

010-6: Tripod support legs (fig. 16h).010-7: Dentate-stamp decorated sherds (fig. 16j, k):

Frequently in vertical or oblique sets between the grooveson jar shoulders.

010-8: Roughened bottom dishes (fig. 161); Mostcommon are bottom sherds from this form, which is ashallow oval vessel with thick, heavily chaff temperedwalls and bottom. The bottom was roughened with atoothed instrument whose teeth are large and may beround, square, or triangular. This form is distinguishedfrom the Hassuna "husking tray" by the small size ofthe roughening impressions.

010-9: Eggshell-ware sherds: Small, very thin sherdswith a fine paste, generally very light brown or whiteand sometimes glazed.

319

oi.uchicago.edu

BA

II

C

E

* . ... .,

F

G

Fig. 14. Ceramics of the Neo-Babylonian and Later period (2/5).RJ10-1, round-rim jar, B: SY 7/3, EP-50.RJ10-1, round-rim jar, B: 5Y 6/4, EP-47.RJ10-2, thick-rim jar, B: 5Y 6/4, EP-50.RJ10-3, club-rim jar, B: 10YR 7/3, EP-50.RJ10-4, large thick-rim jar, B: 10YR 8/4, EP-148.

f. RJ10-5, concave band-rim jar, B: 10YR 7/3, EP-88.g. RB10-1, simple ledge-rim bowl, B: 10YR 7/3, EP-29.h. RB10-2, narrow ledge-rim carinated bowl, B: 5Y 6/4, EP-

50.

320

D

a.b.c.

d.e.

H

II -j -

- --

- --

1

oi.uchicago.edu

* . - q *~-. - - -

A

It.

!-k.Z. W-.S* C

- ~ **40

C

- * - ~ :-.~:~-~-. -

-. £:.

E

D

Fig. 15. Ceramics of the Neo-Babylonian and Later period (2/5).

a. RB10-3, wide ledge-rim carinated bowl, B: 10YR 7/5, EP- d. B10-1, grooved flat base, B: 5Y 8/3, EP-50.50. e. B10-3, hole base, B: 5Y 6/4, EP-81.

b. RB10-4, inverted-rim carinated bowl, B: 5Y 7/3, EP-50. f. B10-2, thin flat base, B: 2.5Y 8/4, EP-72.c. RB10-5, grooved trapezoidal-rim bowl, B: 5Y 6/4, EP-102.

321

B

S.,.,

F

dOW

C

oi.uchicago.edu

CA B

D

E F

H

b_~C

0,0

o oc. 0a

J K

Fig. 16. Ceramics of the Neo-Babylonian and Later period (2

a. B10-4, small incurved flat base, B: 10Y 7/4, EP-88.b. B10-4, small incurved flat base, B: 2.5Y 7/4, EP-102.

c. 010-1, simple loop handle, B: 5Y 5/3, EP-50.

d. 010-2, pinched lug handle, B: 5Y 6/4, EP-162.

e. 010-3, model horse, B: 5Y 6/4, EP-50.

f. 010-3, model horse, B: 5Y 8/4, EP-162.

L

15).g. 010-4, stamped-node decoration, B: 10YR 7/5, EP-162.

h. 010-6, tripod support leg, B: 5Y 7/3, EP-50.

i. 010-5, medallion-stamp decoration, B: 5Y 7/4, EP-110.

j. 010-7, dentate decoration, B: 5Y 6/3, EP-48.

k. 010-7, dentate decoration, B: 2.5Y 4/2, EP-113.

1. 010-8, roughed dish bottom, B: 5Y 7/3, EP-54.

322

f

%.

~,

IL.'r 4 -il

oi.uchicago.edu

The Development of Settlementin the Area of Ur and Eridu

Middle Paleolithic bands hunted on the deserts south ofthe Euphrates alluvium (Wright 1967) and along thewatercourses of a then very different Euphrates River(Voute 1957). Doubtless many campsites of both Middleand succeeding Upper Paleolithic groups are buriedbeneath the recent sediments. However, the relativelycontinuous record of human endeavor on the alluviumdoes not begin until well after the end of the last glacia-tion, during the sixth millennium B.C.

EARLY VILLAGES AND TOWNS OFSOUTHERN SUMER (THE UBAID PERIODSAND BEFORE)

The earliest known communities in the surveyed areaare those of the Eridu period or Ubaid I phase, whoseceramics are related to those of the Samarran communi-ties of the Zagros Piedmont and the Middle Euphratesand Tigris valleys (Oates 1976, pp. 20-22). Whether theserepresent immigrants or, in contrast, acculturated localcommunities, they would have been familiar with thepatterns of cattle-breeding and small-scale irrigation agri-culture pursued in the Samarran villages. On the otherhand, the Eridu period peoples also had on hand copiousmarsh resources. A canoe model and numbers of perfor-ated clay ovoids, perhaps net weights (Lloyd and Safar1948, p. 118, Pl. III), from Eridu period levels suggestthat the marshes were already being utilized in a sophis-ticated manner. Regrettably, there are no samples avail-able of faunal or floral remains from sites of this phase,and the balance between collecting and farming is un-known.

By the end of the Eridu period there are four knownsettlements within the survey area. To the south therewas Eridu (EP-3) and 'Usaila (EP-104), 7 kilometers tothe west; to the north there was Ur (EP-10) and Tell al-'Ubaid (EP-8), 6 kilometers to the west. In no case canthe size of an Eridu period settlement be estimated, butat Eridu and Ur diagnostic sherds were found in only oneof the several excavations that reached apparently naturaldeposits, suggesting small communities of only a hectareor so.

Though there is no evidence of the proliferation ofcommunities-the same four sites remain the only onesoccupied-by the end of the Hajji Muhammad period orUbaid II phase, there is settlement growth. At Eridu thereis a temple platform in the "Temple Sounding," anddebris is evident in the "Hut Sounding" 100 meters tothe east (Safar 1950, p. 28). The distance between the

soundings suggests a site area of more than 2 hectares.At this time, nearby EP-104 had not yet reached its maxi-mum size and was probably about hectare. Thus, thereis evidence from the Hajji Muhammad period in theEridu area of some differentiation both within and be-tween settlements.

These sites remain occupied into the succeeding EarlyUbaid period or Ubaid III phase, and for the first timesome evidence is available on the economic and sociallife of the period. A faunal sample from Eridu has wildonager, which would have been hunted on the nearbyalluvial desert, domesticated cow, and goats, sheep, orgazelle (Flannery and Wright 1967, pp. 61-63). ExposedEarly Ubaid strata at nearby EP-104 have concentrationsof freshwater mollusks, showing that the river or marshenvironment was also used. Plant use is unfortunatelynot documented.

It is during the Early Ubaid period that the settlementof Eridu grows beyond the limits of the central mound,at least to the southwest, though no actual architecture isknown from this area (Campbell-Thompson 1919, p.136). The temple platform becomes more elaborate, andthere is both mud brick and adjacent reed domestic con-struction in the "Hut Sounding" (Safar 1950, p. 28). Thefragmentary evidence does not allow a social characteri-zation of the center. During this period, several new settle-ments are founded. Perhaps early in the phase, whenbowls derived from Hajji Muhammad types are com-mon, the site of Merejib (EP-29) (cf. Woolley 1956, pp.82-85) is occupied near the channel through the Hazim,8.5 kilometers southeast of Ur. Also at this time, therewas use of a campsite on the great sand dunes that borderthe Eridu plain on the southwest (EP-38W), approxi-mately 45 kilometers west of Ur, outside our area ofintensive survey. The site is only a scatter of sherds andretouch flakes from large chert bifaces. Perhaps duringthis period people went to the dunes to hunt, to cutHaloxylon wood for fuel, or to cultivate grain crops inareas with a high water table. Only thorough survey of thedifficulty dune area will tell. Perhaps at a slightly latertime during this phase, as indicated by numbers of finecup sherds, another settlement (EP-141) was founded 8kilometers northwest of al-'Ubaid. Thus there is expan-sion of the settled areas both up and down the develop-ing levee system.

During the Late and Terminal Ubaid periods, the areacan be characterized in both economic and social terms.Several additional food sources are documented. Boatmodels from Eridu show that sailing craft have been

323

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

developed; and marine fish are being brought to thetemple at Eridu as offerings. In the area of agriculture,one can assume that the herding of cattle continues, andfrom Eridu there are identified wheat, six-row barley,and dates (Safar, pers. comm.). Thus the elements offarming known from later texts are all present by the endof the fifth millennium B.C. That irrigation canals havereached substantial proportions is indicated by both sitelocation and the distribution of evidences of cultivation.In the first place, in the northern part of the survey region,Ur, al-'Ubaid, and newly occupied site EP-73 are locatedon the corners of a triangle oriented in such a way thatit is unlikely that one channel watered all three of themgiven the later documented amplitude of channelmeanders in the area (fig. 17). One of the smaller siteswas probably watered by a small branch canal. In thesecond place, clay sickles, whose period of maximum usewas in Late and Terminal Ubaid times, are found on latersites above Eridu and Ur over belts about 5 kilometerswide. To cultivate fields in such belts would requirebranch canals 3 to 5 kilometers long. Needless to say,

one would prefer to have air photograph traces and crosssections of such canals, but the indirect evidence of sickledistribution is better than none at all. Such canals arewithin the abilities of extended kin groups to build andmaintain. Although they would create no particular de-mands for managerial control, they would provide suchkin groups with more stable and perhaps more nucleatedagricultural resources (Fernea 1970, pp. 77-142).

Whatever the situation in the Early Ubaid period, thereis no question that by Late Ubaid times we are dealingwith a variety of large and small settlements. The aggre-gate of the evidence from the various excavations at Eriduindicates that it had grown to about 12 hectares. Theearlier mound now forms the nucleus of a broad plat-form on which was a temple on a raised terrace andsome substantial buildings of a residential character(Lloyd and Safar 1948, pp. 119-21, pl. VI; Safar 1950,p. 29, fig. 1). In the surrounding town below were moremodest buildings, also probably residential (Campbell-Thompson 1919, p. 136, trench XIV, 3). The Eridu ceme-tery contained both individuals with substantial brick

////

\3

\\

\

e73

Sur

96

\>

2/

/I\(.

.2929 N°134

104

N

*eridu N

K.

Fig. 17. Settlements of the Late Ubaid period and the distribution of clay sicklesfound on later sites.

324

141 vA

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

tombs and individuals with none (Lloyd and Safar 1948,pl. IV), but one cannot say what type of social differentia-tion this represents until the full burial inventory is pub-lished. Ur was apparently similar in scale to Eridu, cover-ing about 10 hectares, and it also had modest buildingsand a cemetery on its margins (Woolley 1956, pp. 69, 87-102).

Subsidiary to these towns were a variety of smallersettlements. Upstream from Eridu were at least twosmaller settlements giving a minimum settlement area forthe enclave on the southern most branch of the Euphratesof 17 hectares. Doubtless some other small settlementsremain undiscovered. The cultivated zone around thesesettlements, indicated by stray sickles, was about 100square kilometers. Upstream from Ur, there is evidenceof four small settlements, two of which present interest-ing surface features. Al-'Ubaid itself is on a low sandknoll covering about 5 hectares. The site has a thin de-posit composed mostly of ash lenses. The excavator'sopinion that it was a settlement of mud huts is quite rea-sonable (Hall and Woolley 1927, pp. 149-51). Probablyonly a portion of the knoll was occupied at any one time.In contrast, site EP-141 is a small, relatively high moundof mud construction debris, perhaps a single building ona platform. Certainly the small subsidiary settlementswere not uniform village communities. Assigning anaverage value of 2.1 hectares to those communities thatcannot be measured because of later obscuring deposits,the enclave on the northern branch of the Euphratescovered a minimum area of about 18 hectares. The cul-tivated zone also would have comprised about 100 squarekilometers. Smaller cultivated areas existed to the eastaround Merejib and to the west around site EP-36 (fig.17), and the campsite on the dunes continued in use.

It is unlikely that southern Sumer as a whole con-tained more than 2,500 to 4,000 persons, or about 20persons per square kilometer of enclave, far fewer thanwere later supported with a similar technology. Notethat this is a revision of the figure of 5,000 proposedearlier (Wright 1969, p. 25), resulting from a reconsidera-tion of the dating evidence and assumption of between125 and 200 persons per hectare of settlement.

From this modest first apogee of settlement there was,in the succeeding Terminal Ubaid period, a diminutionin the number of settlements and the extent of cultivatedzones. Outlying Merejib and nearby sites are abandoned,and site EP-141 in the Ur enclave was no longer occupied.Others may also have been abandoned. There is only onenew site, EP-98, 10 kilometers northwest of Eridu. Thisnew site forms a triangle with Eridu and site EP-104 suchthat one of the smaller settlements probably was wateredby a branch canal. Though it covers only 1 hectare, thissettlement has evidence of the manufacture of bothpottery vessels and clay sickles. At Ur, a large area ofceramic manufacturing debris has been exposed on the

site margin, but the central areas have not been excavated.At Eridu, the substantial residences continue on the cen-tral platform, and modest structures continue at its foot.

The evidence of the fifth and sixth millennia suggestsa long period of gradual growth, but this may well bean illusion created by the small sample of sites combinedwith our limited understanding of the ceramic sequence.Some early sites may be hidden beneath massive latersites, while others doubtless have been removed by themeandering of river channels. Of those visible, we havecomprehensive ceramic samples only from Eridu. Whenmore of the sites have been excavated, I suspect that themajor period of town growth will prove to be the be-ginning of the Early Ubaid period or Ubaid III phase.Before that time there was a range of relatively smallsettlements. After that time a two-level settlement hier-archy was present in each enclave, the towns of Eriduand Ur being the loci of authorities whose powers are yetunknown.

THE FIRST STATES (URUK TO EARLYDYNASTIC)

If the Terminal Ubaid period was one of slight diminu-tion in settlement, the Early Uruk period was one offundamental change. Early Uruk ceramics are knownonly from the two towns of Eridu and Ur. Regrettably,the evidence from these two is not comparable. Ur isknown primarily from exposures on its south margin,where there is evidence of an area for ceramic manufacturecontinuing from the previous period (Woolley 1956, p.66, pl. 73). The various excavations of Eridu have ex-posed a number of substantial buildings, some decoratedwith cone mosaic, all on the central platform, which hadgrown to cover 4 hectares (Taylor 1855, pp. 404-15;Hall 1930, pp. 208-13; Lloyd and Safar 1948, pp. 108-10, fig. 7). Around this must have been an extensive lowertown covering perhaps 40 hectares, as indicated by thesherd scatter. An early test excavation visible approxi-mately 700 meters northwest of the ziggurat revealedEarly Uruk ceramics and copper fragments. Little else isknown about this lower settlement. However, it is clearthat while small settlements were abandoned, Eridu hadexpanded greatly. Such nucleation may imply unsettledconditions. If our estimates of the town sizes are correct,population would have been at most 6,200 to 10,000 inthe area.

In the succeeding portion of the Uruk period, Ur con-tinued as a small town, but Eridu was abandoned (fig.18). The buildings on its acropolis filled with more than 2meters of windblown sand. Periodically deposited as thesand accumulated were groups of crude bowls and smalljars that the excavators variously interpret as graves oras votive caches (Campbell-Thompson 1919, pp. 110-12).Whether the assorted Late Uruk sherds from Eridu also

325

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

/ \/060

/1

\

30'

^. N

.7 N°,o ^.*1

I8 Ur /

IIo.

29

\ \

\\K.

044

km00 5

Fig. 18. Settlements of the Late Uruk and Jemdet Nasr periods.

represent such deposits or indicate actual reoccupationcannot be ascertained because of yet later leveling andreconstruction. The encroachment of sand on the centralprecinct of Eridu may represent a cutting off of thesouthermost channel of the Euphrates and desiccationof the Eridu plain. However, one suspects that even withvegetation a great deal of sand would blow into the areafrom the large dunes to the southwest, and that lack ofmaintenance alone would lead to an accumulation.

Whatever the situation around Eridu proper, to thenorth and east the Late Uruk period is one of proliferationof smaller settlements. Ur itself remains a small town ofabout 10 hectares. The pottery-producing area continues,and no expansion of the town to the south occurs. South-east along what may have been the southerly course ofthe Euphrates are a small settlement at Merejib about 12kilometers away (EP-29), where Woolley found gypsumcement block buildings (Woolley 1956, p. 83), and acenter of 3.5 hectares (EP-171) near Tell al-Lahm, about32 kilometers away. Northwestward, several small settle-ments are occupied: one at al-'Ubaid itself, only 6 kilo-meters from Ur, one at site EP-141, 15 kilometers distant,

and perhaps another at site EP-60, 18 kilometers distant.This low density of settlements strung out along the Eu-phrates contrasts to the dense network of settlementsaround Late Uruk Warka to the northwest. More inter-esting than this linear pattern is a small settlement about12 kilometers southwest of Ur and 12 kilometers northof Eridu, just inside the limits of the Eridu depression atthe point where the Hazim dips under the alluvium (EU-7). It consists of six small mounds of debris, most withconcentrations of ceramic slag, three aligned north-north-west to south-southeast and three aligned at a near rightangle to this, extending east-northeast (Fig. 26). Appar-ently this site was a series of structures and kilns at thejuncture of a canal and one of its branches. The slopes ofthe land seems such that this canal would flow from thenorth-northwest and would have left the Euphrates nearor above site EP-141, about 10 kilometers away. It contin-ued an unknown distance in the direction of Eridu, butthere is no substantial settlement farther on unless it is atEridu itself. This canal, founded during the Late Urukperiod and developed during the Jemdet Nasr period, maywell have been excavated to provide access and sustenance

326

N\

N

''*171

"-? ? /

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

to the shrine at Eridu. It is notable that this is the onlycanal that follows the course from Uruk to Eridu dis-cussed by Jacobsen (1960, pp. 180-83) on the basis oflater ritual texts. Cultivated area on the Eridu plain cannotbe estimated, but along the Ur channel about 200 squarekilometers would have been available to a population ofabout 2,500 to 4,000, giving a maximum of 20 persons persquare kilometer during Late Uruk times.

Recognizing sites of the Jemdet Nasr period is difficultnot simply because of the covering of silt and later occu-pational layers that plagues all efforts at the geographicalstudy of ancient Mesopotamia, but also because of thepoor understanding of its nondescript ceramics (Adamsand Nissen 1972, pp. 99-103). However, while somesmaller Jemdet Nasr sites may be unrecognized, the largersites can be discussed (fig. 18). Ur itself has grown, theformer ceramic producing area being overridden by do-mestic buildings. Presuming similar expansion in otherdirections, 15 hectares may have been occupied. A ceme-tery was on its south edge. The central precinct, later to bethe site of the ziggurat and the temple of Nanna, hadbuildings with cone mosaic decoration (Woolley 1939,pp. 1-6, pl. 11). Up the Euphrates channel 18 kilometersto the northwest lies Ishan Khaiber (EP-60), a small centerof 3.5 hectares with evidence of a cone-decorated buildingand pottery kilns. Intermediate Sakheri Kabir (EP-30),only 8 kilometers from Ur, may have been founded in thisperiod, but the evidence is uncertain. The previously dis-cussed Late Uruk canal waters the town of Rejibah (EP-4,5, and 93), 13 kilometers southwest of Ur and coveringabout 23 hectares. Buildings with cone decoration andpottery kilns are indicated here. If the journey of Inannadiscussed by Jacobsen refers to this canal, Rejibah mightbe one of the places mentioned (Jacobsen 1960, pp. 180-83). This canal would have continued southward to thenorthern suburbs of Eridu (EP-46), for here, beneath theEarly Dynastic palaces cleaned by Safar, are possible Jem-det Nasr sherds and clay cone occurrences. This moundcovers about 7.5 hectares. In contrast to the larger centerswith public buildings is a small unnamed site (EP-156)11 kilometers northwest of Rejibah and 12 kilometerssouthwest of Ishan Khaiber. Although wall cones are ab-sent, there are many stone vessel fragments and somemetal artifacts suggesting that this 0.6 hectare settlementcontained individuals of higher status or wealth. Thealignment of the site suggests that water came to it by acanal from the main Euphrates channel to the north.Perhaps this was an isolated elite residence. Another pos-sible Jemdet Nasr site is EP-142, equidistant from Sak-heri Kabir and Ishan Khaiber, though it may be of EarlyDynastic I age. These, however, are the only recognizedexceptions to the pattern of larger settlements during thisperiod. The aggregate of 53 hectares of settlement suggestsa regional population of no more than 6,600 to 10,500.Unfortunately, the absence of a network of small sites pre-

cludes any estimation of the area used for cultivation, andtherefore a population density estimate cannot be made.

The succeeding Early Dynastic I period has been thefocus of a previous lengthy study (Wright 1969), and onlya brief summary is necessary to this consideration of thelong-term changes in settlement. During this period thereis no evidence of occupation on the plain of Eridu, but theUr area prospered (fig. 19). Ur itself expanded farther tothe south, and if similar growth occurred elsewherearound its margins it covered about 21 hectares. The cen-tral precinct was replanned on a larger scale. Small,densely packed residences are known, and there mayhave been larger residences and administrative buildingsfrom which came the many clay sealings and tablet frag-ments found dumped over the south revetment of thetown. Sakheri Kabir (EP-30), 8 kilometers up the Eu-phrates channel, was a smaller town covering about 8hectares. Between Ur and site EP-30 were three small vil-lages ranging from 0.4 to 1.5 hectares. Test excavations onthe largest of these, Sakheri Sughir (EP-47), indicated thatit was a large community of structures dispersed alonga canal bank, not unlike the villages of the area today.There was evidence of grain cultivation, the herding ofsheep and cows, fishing, and some ceramic production.Near these villages at the abandoned village of Tell al-'Ubaid (EP-8) there was a cemetery and an oval structure-perhaps a shrine and rural administrative center. Twoshort canals 2 to 4 kilometers long are in evidence nearthese settlements, and doubtless others existed that arenow covered by silt. Farther down the Euphrates channelsand on side canals below Ur there were at least two moresmall rural settlements. The cultivated enclave around Urprobably totaled 90 square kilometers, and the settledpopulation no more than 6,000, suggesting no more than66 persons per square kilometer of possibly cultivatedland.

During the later portion of the Early Dynastic period, orEarly Dynastic III, there is a reorganization in the patternof settlement. Ur itself certainly grows. Areas of domestichousing disappear from their former locations, and largebuildings of various uses replace them. The rebuiltcentral precinct, the royal cemetery, and various cunei-form documents attest to the accomplishments of Ur'sdynasty. It is possible that Ur approached its full extent of50 hectares at this time. Sakheri Kabir remains occupied,and the rural shrine at al-'Ubaid is rebuilt, but the nearbyvillages show no trace of occupation. Rural settlementsurvives only on the channels or canals below Ur. Site EP175-76, 9 kilometers distant and covering 3.5 hectares,may be of this period. Within the Eridu plain there arenew developments. First the north mound at Eridu (EP-46) becomes the site of two palaces, though there is littleother evidence of occupation (Safar 1950, pp. 31-33). Var-ious repairs on the main temple platform may date to thesame time. Second, at the breach in the Hazim, a tiny site

327

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

/

K. 7

/4? N~.

142 I'I ~%

II ~'.

iij~'* 'r~~

\gUr,~N.

KN

===.=-=.--=.-- -,

\/

*1~~

1\

*N

'. -w

' *i . ./

Fig. 19. Settlements of the Early Dynastic I period.

(EP-134) is reoccupied, and several buildings with founda-tions of baked planoconvex brick are erected along thechannel's bank. These sites, the sporadic occurrence ofEarly Dynastic artifacts on predominantly later sites (EP-31 and EP-96), and textual mention of administrators ofEridu as sending goods to Ur (Burrows 1935, suppl. 16,44) indicate that water was once again flowing in thesouthernmost channel of the Euphrates. Whether this wasa planned alteration of the channel or not, the palaceconstruction suggests that the opportunity to expand fieldarea was seized by the authorities at Ur.

THE FIRST EMPIRES (AKKADIAN TOPOST-CASSITE)

Ur's period of independence and prosperity was brief;toward the end of the Early Dynastic period the town wasunder the control first of the dynasty of Lagash, then thatof Uruk (Gadd and LeGrain 1928, p. 3). It then passed un-der the dynasty of Agade, a footnote to Sargon's conquestof Uruk. The polity of Sargon and his successors was ap-parently the first to rule not only all of alluvial Mesopo-

tamia, but-at various times-the Susiana, Assyria, andparts of Syria through the agency of appointed governorsand military garrisons. It controlled most of the resourcesused in its crafts and trades; it ruled most of southwestAsia's urban agglomerations; and it made a claim of uni-versal suzerainty that would be repeated by would-bestate-builders for centuries: "King of the Four Quarters"(Bottero, Cassin, and Vercoutter 1965, 109), a claim ofsuzerainty over peoples speaking diverse languages andfollowing diverse cults. Thus, in the economic, political,and ideological senses, we may speak of the Akkadianpolity as an empire, even though it was plagued with re-bellion and of only short duration.

It is doubtful that the Akkadian rulers found it difficultto control the region of Ur and Eridu, for there were fewtowns and settlements on this southern desert border oftheir domain. The southernmost or Eridu channel had nosettlements with what we have termed Akkadian ceramics,and it was perhaps dry. Ur itself was inhabited, as wasits northern suburb of Diqdiqah (EP-12), but almost noarchitecture of the period has been excavated (Woolley1956, p. 50), and even an approximate estimate of its size

328

oi.uchicago.edu

Appendix: The Southern Margins of Siumer

is difficult. Sakheri Kabir (EP-30), upstream from Ur, ap-parently shrank to a village-sized settlement of only 2.0hectares. A new village-sized settlement was inhabitedonly 4 kilometers northeast of Ur (EP-17). The only othercertain Akkadian foundation is 26 kilometers northeastof Ur, apparently on the outer sweep of a large meanderin the Euphrates channel (EP-122E). This tiny village, littlemore than a hectare in area, was to become important inthe following centuries.

During the Akkadian period, Ur seems to have becomeeven more nucleated and isolated than it had been duringthe Early Dynastic period. It seems surprising that it roseso quickly to imperial control. Ur's ascendancy will re-main unexplained until more is known about the actualsocial and economic effect of the Gutian invasions andabout the local problems faced at this time by such com-peting cities as Uruk and Lagash. Although Lagash hascuneiform sources richer than those at any contemporaryMesopotamian site (cf. Genoulliac 1910), its hinterlandhas never been systematically surveyed; while Uruk'shinterland has been completely surveyed (Adams andNissen 1972), there is little textual or architectural

.1

/ ,

\ "\N

122

evidence of the( Akkadian period from the city itrself.At the end of the twenty-second century B.C., the 7?"hird

Dynasty of Ur, more stable than that of Agade tho,ugh ofshorter duration,, gained control of the lowlands of GreaterMesopotamia. The utility of the southern Sumer surveyin further elucidating this copiously documentced hegem-ony is limited by four factors. (1) Our ceram:ic indicatorsmay not have become widespread until -swell after thefounding of the dynasty and certainly continued in useafter its demise. (2) Ur, though the seat of the dynasty, wasnot conveniently ,situated for administration, and manyimperial functions were discharged from more central fa-cilities near Nippu r, Lagash, and other towns. (3) Theagricultural regime reestablished by the Ur III rulers wasso successful that th e intensive cultivation of the succeed-ing period has obscu red not only the smaller canal traces,but the main channels of the late third millennium B.C.For this reason they are symbolized only by dotted lineson figure 20. (4) The extensive excavations at Ur haveproduced few texts redevant to rural life.

The monumental buildings of Ur itself are well known,but little ordinary housing is exposed (Woolley 1974).

/ ,\7 1

(

, .

* -•

•p7

Fig. 20. Settlements of the Ur III and Early Larsa period.

329

K.

m lu

K

11

//

\Ur

oi.uchicago.edu

App endix: The Southern Margins of Sumer

The Third Dynasty walls enclose about: 50 hectares ofsettlefA area, a small area compared with (other sites to thenorth a.nd northwest. Of the other twenty-two settlementsoccupieal during the period, only the temple-ziggurat ofEridu has been cursorily examined, and. we must rely onsurface evidence. In contrast to Ur itself, there are twosmall centersI of about 8 hectares. On e is Tell al-Lahm(EP-172), perhaps occupied during the Early Dynasticperiod and certainly extensive by the early Larsa period,39 kilometers to\ the east-southeast of Ur (Safar 1950),which may be ancient Kisiga (Jacobsei. 1960, pl. 28). Theother is the site founded in Akkadian times (EP-122),which reached its mn\aximum size in Ur III-Early Larsatimes. The other settlements are sma11 and large villages.Those near Ur itself are\strung along the main channel anaverage of 4 kilometers apart, exce:pt for a vacant areasoutheast of Ur, perhaps the fields arid gardens of the city.Those on the now reopened southernmost or Eridu chan-nel average 8 kilometers apart, pe-rhaps an index of thethen recent redevelopment of this part of the survey area(fig. 20).

The total of site area occupied throughout this periodis about 107 hectares. Slightly less than half of this is thecapital itself, with almost all the remainder being in smallrural settlements (fig. 25). This suggests a doubling ofpopulation over that of the preceding centuries of the thirdmillennium, to at most 13,400 to 21,400. It is perhaps anindex of the prosperity of the period that 60 percent of theten sites whose last major occupation was in this period

had baked brick fragments on their surfaces.During the Late Larsa and Old Babylonian periods the

apogee of settlement in southern Sumer was reached.

Though Ur had lost its autonomy during the strugglesbetween successor states reacling for control of the rich-

est parts of the empire, its environs became even more

heavily populated and productive. Ur grew, its houses

spreading beyond the Third .Dynasty walls, though cer-

tainly not covering the full 500 hectares suggested by

Woolley (1965, p. 193); a minlimum size for this period

would be 60 hectares. Many pu blic buildings were main-

tained and rebuilt, and extensive areas of dense multistory

housing have been excavated. A large number of cunei-

form documents attest to the economic and political life

of the city (Figulla and Martin 1953). Integrated re-

analysis of these data from Ur and the archaeological sur-

vey data could solve a number of problems.The remaining fifty-seven recorded settlements with the

ceramics termed Late Larsa-Old Babylonian are, with one

exception to be discussed further below, relatively small.

Thus, the general size distribution established in the pre-

ceding period is duplicated (fig. 25) even though few of

the earlier sites continued, 89 percent of the Late Larsa-

Old Babylonian sites having no Ur III-Early Larsa occupa-

tion. There are six small centers of 5 to 10 hectares. One

of these (EP-77) is on the Ur channel, situated only 10

kilometers upstream from Ur (fig. 21). On the Eriduchannel, however, there is a continuous distribution ofsuch small centers. Tell al-Lahm (EP-172) is at the eastern-most limit of the survey area. Abu Ras'ain (EP-27) is 18kilometers upstream. Eridu South (EP-108), Safar (1950,p. 28), 3 kilometers southwest of the ancient town, isabout 23 kilometers farther. On the west edge of the sur-vey area are the two smallest sites of this size grouping,one (EP-31S) 26 kilometers farther up the channel, andthe other (EP-158) 38 kilometers up the channel fromEridu South. Most of these are on the north or left bank,away from the dunes and the Southern Desert beyond, asif this quarter was unsafe. It is in this period that the firstdefinite example of a small rectangular fort with cornerbastions is known (EP-166), on the left bank of the Eriduchannel (fig. 26). Across the river to the southwest risethe great dunes. To explain this aspect of settlement con-figuration in terms of a nomad "threat" almost a millen-nium before the camel gave peoples desert fast mobilityis reasonable, since there is a continuous distribution ofwells from the area just behind the dunes northeastwardtoward better-watered areas along the middle Euphrates.Furthermore, even in the third millennium there werenomads in the area, since, from Ur, there is a Late EarlyDynastic receipt for lard from Amorites (Burrows 1935,suppl. 29). Smaller than the centers are a range of regu-larly spaced small and large villages. Around Ur, settle-ments average only 2 kilometers apart, but those along theEridu channel occur, on the average, every 5 kilometers.Similar settlements are also spaced evenly throughoutthe area between the two channels in the northwest endof the survey area; given the slope of the land, it is rea-sonable to reconstruct a canal in this area, as is shownwith dotted lines on figure 21.

The exceptionally large site mentioned above is anunnamed town site on the west edge of the survey area(EP-34). Since it is low and does not appear clearly on theair photographs, it is difficult to measure; however, itcovers more than 45 hectares. On the ground, winderosion has emphasized small canals within the settlement,drains lined with baked brick in former streets, buildingfoundations of both baked and mud brick, and localizedconcentrations of basalt, copper, ceramic slag, and otheritems perhaps indicating workshops. Regrettably, no in-scribed pieces were found, and we have no idea of the

ancient name of this short-lived but substantial town. Theceramics from the site seem to have somewhat finer dec-oration and appendages than those on other sites of theperiod, and it is possible that the site represents a lateoccupation during the Old Babylonian period proper.Certainly this was a period in which there was ample ex-cuse for resettling populations in new areas. If this site islater, perhaps some of the small settlements contributingto the marked density of sites near EP-34 are of the laterpart of the period as well. Unfortunately there are other

330

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

//

K

k

N'\

/

Fig. 21. Settlements of the Late Larsa and Old Babylonian period.

possible explanations of the ceramic differences betweenthis large town and other sites of the period, and only fur-ther work on this interesting settlement cluster will clarifythese issues.

Even if these possibly later settlements are removedfrom consideration, there is an almost continuous net-work of settlements south and west of Ur. Assuming aminimum size for Ur of 60 hectares, the total settlementarea occupied is about 198 hectares, indicating a popula-tion of almost 25,000 to 40,000, less than a third of it inthe city itself. Even if 60 hectares is discounted for thepossible later group to the west, this is a greater area thanwas covered in the preceding Ur III-Early Larsa period,and population would have been at most 17,100 to 27,500.This is at variance with Woolley's suggestion of 250,000persons for Ur and the nearby villages alone (1965, p.193).

Given a total area potentially cultivable-that is, siltyand salt-free soils accessible to water-of about 130,000hectares, the available amount of land is quite sufficient tosupport even the maximum population. It may be thateither this approach toward near total use of the land sur-

face without the benefit of truly centralized agriculturalplanning, or the incessant political vicissitudes of the time,or the loss of unearned imperial revenues expended nearthe capital, led to reduced prosperity: only 39 percent ofthe thirty-three settlements whose last major occupationwas during the Late Larsa-Old Babylonian period hadbaked bricks on the surface, a marked decrease from thesixty percent of the preceding period.

The extensive abandonment of settlements between theUr III-Early Larsa period and the Late Larsa-Old Baby-lonian period raises the problem of social continuity. Eventhough the ceramic complexes are similar and clearly aredevelopmentally related, only about 10 percent of thesettlements of the earlier period were occupied in the lat-ter period. It is possible that there was a hiatus in occu-pation between the two periods, but there is little textualevidence for such. Alternatively, there may have been afairly abrupt series of channel shifts, perhaps precipitatedby irrigation works undertaken farther upstream, requir-ing much movement of villages.

Between the Late Larsa-Old Babylonian period and theCassite period there is also extensive abandonment. In this

331

)

~

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

case, however, abandonment is accompanied by a discon-tinuity in ceramic styles. The writing of legal and eco-nomic texts ceases at Ur about 1740 B.C. (Stone 1977, fig.2), and commercial life does not revive until about 1400B.C., approximately 250 years after the beginning of theCassite dynasty. Given this evidence, it is possible that thesmall settlements in the hinterlands south of Ur were littleoccupied for two or three centuries.

In any event, there is an impressive Cassite regenerationof settlement. Ur itself appears to remain contractedwithin its Third Dynasty walls, but many of its templeswere restored under the sponsorship of Kurigalzu II. Ofthe thirty-nine other recorded settlements, 79 percent hadno preceding Late Larsa-Old Babylonian occupation.There are four smaller towns, all on the southern channelthat runs past the ruins of Eridu (fig. 22). As before, thelowest is Tell al-Lahm (EP-172). Thirty kilometers abovethis and only 15 kilometers southwest of Ur is a town ofalmost 18 hectares (EP-89); 25 kilometers above this isthe next town (EP-100), and 35 kilometers beyond liesanother (EP-36). Thus, towns are slightly larger and morewidely spaced than during the Late Larsa-Old Baby-

lonian period. Several of these are on the south side of thechannel, as are a number of villages, indicating that therewas little threat from the desert at this time. Small andlarge villages are found on both the Ur channel and theEridu channel, but not in the land between. Spacing isirregular, and closely associated pairs are notable, suggest-ing that some sites may not be contemporary. This is theearliest period for which a number of small canals couldbe defined on the air photographs. Most striking are those7 kilometers east-northeast of the ruins of Eridu, wheresmall field canals and even possible furrows could be seenfrom ground level, outlined in the salt crust. This areadeserves detailed study (EP-23).

The total settled area of the Cassite period coveredabout 179 hectares, suggesting a population at most of22,500 to 36,000. However, with the abandonment of thearea between the two channels, only about 85,000 hec-tares would have been available for cultivation-still ade-quate but close to the limit. As in the preceding period,baked brick is not common. Only 43 percent of the twen-ty-three sites without major occupation after the Cassiteperiod have baked brick fragments.

/ -.----

/

\

/

KMS

K

Fig. 22. Settlements of the Cassite period.

332

/ °

IN

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

The rapid decline of the Cassite dynasty after 1200 B.C.

did not bring immediate ruin to the southern margins ofSumer. Though there is population decline and localabandonment, there is much continuity in both the ce-ramic technology and the settlements. Almost half the set-tlements with the ceramics termed "Post-Cassite" wereoccupied in the preceding period; only 53 percent of thesettlements are at locations unoccupied in Cassite times, astrong contrast to the lack of continuity between the pre-viously discussed time periods. However, this continuitydoes not obscure the evidence of decline: Ur itself hasonly a few evidences of the repair of temples, and none ofactual rebuildings. Its precise size is unknown, but it wasprobably smaller than in Cassite times. The former Urchannel survives only as a small canal (fig. 23) wateringthe immediate vicinity of Ur and a few nearby villages. Onthe southernmost Euphrates channel, only the three east-ernmost towns survive. Villages are spaced along thisstretch every 5 kilometers; the western portion of thischannel within the surveyed area was little occupied. Hereand there, small canal systems are visible on the airphotographs.

In the survey area as a whole, the 139 hectares of Post-Cassite settlement indicate a population of at most 16,800to 27,000 persons. There was little relative change in theprosperity of these communities, since, of the twenty-threesites without subsequent major occupation, about 39 per-cent had baked brick fragments on their surfaces. In addi-tion, a number of settlements, some of them quite small,were on the south side of the southern channel, unpro-tected from the deserts beyond, indicating little fear ofattack from that quarter. It is possible that this period ofcontinued prosperity and relatively peaceful local condi-tions southeast of Ur is related to the brief rise of the so-called Second Dynasty of the Sealands in the area cen-

tered to the east of Tell al-Lahm around the present Horal-Hammar (Brinkman 1968).

Dating to this period of the decline of settled life on thelower alluvium is a striking channel feature visible on theair photographs, beginning just east of site EP-183 (fig. 1).Here a channel, similar in form and size to the occupiedPost-Cassite channel, cut through the left bank levee andflowed for at least 20 kilometers. The location at which itdiverges slopes downward to the northeast, at a point

/*N

N

/

I\.

N

Fig. 23. Settlements of the Post-Cassite period.

333

/'

/

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

where it would be easy to break the levee, flooding a largearea to the east. Though flow continued for long enoughto establish a definite channel, there was no settlementalong its banks. Subsequently the river was diverted backinto its former course running by Tell al-Lahm, perhapsancient Kisiga (Jacobsen 1960, pl. 28), a known Sea-lands town (Brinkman 1968). This repair must have oc-curred during the Post-Cassite period, since a small canalrunning due north of a Post-Cassite village (EP-189)clearly crosses this channel. Without excavation, it cannotbe demonstrated that there was a deliberate cutting of themain Euphrates channel for military purposes, rather thana natural calamity. However, such a tactic would be ex-pected to leave precisely this kind of evidence.

The decline at the end of the Post-Cassite period can beseen as the end of single cycle of growth and decay span-ning a millennium and a half. Growth began during theAkkadian period about 2300 B.C., reached an apogee inLate Larsa times about 1800 B.C., and declined to anotherlow point before 900 B.C. However, this seeming demo-graphic cycle does not represent the trajectory of a singletype of cultural system or even of a single type of cul-tural system. In the first place, the growth and declineof settlement is apparently episodic. There was certainlya major decline of rural settlement between the LateLarsa-Old Babylonian peak and the Cassite peak,and there may have been a similar decline between UrIII-Early Larsa and Late Larsa-Old Babylonian. Otheroscillations may be obscured by our yet elementary under-standing of ceramic development. This episodic charactercontrasts with the apparently more continuous pattern ofgrowth and decline to the north and northeast (Adams,this volume), perhaps reflecting the marginal position ofsouthern Sumer, with its communities exposed to desertraiders and the difficulties of maintaining the flow ofwater in the southernmost Euphrates channel. In the sec-ond place, this long period of oscillating growth and de-cline is apparently one of considerable social and eco-nomic change: the earlier portion of the cycle is one inwhich central institutions, many of them direct organs ofgovernment, dominate the economy, while from LateLarsa times onward independent entrepreneurs and cor-porations seem dominant (Oppenhiem 1964, pp. 83-109).Perhaps the rise in importance of small towns in Cassiteand Post-Cassite times is a consequence of such basic so-cial and economic changes. Alternatively, however, therise of such towns may be a local phenomenon resultingfrom the loss of Ur's political importance after the fall ofthe Third Dynasty. Further textual and comparative set-tlement studies will eventually settle such issues.

DEATH OF A REGION (NEO-BABYLONIAN

AND LATER)

Settlements occupied during the period from Neo-

Babylonian to Seleucid (and perhaps early Parthian) timesare common in the environs of Ur and Eridu. It seems thatprosperity once again returned to the region and that Urand many small towns were thriving. But close examina-tion of the evidence indicates that this would be a mis-leading conclusion.

Ur itself had an incomplete and relatively low-densityoccupation at the beginning of this period. Though thetemples were restored on a grand scale by the Neo-Baby-lonian rulers and subsequently maintained by their suc-cessors (Woolley 1962), the city was difficult to supplywith water. Early in the period, water could still bebrought to Ur, and even to villages beyond, by a smallcanal running along the old Euphrates levee. However,this earlier construction is cut by one that leads water tothe gardens around Ur by a canal coming from the northon a raised causeway (fig. 24). Examination of the south-ern end of this causeway reveals that its fill was consoli-dated with ceramic slag. Water must have been raisedfrom the level of the Euphrates to the level of the cause-way by a device such as a current-powered waterwheel.

While the old Ur channel gradually died, the southern-most channel, running by the ruins of Eridu, still flour-ished. There were three substantial towns. The eastern-most was the northern extension of Tell al-Lahm(EP-173) covering 14 hectares; 43 kilometers upstreamwas Eridu South (EP-108), covering 22 hectares; 45 kilo-meters farther by the circuitous river was Tell ud-Dahaila(EP-148), also 14 hectares. Villages are irregularly spacedalong the channel between the towns, averaging 12 kilo-meters apart. Most of these settlements are on the northor left bank, one of the exceptions being a fort with cornerbastions (EP-163) opposite Dahaila (fig. 26). Apparentlydesert peoples were once again a threat. The area betweenthis southernmost channel and Ur was again watered by aseries of small canals, as it was in Late Larsa-Old Baby-lonian times. Several of these canals begin as straight,probably planned constructions but become quite sinuousat their tails, indicating that growth late in their historyof use was by accretion. Field patterns of this last majorperiod of settled occupation are visible in several places.

The total settled area of this period is 160 hectares, in-dicating a maximum population of 20,000 to 32,000.However, this period of time is long, and the irregularsettlement spacing and overlapping small canals suggestthat the settlements were not all contemporary. Of thethirty-six settlements without subsequent occupation, 31percent had baked bricks on their surface, which suggeststhat the villagers were not as prosperous as those of pre-ceding periods. However, when it is possible to date sitesto shorter spans of time within this period, it may bepossible to show that the initial time of canal restorationwas relatively prosperous, and there was a gradual oroscillating abandonment thereafter.

Within the surveyed area, there is only one small village

334

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

/

//'

* i

)

7

KmMS

Fig. 24. Settlements of the Neo-Babylonian and later periods.

site with definite Parthian ceramics (EP-54). This site isvery near the above-mentioned causeway leading watertoward Ur, and it is possible that this watercourse wasmaintained after settlement was no longer possible in therest of the area. Apparently the southernmost channel wasunoccupied; but one must remember that the three townsof the preceding period were high and badly erodedmounds. It is possible that evidence of lingering reducedoccupation in Parthian times has been removed by naturalforces.

Two Sasanian settlements were recorded. One is asquare fortified settlement of 5.0 hectares close to thepresent Euphrates near modern Bat-ha (EP-65) (fig. 26).A short distance to the west, and on the flanks of a Neo-Babylonian-Achaemenian village site, is a small moundcovered with glass sherds and glass cullet of a type thatis probably Sasanian in age (Robert Brill, pers. comm.).Apparently the settlement had resident craft specialists aswell as farmers and/or border guards. The other site withSasanian ceramics is a small possible cemetery (EP-32)beside the southernmost channel at the extreme westernend of the survey area. Is this the isolated cemetery of no-

mads, or was there some water moving down the southernchannel watering villages just beyond the survey's limits?Visible on the air photographs, which I did not see untilafter I returned to Baghdad, about 8 kilometers west ofthe survey border is a roughly rectangular walled town,covering 55 hectares and divided into three parts. It issimilar in plan to Jundi Shahpur and Iwan-i Karkheh insouthwestern Iran (Adams 1961, figs. 5, 7) and is proba-bly a Parthian or Sasanian center. Perhaps when the pres-ent levee of the Euphrates is examined in detail we willfind that EP-65 is one of a series of fortified settlementslinked to this larger town, with installations such as al-Qusair (Finster and Schmidt 1976, pp. 49-54) beingeither elements in an outer line of defense or posts de-signed to protect routes coming from the south.

There is no indication of Islamic occupation until rela-tively recent times. It is therefore reasonable to ask whythe southernmost channel of the Euphrates ceased tocarry water. It is possible that the reasons are purelygeological. For example, the causeway canal that allowedlife to continue at Ur implies that the Euphrates waterlevel was lower than previously. Such geological explana-

335

1 ý3I f

/

I\

V.N111

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

LATE 'UBAID4

1 12 n.0 4 8 12 16 20

in , . _____

40

0- 3.6- 9.6- 22.6-3.5 9.5 22.5 60

EARLY DYNASTIC

UR III - EARLY LARSA

UR - 50 H

NUMBER

1 4 LATE LARSA -

1210 o0LD BABYLONIAN6 OUR -60. H

6 34-45 H

0 4 8 12 16 20

AREA IN HECTARES

CASSITEUR-50H

POST - CASSITE

UR -50 H

NEO BABYLONIAN

ACHAEME N IAN10

8

UR-40 H

0 4 8 12 16 20

TOTAL

HECTARES

120100

6020

0- 3.6- 9.6- 22.6-3.5 9.5 22.5 60

AREA CATEGORY

1 n n

60

0- 3.6- 9.6- 22.6-3.5 9.5 22.5 60

Fig. 25. Patterns of settlement and population in southern Sumer.

tions can be tested only by a competent geomorphologistwith accurate maps in hand. Alternatively, it is possiblethat the reasons are cultural. Camel-mounted nomads ofthe first millennium B.c. and after were doubtless far moredifficult to handle than their predecessors. It is likely thatthe evidence needed to answer the problem will lie notwithin the limits of the Ur-Eridu survey area, but to thenorthwest in the land between modern Samawa and mod-ern Hilla, one of the three large tracts of the lower al-luvium (along with that along and east of the Shatt al-Gharraf and that along the Tigris) that have not yet beensystematically surveyed.

CONCLUDING REMARKS

The geographical history of southern Sumer is forthe most part a footnote to events occurring elsewhere.While specific events here may be relevant to particularperiods of development in the more densely settled cen-tral regions of the alluvium, there is at least one generallesson to be drawn from this survey. This is that settlementin this border area is distinctly episodic, with frequent

periods of collapse. While there are such interludes inthe central regions, they seem less frequent. If we areto move beyond the simple recording of such historicalepisodes to a processual understanding, two lines offuture research must be pursued.

One such line is more detailed survey, with a focuson more limited periods of time. The value of resurveyusing more refined chronologies and more rigorous fieldobservations of sites is well illustrated by Adams's re-consideration of the Uruk period presented in this volume.Furthermore, this reconsideration has raised issues thatcan be answered only with even more intensive surveyinvolving small-scale excavations on samples of rural sitesand on canal and field remains. In southern Sumer, com-bined resurvey/excavation programs might focus onUbaid to Jemdet Nasr period settlements and canals be-tween Eridu and Rejibah, on Late Larsa and Old Baby-lonian remains around the town site of EP-34, or on Cas-site and Post-Cassite settlements, canals, and fieldsbetween Eridu South and Tell al-Lahm.

Another future line of research is the survey of theSouthern Desert beyond the great dunes. It is there that

336

oi.uchicago.edu

&

7

\ 1.5 M. Contours

s * C honnel?

N,' Mud Brick

Sherd

Co ncentration

8g

N

166

* N*

\- \-9K \ "

Fig. 26. Selected settlement maps.

337

-N-

72

163

1%

MC

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

direct evidence of nomadic people must be sought. Al-though the methods needed to do such survey are onlynow being developed, and though it seems certain that

their application will prove time-consuming and tedious,such survey must be undertaken if we are to gain anunderstanding of the oscillation of imperial control.

Eridu-Ur Survey Data: Site CatalogINTRODUCTION

This catalog presents most of the actual data uponwhich the foregoing is based. For each site it gives the sitenumber, the local and ancient names when known, thelength (plus orientation), width, and height in meters, andthe existence of any baked bricks (with measurements incentimeters), slag, stone, or other features of interest.After this, for each period on the site, it notes the areasin hectares estimated to have been covered during thatperiod, the artifact categories ascribed to that period,using acronyms previously defined, and the counts of eachcategory in the samples. If an artifact category is foundin several periods, the counts are entered in the earliestsuch period thought to be present on the site. These countsinclude neither all categories of diagnostic sherds foundon a site-since very large vessel rims and bases were notcollected-nor even all categories present in a sample,since unique pieces were not categorized in the precedingdiscussion. However, individual drawings and measure-ments were made of such pieces, and they are included inthe field records filed at the Iraq Museum in Baghdad, theOriental Institute in Chicago, and the Museum of Anthro-pology in Ann Arbor. Given the collection procedures dis-cussed in this introduction, interested scholars shouldprobably first reduce the counts to a more general scalesuch as "rare," "present," and "common" before engagingin statistical manipulations.

CATALOG

1 Ishan Umm Fushij. 150 NE X 115 X 3. Brokenbrick; ceramic slag. Late Larsa: 1 ha (RJ7-5:4, RJ8-1:4, RJ8-3:1, RB8-4:1).

2 Tell Ghaghla Gharbi. 350 WNW X 180 X 3.Brick, 28 X 28 X 7 cm. Late Larsa: 3:5 ha (RJ7-5:2, RJ8-1:6, RJ8-2:1, RJ8-3:1, RB8-1:1, RB8-2:2,RB8-4:8, B7-1:4, B8-1:2); Post-Cassite: trace (RJ9-2:1, RB9-2:3).

3 Abu Shahrain, Eridu. Eridu, Hajji Muhammad,Early Ubaid, Late Ubaid: 12 ha; Terminal Ubaid,Uruk, Jemdet Nasr-Early Dynastic: trace; Ur III:trace (see Campbell-Thompson 1919; Hall 1930;Lloyd and Safar 1948; Safar 1947, 1950; Taylor1855).

4 Rejibah Jinub. 450 NW X 280 X 6. Stone foot-ings; brick, 30 X 30 X 6.5 cm, 29 X 29 X 6.5 cm,30 X 15 X 7 cm. Jemdet Nasr-Early Dynastic: 8.2ha (RJ5-2:2, RB5-2:8, B5-3:4, B5-4:2, B5- 5:1, 05-

2:1); Late Larsa: 3 ha (RJ7-5:8, B7-1:2, RJ8-1:5,RJ8-2:3, RB8-1:4, RB8-2:2, B8-1:3, RB9-2) (seeWoolley 1956).

5 Rejibah Shamal. 550 X 225 X 2. Ceramic slag;limestone wall footings. Late Ubaid? (03-1 small,clay axe, interior combed sherd); Uruk? (RB4-1:1);Jemdet Nasr: 12.5 ha (RJ4-1:4, RJS-1:1, RJ5-2:5,RB5-2:17, RB5-2:20, B5-3:13, B5-4:4, 04-4:1);Late Larsa: trace (RB8-1:2, B7-1:1).

6 - 125 NNW X 85 X 1. Post-Cassite. 0.9 ha(B7-2:4, RB8-1:1, RB9-2:2, B9-1:4).

7 - See map, figure 26. Much limestone; ceram-ic slag; chert flakes. Late Ubaid: trace (RB3-11:1.stone hoe); Uruk: 0.7 ha (RJ4-1:2, RJ4-2:3, RB4-1:4, 04-1:2, 04-2:4, 04-5:1, 04-8:1).

8 Tell al-'Ubaid. Eridu, Hajji Huhammad, EarlyUbaid, Late Ubaid, Uruk, Jemdet Nasr, Early Dy-nastic I, Early Dynastic III, Ur III (see Hall 1930;Hall and Woolley 1927).

9 Not assigned.10 Tell al-Muqayir, Ur. Eridu, Hajji Muhammad,

Early Ubaid, Late Ubaid: ca. 10 ha; TerminalUbaid: 10 ha; Uruk: 10 ha; Jemdet Nasr: ca. 15ha; Early Dynastic I: 21 ha; Early Dynastic III,Akkadian, Ur III: 50 ha; Late Larsa: 60 ha; Cassite:50 ha; Post-Cassite, Neo-Babylonian: 40 ha (seeBrinkman 1968; Burrows 1935; Gadd and Legrain1928; Hall 1930; Oates 1960; Woolley 1939, 1956,1962, 1965, 1974).

11-12 Diqdiqah: 12. 305 NE X 230 X 4. Ceramic slag.Akkadian: 5.9 ha; Ur III (see Woolley 1956, 1974).

13 - 170 N X 130 X 2. Late Larsa: 2.5 ha (B7-1:5, RJ8-1:4, RJ8-2:1, RB8-2:1, RB8-4:3, RB8-5:2,B8-2:1); Cassite: trace (RB9-2:1, RB9-5:1); Neo-Babylonian+: trace (RJ10-1:3 without ridge, B10-1:1, B10-2:1, B10-3:1).

14 160 WNW X 70 X 0.5. Late Larsa: 1.1 ha(RJ8-1:1, RB8-2:1, RB8-4:2 [1 with hatched stripand incised curved line], RB8-5:2 thick, B7-1:1,B8-1:1, 08-1:2).

15 - Small sherd scatter around recent gravelpit. Cassite: trace (RB9-1:1, RB9-2:1, B7-2:1,B9-1:1).

16 Tell Ba'arura Jinub. 225 NE X 200 X 3. Brokenbrick. Early Dynastic (RJS-2:1, RJ5-5:1, RB5-2:1,RJ6-1:1?); Ur III-Early Larsa: 3.7 ha (RJ7-3:9,RJ7-5:3, B7-1:8, B7-2:1, B7-5:1, 07-1:6); Late

338

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

Larsa (RJ8-1:9, RJ8-3:2, B8-1:1); Cassite-Post-Cassite (RJ9-1:1, RB9-1:2, RB9-2:2, RB9-3:1,RB9-5:1, B9-1:3).

17 Tell Sughariyya. 110 NW X 80 X 3. Broken brick.Akkadian?: 0.8 ha.

18 Tell Ba'arura Shargi. 230 NE X 130 X 2. Ur III-Early Larsa (RJ7-1:1?, RJ7-5:2, B7-1:3, B7-2:4; UrIII inscribed brick fragment); Late Larsa? (RJ8-3:2,RB8-5:2, B8-3:1); Cassite: 2.5 ha (RB9-1:6 [1thick], B9-1:1).

19 Tell Ba'arura Sughir. 170 N X 120 X 1.5. EarlyDynastic? (stone bowl sherd, chert core); Cassite:1.8 ha (RB9-2:5, B7-1:1, B7-2:2); Neo-Babylo-nian +? (B10-2:1).

20 Tell Ba'arura Shamal. 300 NE X 150 X 2.5. Cas-site: 3.5 ha (RJ7-5:5?, RJ8-1:1, RB8-1:4, RJ9-1:1,RB9-1:3, RB9-2:4, RB9-3:1, B9-1:1). Recentqual'a ruin.

21 Ishan Beni Sa'id. 90 WNW X 65 X 3. Early Dy-nastic (05-6:1); Cassite-Post-Cassite: 0.6 ha (RJ7-5:1, B7-1:1, RJ9-1:3); Neo-Babylonian +: 0.6 haRJ10-1:2, RJ10-4:1, RB10-2:1, B10-1:1, B10-2:1,010-4:1).

22 Maftul Shuwemi. 120 NW X 75 X 1.5. JemdetNasr-Early Dynastic: 0.6 ha (RJ4-1:1, RJ5-1:2,RJ5-3:2, RB5-2:10, RB5-3:1, B5-1:1, B5-2:1, B5-3:6); Recent occupation and mud-brick tower.

23 - Area of small canal and plow marks visiblein salt. Cassite? (B9-1:2, 7 large vats or coffins).

24 - 150 NE X 90 X 3. Brick fragments; ce-ramic slag; carnelian fragments. Late Larsa: 10 ha(RJ7-5:2, B7-1:1, RJ8-1:3, RJ8-2:1, RJ8-3:4. RB8-1:1 thick, RB8-3:3 [2 with incised curved line],RB8-5:3 [2 thick], B8-1:5, 08-1:1); Post-Cassite:1.0 ha (RJ9-2:1, RB9-1:3 [1 with incised wavy line,2 thick], RB9-2:7, B9-2:1).

25 310 NE X 220 X 2. Brick fragments. Cas-site: 3.5 ha (RJ7-5:2, B7-1:2, B7-3:1, RJ9-2:3,RB9-1:1, RB9-2:4, B9-1:2, 09-1:1).

26 - Thoroughly plundered cemetery. EarlyLarsa to Cassite.

27 Tell Abu Ras'ain. 310 NNE X 280 X 4. Brokenbricks. Early Larsa (RJ7-1:1, RJ7-5:1, B7-1:2, 07-1:1); Late Larsa: 6 ha (RJ8-1:10, RJ8-2:3, RJ8-3:1,RB8-1:2 [1 with incised curved line], RB8-3:3,RB8-5:5 thick, B8-1:1, B8-2:1, 08-1:6); Cassite:trace (RB9-1:3 with wavy combed decoration,RB9-5:1) (see Campbell-Thompson 1919).

28 Tell Abu Salabikh. 270 N X 150 X 2.5. Brokenbricks; ceramic slag; carnelian debris. JemdetNasr-Early Dynastic? (RB5-2:1, stone bowl sherd);Late Larsa: 3 ha (B7-1:1, B7-3:1 without button,RJ8-1:12, RJ8-2:3, RJ8-3:1, RB8-2:1, RB8-3:2,RB8-4:2 with large groove, RB8-5:1, B8-1:3, B8-2:3, 08-1:7, 08-2:1); Neo-Babylonian +: trace

(RJ10-1:1 without ridge, B10-1:1) (see Campbell-Thompson 1919).

29 Merejib. 170 E X 150 X 2. Gypsum cement brickfootings. Early Ubaid (RJ2-1:2, RB2-1:8, RB2-3:1); Late Ubaid (RB3-2:2, RB3-11:1, RB3-12:6,RB3-14:1, 03-1:14, 03-2:2); Uruk: 1.6 ha (RJ4-1:3, 04-1:1, 04-2:2, 04-3:1, 04-5:1); JemdetNasr-Early Dynastic: 1 ha (RJ5-1:4, RJ5-2:2, RJ5-3:1, RJ5-4:2, RB5-2:1, RB5-3:1, RB5-7:2, B5-3:2,05-2:3, 05-4:3, 05-5:1, 05-6:2, 05-7:2); Cassite-Post-Cassite?: 0.2 ha (RJ7-5:1, B7-1:1, B7-2:1).(See Woolley 1956.)

30 Tell al-Sakheri. 360 N X 300 X 6. Bricks, 31.5 X31 X 6.5 cm, 27 X 16.5 X 7.5 cm, 25 X 16.5 X7.2 cm, 23 X 14 X 6.6 cm; ceramic slag. JemdetNasr-Early Dynastic I, ca. 8 ha (RJ5-1:6, RJ5-2:1,RJ5-3:20, RJ5-4:4, RB5-2:44, RB5-7:2, B5-1:6, B5-3:7, B5-4:10, 05-1:1, 05-2:3, 05-3:1, 05-5:1, 05-7:3); Early Dynastic III, ca. 4 ha; Akkadian: ca. 2(RJ6-1:1, RJ6-2:1, RJ6-3:2); Ur III-Early Larsa:ca. 3.5 ha (RJ7-1:6, B7-4:1); Late Larsa?: trace(RB8-4:1).

31 south Tell Khorsadah. 310 N X 240 X 4. Bricks, 36X 36 X 6 cm, 35 X 35 X 7 cm, 37 X 37 X 7 cm,35 X 17 X 8 cm, 24 X 19 X 8 cm. Early Dynastic(RB5-1:1, stone bowl sherd); Late Larsa: 6:5 haRJ7-2:1?, RJ7-5:3, B7-1:1, RJ8-1:3, RJ8-3:1, B8-1:1, B9-3:1); Neo-Babylonian +: 3 ha (RJ10-1:8[5 without ridge], RJ10-2:4, RB10-2:1, RB10-3:1,RB10-4:1 thick, B10-2:3, B10-4:1, 010-3:1).

31 center - 230 E X 100 X 2. Uruk cultivation (04-8:1); Neo-Babylonian +: 1.9 ha (RJ10-1:9 [7 with-out ridge], RJ10-5:1, B7-1:2, B10-2:1, B10-4:2,010-4:1).

31 north - 90 E X 70 X 1. Late Larsa: 0.5 ha (RJ8-1:3,RJ8-2:1, RJ8-3:1, B8-1:1, 08-1:1; carnelian frag-ments).

32 - Ca. twenty small pottery concentrationsin an area ca. 300 m in diameter, perhaps aneroded cemetery. Sasanian?

33 230 E X 170 X 1. Ceramic slag. Uruk cul-tivation (04-8:1); Late Larsa? (RJ8-1:1, RB6-1:3,RB8-5:2, RB9-5:1, 08-1:1; Neo-Babylonian+: 2.7ha (RJ10-1:5, RJ10-2:3, RJ10-3:1, RB10-1:1, B10-1:1, B10-2:1, RB10-1:1, B10-1:1, B10-2:1, 010-3:1, 010-4:1, 010-9:1).

34 Ca. 1,300 ENE X 900 X 2. Bricks, 32 X 32 X6.5 cm, 38 X 38 X 6 cm, 33 X 16 X 6.5 cm,51 X 50 X 8 cm, 65 X 64 X 7 cm; ceramic slag;carnelian; flat cuprous fragments. Late Larsa: ca.85 ha (RB5-2:1, RB6-1:2, B7-1:5, RJ8-1:23, RJ8-2:3, RJ8-3:1, RB8-4:12, RB8-5:1, B8-1:8, B8-2:2,B8-3:1, 08-1:1, 08-2:2).

35 490 N X 275 X 2. Brick fragments. Ur Ill-EarlyLarsa (RJ7-1:2?, RJ7-3:4, RJ7-4:2 large, B7-2:2

339

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

large, B7-5:2); Late Larsa? (RJ8-1:2, RJ8-3:3);Cassite: 9.8 ha (RJ9-1:1, B9-1:1, 09-1:2).

36 320 N X 275 X 1. Ubaid-Uruk? (03-1:1, 04-7:1);Ur III-Early Larsa: 5.2 ha (RJ7-3:1, RJ7-5:2, B7-1:2, B7-2:2, B7-5:2); Late Larsa? (RJ8-1:2, RB8-4:1); Cassite: 5.2 ha (RB9-1:3, RB9-2:4, RB9-3:2,B9-1:3).

37 - Outside survey area.38 Wahashit al-'Ash'ali. Outside survey area. Sherds

scattered on sand dune. Ubaid (RB2-1:1, RJ3-1:4,RJ3-2:2, RB3-1:5, RB3-2:4, RB3-4:4, RB3-5:2,RB3-7:1, RB3-8:2, RB3-14:1, RB3-15:1, 03-1:15).

39 150 E X 120 X 1. Bricks in square pit. Late Larsa:1.6 ha (B7-1:2, B7-3:1, B7-5:2, B7-6:1, RJ8-1:5,RJ8-2:4, 08-1:4); Neo-Babylonian +? (R10-1:1,without ridge).

40-45 Outside survey area.46 Eridu Shamal. 250 NNW X 225 X 2. Jemdet

Nasr-Early Dynastic: 2.7 ha (RJ4-1:4, 04-4:4,RJ5-3:5, RJ5-4:1, RB5-2:27, RB5-3:1, B5-3:5, B5-4:19, 05-2:4); Ur III-Early Larsa: trace (RB8-1:1)(see Safar 1950).

47 Sakheri Sughir. 230 E X 110 X 0.5. Bricks. EarlyDynastic: 1.7 ha (RJ5-1:5, RJ5-3:5, RJ5-4:1, RJ5-5:4, RB5-2:4, RB5-7:6, B5-3:2, B5-4:1, B5-5:1, 04-1:1, 05-2:5, 05-3:3, 05-4:2, 05-5:1, 05-7:4);Larsa? (RB6-1:1); Neo-Babylonian+ ? (RJ10-1:1)(see Wright 1969).

48 Tell Shaman Gharbi. 130 E X 120 X 2. Bricks.Late Larsa? (RB6-1:1, RJ7-5:1, RJ8-1:2, B8-2:1);Neo-Babylonian+: 1.1 ha (RJ10-1:9 [1 hatchedridge, 1 without ridge], RJ10-2:2, RJ10-3:1, RB10-1:3, RB9-2:1, B10-1:2, 010-3:1 leg, 010-4:2, 010-7:1, 010-8:1, 010-9:2).

49 Line of ceramic slag piles.50 Tell Sakheri Shargi. 285 N X 250 X 3.5. Bricks.

Neo-Babylonian+: 2.9 ha (RJ10-1:16 [1 glazed],RJ10-2:6 [1 glazed], RJ10-3:2, RJ10-4:1, RB10-2:9, RB10-3:4, RB10-4:3, RB10-5:2, B10-1:7, 010-3:1, 010-6:3).

51 290 E X 140 X 2. Bricks. Ubaid cultivation(03-1:1); Ur III-Early Larsa: 3.9 ha (RJ7-1:1, RJ7-2:4, RJ7-3:4, B7-2:5, B7-3:3, B7-5:3); Cassite: 3.9ha (RJ8-3:5, RB8-1:1, RB9-1:3, RB9-2:5, RB9-3:1,B9-1:2, B9-2:1, 09-1:1), inscribed brick.

52 (71) - 90 NW X 70 X 2. Bricks, 23 X 15 X6.5. Jemdet Nasr-Early Dynastic cultivation (05-7:1); Late Larsa: 0.4 ha (B7-1:1, B7-3:1, RJ8-1:12,RJ8-2:2, RJ8-3:8, RB8-4:12 [10 with large groove],RB8-5:4 thick, B8-1:2, B8-2:1, 08-1:8); Post-Cassite: trace? (RJ9-2:1, RB9-2:4); Neo-Baby-lonian+: trace? (RJ10-1:1 without ridge, RJ10-2:1, B10-4:1).

53 Tell Sakheri Jinub. 180 WNW X 150 X 3. Bricks;brick drain. Ur III-Early Larsa: 1.8 ha (RJ7-1:3,

54

55

56

57

58

59

60

61

62

63

RJ7-2:1, RJ7-3:5, RJ7-4:3, RJ7-5:2, B7-2:3, B7-5:2); Late Larsa: 1.8 ha (RJ8-1:5, RB8-2:1, RB8-4:1, RB8-5:1, 08-1:1); Cassite-Post-Cassite: 1.8ha (RJ9-1:1, RB9-3:1, RB9-1:1, RB9-3:2, RB9-4:2,B9-1:2); Neo-Babylonian+: trace (B10-2:1, B10-3:1).

160 N X 80 X 0. A scatter of sherds. LateLarsa?: trace (RB8-3:1, RB8-5:1, B8-2:1); Neo-Babylonian+: 0.8 ha (RJ10-1:12 [1 without ridge,1 with hatched ridge], RJ10-2:6, RJ10-4:2, RB10-1:2, RB10-2:2, RB10-3:3, B10-1:1, 010-3:1, 010-4:3, 010-8:2).

- 310 W X 125 X 2. Bricks. Ubaid? (03-1:1,03-3:1); Ur III-Early Larsa: 2.8 ha (RJ7-3:4, RJ7-5:5, B7-1:4, B7-2:1, B7-3:3, B7-5:1); Cassite: 2:8 ha(RJ8-1:2, RB8-1:4, RJ9-1:3, RB9-1:5, RB9-2:7,B9-1:3).

-- 220 WNW X 130 X 2. Bricks. Ur III-Larsa? (RJ7-5:4, B7-2:3, B7-5:2, RJ8-3:1, RB8-1:1,B8-3:2); Post-Cassite: 2.2 ha (RJ9-1:3, RJ9-3:1large, RB9-1:7, RB9-2:10, RB9-3:1, RB9-5:1, 09-1:1).

-- 100 NNW X 80 X 0.5. Bricks .Late Larsa:0.4 ha (B7-1:1, B7-2:1, B7-6:1, RJ8-1:2, RJ8-3:1,08-1:1); Cassite: 0.4 ha (RB9-1:3, RB9-2:5, RB9-3:1, B9-1:2, 09-1:1).Tell Umm al-Dhab. Bricks. Ur III-Early Larsa: 3.1ha (RJ7-1:1, RJ7-2:9, RJ7-3:11, RJ7-4:4. B7-2:1small, B7-4:3, B7-5:2, B7-6:5, 07-1:1); LateLarsa+: trace (RJ8-1:3, RJ8-3:5, RB8-2:2, 08-1:3, RB9-2:2) (see Jacobsen 1960, "SugheriWest").

60 E X 40 X 1. Late Larsa? (B7-1:2, B7-2:3, RJ7-5:6, RJ8-1:2, RB8-1:2); Cassite: 0.2 haRJ9-1:3, RB9-1:3, RB9-2:8, RB9-3:1, B9-1:1, B9-2:1).Ishan Khaiber. 265 NE X 245 X 3. Bricks, 27 X18 X 7.5 cm; ceramic slag. Late Ubaid (03-2:1?,03-3:1, painted sherd); Jemdet Nasr: 3.5 ha (RJ4-1:3, 04-2:1, 04-3:1, 04-4:1, RJ5-1:2, RJ5-4:1,RB5-2:23, RB5-3:15, RB5-4:12, RB5-5:7, B5-3:1,B5-4:13, B5-5:1, 05-2:1, 05-7:2); Ur III-EarlyLarsa: 0.8 ha (RJ7-1:3, RJ7-2:6, RJ7-3:3, RJ7-4:2,RJ7-5:2, B7-2:1, B7-4:1, RJ8-1:1); Cassite: trace(B9-1:1).Tell Gurra. 290 NW X 100 X 3. Late Larsa? (RJ7-5:2, B7-1:2, B7-2:2 small, B7-4:1, RB8-4:2); Post-Cassite: 2.3 ha (RB9-1:1, RB9-2:3, RB9-5:1, B9-1:2, 09-1:1).

- 80 NE X 60 X 0. Sherd scatter. Neo-Baby-lonian+: 0.3 ha (B7-1:1, B9-3, RJ10-1:8 [1 with-out ridge], RJ10-5:2, B10-4).Umm al-Ghemimi. 170 NW X 100 X 4. Uruk?(O4-2:1, RB5-6:1, stone sherd); Neo-Babylo-nian+: 1.5 ha (RJ8-1:3 small, RJ10-1:3, RJ10-2:2,

340

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

RJ10-4:3, B10-1:5, 010-3:2, 010-4:1, 010-5:1,010-8:1, 010-9:1).

64 - 100 NW X 60 X 2. A small site with fewsherds but much glass cullet and two cementbasins. Perhaps associated with site 65, thoughclose to site 63.

65 260 X 260 X 2. See map. Sasanian: 5 ha.66 270 NW X 120 X 1. Late Larsa? (BJ8-1:1,

RB8-3:2 thick, RB8-5:6 [4 glazed]); Neo-Baby-lonian+: 2.1 ha (RJ10-1:20 [4 without band, 1with hatched band], RJ10-2:5, RJ10-4:1, RB10-1:2, RB10-2:1, RB10-3:2 glazed, B10-1:3, 010-3:2, 010-4:3).

67 Tell Abu Khumoyis. 150 X 150 X 3. Only a fewSasanian and Islamic sherds on what may be asolid mud-brick mass.

68 - 50 NW X 40 X 0.5. Late Larsa: 0.2 ha(B7-1:3, RJ8-1:2, RJ8-2:2, RB8-2:1, RB8-5:2 thick,08-1:4); Neo-Babylonian+: 0.2 ha (RJ10-1:2without ridge, RJ10-5:1, RB9-2:2, B10-1:1).

69 Ishan al-Kharita. 155 E X 150 X 1. Bricks, 25.5 X25.5 X 6 cm, 25 X 10 X 6 cm. Neo-Babylonian +:1.6 ha (RJ8-1:1 small, RB8-5:2, RB9-2:1, RJ10-1:4[1 without ridge], RJ10-2:1, RB10-1:2, 010-4:1,010-9:2); possible Recent occupation.

70 145 N X 140 X 1. Ubaid cultivation? (03-1:1); Late Larsa: 1.5 ha (B7-1:1, RJ8-1:4 [1 small],RJ8-2:2, B8-1:1, 08-1:4, 08-2:1?); Neo-Baby-lonian+: 1.5 ha (RB9-1 thick, RJ10-1:3 [2 with-out ridge], B10-4:1, 010-9:1).

71 Same as site 52.72 85 E X 80 X 1. Bricks, 23 X 23 X 6 cm;

ceramic slag; possible square wall around site (seemap, fig. 26). Late Larsa: 0.4 ha (RJ8-1:8, RJ8-3:1,RB8-1:1, RB8-3:1, RB8-4:8, B8-1:2); Neo-Baby-lonian+: trace (B10-2:1, B10-3:2).

73 Ishan Karib Makina Muhammad. 390 WNW X16 X 2. Bricks, 26 X 16 X 6 cm; ceramic slag.Ubaid cultivation (03-1:1); Jemdet Nasr-EarlyDynastic: ca. 1 ha (RJ5-1:1, RJ5-2:1, stone bowlsherd); Ur III-Early Larsa: 3.7 ha (RJ7-3:7, RJ7-5:9, B7-1:3, B7-2:2 large, B7-5:3); Late Larsa: 3.7ha (RJ8-1:2, RJ8-3:1, RB8-1:2, 08-1:1); Cassite-Post-Cassite: 3.7 ha (RJ9-1:4, RJ9-3:2, RB9-1:7[3 with double carination], RB9-2:8, RB9-3:1,RB9-5:1, B9-1:2); fragments of inscribed ceramiccylinder referent to watercourses.

74 - Piles of baked bricks, 33 X 33 X 7 cm.Only one green glazed sherd associated.

75 Ishan Abu Dhib. 110 E X 105 X 1. Bricks, 27 X16 X 7 cm. Late Larsa: 0.8 ha (RB6-1:3, B7-1:2,RJ8-1:1, RJ8-2:7, RB8-3:2, RB8-2:1, RB8-4:7 [2thick, 3 with heavy groove], B8-3:1, 08-1:4, 08-2:1).

76

77

78

79

80

81

82

8384

85

86

220 N X 180 X 2. Bricks, 34.5 X 34.5 X8.5 cm, 19 X 19 X 7 cm. Late Larsa: 2.9 ha (B7-1:3, RJ8-1:10, RJ8-2:3, RB8-4:7, B8-1:5, 08-1:6),Neo-Babylonian+: trace (RB9-2:1, RJ10-1:3 with-out ridge, RJ10-2:1, B10-2:1, B10-3:1).Ishan Mazra 'a 'Ubaid. 350 NE X 320 X 3. Brickfragments. Late Larsa: 6.9 ha (B7-1:2, RJ8-1:5,RJ8-2:1, RB8-1:1 with incised curved line, RB8-4:3, 08-1:1 with incised curved line); Neo-Baby-lonian+: 6.9 ha (RJ10-1:11 [6 without ridge],RB10-3:1, RB10-4:2, B10-4:1).

- 145 NW X 85 X 1. Late Larsa: 1.2 ha (B7-1:1, RJ8-1:3, RJ8-2:1, RB8-3:1, RB8-4:3, RB8-5:1, 08-1:1); Post-Cassite: 1.2 ha (RJ9-2:1, RJ9-3:3, RB9-1:1).

100 NE X 90 X 0.5. Planoconvex brickfragments. Ubaid cultivation (03-1:1, 03-3:1);Early Dynastic: 0.7 ha (RJ4-1:1, RJ5-1:5, RJ5-2:2,RJ5-4:1, RJ5-5:1, RB5-6:9, B5-1:3, B5-2:4, B5-3:2,B5-4:2, 05-1:1, 05-2:4, 05-3:2, 05-4:4, 05-5:1,05-6:1, 05-7:4); Cassite-Post-Cassite: trace (B9-1:1).

70 X 70 X 0.5. Late Larsa: 0.4 ha (B7-1:1,RJ7-5:1, RJ8-1:5, RJ8-2:2, RB8-2:1, RB8-4:2 withlarge groove, B8-1:1); Cassite-Post-Cassite: trace(RB9-1:2); Neo-Babylonian+: trace (RJ10-1:1without ridge, B10-2:1). Indications of former ex-cavation.

- 115 NNE X 80 X 0.5. Late Larsa: 0.8 ha(RJ7-5:1, B7-1:1, RJ8-1:13, RB8-2:1, RB8-3:1,RB8-4:3, B8-2:4, B8-3:2, 08-1:4); Neo-Baby-lonian+: trace (RJ10-1 without ridge, B10-3:2).

-- 190 W X 120 X 0.5. Eroded sand knoll;planoconvex bricks, 21 X 13 X 8 cm. JemdetNasr-Early Dynastic I: 1.5 ha (04-1:1, 04-2:1,RJ5-1:7, RJS-2:4, RJS-4:1, RJ5-5:1, RB5-2:7,RB5-3:1, RB5-6:8, B5-3:5, B5-4:4, B5-5:2, 05-1:1, 05-2:6, 05-3:1, 05-4:1, 05-5:1, 05-7:1).Many looted graves.

Isolated Ubaid artifact (03-2).See map, fig. 26. 100 N X 80 X 0. Brick

fragments. Early Dynastic (05-6:1); Ur III-EarlyLarsa (RJ7-1:6, RJ7-2:3, RJ7-3:1, B7-1:2, B7-3:3,B8-2:1). Apparently a circular mud-brick plat-form, perhaps Early Dynastic in date, once coveredwith Ur III-Early Larsa refuse, now completelyleveled by erosion.

- 180 NNE X 115 X 0.5. Sherds on erodedremnant of Hazim. Cassite: 1.4 ha (RJ7-5:3, B7-1:1, B7-4:1, RJ8-3:1, RB8-1:2, RJ9-1:1, RB9-1:7,RB9-2:2, RB9-5:1, B9-1:1, B9-2:1, RJ10-2:1).Tell Ur Juncshen: 145 X 140 X 2. Late Larsa: 1.8ha (RJ7-5:1, B7-1:2, RJ8-1:2, RJ8-2:1, RB8-3:1,RB8-4:2, B8-2:1, B8-3:1 small, RJ10-1:1 withoutridge).

341

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

87 70 X 70 X 0.5. Ceramic slag; few sherds.Late Larsa?: 0.3 ha.

88 - 125 NNE X 75 X 0.5. Neo-Babylonian±+:0.8 ha (RJ8-1:1 small, RJ8-5:3, RJ10-1:7 [6 with-out ridge], RJ10-5:4, B10-1:1, B10-1:2, B10-4:2).

89 Merejib. 680 NNE X 280 X 3. Bricks, 28 X 28 X6 cm. Ubaid cultivation? (03-1:1); Late Larsa?(RJ7-5:5, B7-1:4, B7-2:3, B7-3:2, RJ8-1:1, RJ8-3:2, B8-3:1); Cassite-Post-Cassite: 17.8 ha (RJ9.1:1, RJ9-3:2, RB9-1:4, RB9-2:4, RB9-3:1, RB9-4:1, B9-1:3, B9-2:2, 09-1:1). (Not Woolley'sMerejib.)

90 - 80 NNW X 65 X 0.5. Late Larsa: 0.4 ha(RJ8-1:3, RJ8-2:2).

91 80 N X 70 X 0.5. Late Larsa?: 0.5 ha(RB5-2:1, B7-5:3, RJ8-1:1, RB8-3:2, RJ10-1:1without ridge).

92 - Eroded sand knoll. Late Ubaid cultivation(03-1:1, 03-2:1); Jemdet Nasr, Early Dynastic?:trace (stone bowl sherd).

93 220 NW X 125 X 0.5. Stone footings.Ubaid cultivation? (03-1:1); Jemdet Nasr-EarlyDynastic: 2.5 ha (RJ5-3:1, RB5-3:1, B5-4:2, 05-2:1); Post-Cassite: 2.5 ha (RJ7-5:4, B7-2:2, B7-3:1,RJ8-3:1, RJ9-2:2, RJ9-3:1, RB9-1:1 large, RB9-2:5, B9-3:2).

94 135 N X 125 X 0.5. Ubaid (03-1:1, 03-3:1); Ur III-Early Larsa: 0.6 ha (RJ7-1:2, RJ7-2:1,RJ7-3:5, B7-1:1, B7-2:4 large, B7-3:4 large, B7-4:4large, B7-5:3); Late Larsa?: trace (RJ8-1:1, RJ8-3:6); Post-Cassite: 0.4 ha (RJ9-1:1, RJ9-2:1, RJ9-3:1, B9-1:1).

95 - 145 N X 85 X 1. Recently looted gravesto southwest. Late Larsa? (RJ7-5:2, B7-1:3, B7-2:1, B7-4:2, RJ8-1:2, RB8-1:1); Cassite-Post-Cassite: 2.0 ha (RJ9-2:1, RB9-3:1, RB9-1:1, RB9-2:6, RB9-5:1, B9-1:4).

96 310 NE X 240 X 3. Brick, 29 X 28.5 X6 cm; concentration of copper debris. Ubaid: ca.2.4 ha (03-1:1, 03-3:1); Jemdet Nasr-Early Dy-nastic? (stone bowl sherd, cores, and blades); UrIII-Early Larsa: 4.5 ha (RJ7-1:5, RJ7-3:2, RJ7-4:3,B7-1:1, B7-2:2, B7-3:1 with paint, B7-5:2, 07-1:1); Cassite: 4.5 ha (RJ8-3:2. RJ9-1:2, B9-1:2,09-1:1).

97 80 NE X 70 X 1. Ubaid cultivation? (03-1:1); Late Larsa?: 0.5 ha (RB6-1:1, RJ7-5:1, RJ8-1:1, RB8-3:1, B8-2:4); Cassite?: trace (RB9-1:1).

98 - 125 NW X 100 X 0.5. Bricks in lootedtomb, 26 X 17 X 7.2 cm; ceramic slag. TerminalUbaid: 1.0 ha (RJ3-1:1, RJ3-2:1, RB3-1:1, RB3-4:3, RB3-5:5, RB3-6:2, RB3-7:1, RB3-8:7, RB3-11:1, RB3-15:4, RB3-16:3, O3-1:10 [some fusedtogether in manufacture], 03-2:8); Ur III-Early

Larsa: looted tomb (B7-2:1, B7-3:1). (PerhapsWoolley's "Rejibah X," 1956.)

99 100 N X 50 X 0. Ceramic slag. Late Larsa:0.5 ha (B7-3:1, RJ8-1:2, RJ8-3:1, RB8-1:1 withwavy line, RB8-4:1 with large groove, B10-2:1).

100 Eridu Jinub. 390 NW X 305 X 5. Bricks, 30.5 X30.5 X 6 cm, 25 X 24 X 6 cm. Ur III-Early Larsa?(RJ7-3:1, B7-1:1, B7-3:1, B7-4:2); Late Larsa: ca.5.0 ha (RJ8-1:3, RB8-1:1, RB8-4:3, B8-1:5, 08-1:4); Cassite-Post-Cassite: 9.5 ha (RJ9-2:1, RJ9-3:1, RB9-2:3, RB9-3:1, RB9-4:1, B9-1:4, B9-3:1,09-1:1); Neo-Babylonian+?: trace (RB10-5).

101 - 280 N X 155 X 2. Late Larsa: 2.5 ha(RJ8-1:5, RB8-2:1, RB8-5:3, B8-1:1?, 08-1:1),Neo-Babylonian+: trace (RJ10-1:1 without ridge,010-3:1).

102 325 N X 300 X 2. Neo-Babylonian+: 8.5ha (RJ10-1:5 [3 without ridge], RJ10-2:1, RB10-2:1, RB10-5:1, B10-4:1, 010-3:3, 010-4:1).

103 330 E X 280 X 2. Brick fragments; ceramicslag; mud-brick room block footings visible onsummit. Late Larsa? (RJ7-5:3, B7-2:4, B7-3:1, B7-5:2, RJ8-3:5); Cassite: 6.4 ha (RJ9-1:1, RB9-1:1,RB9-2:2, RB9-3:5, B9-1:3, 09-1:1); Neo-Baby-lonian+?: trace (RJ10-2:3).

104 235 NW X 165 X 3. Ceramic slag; possi-ble stone footings; concentrations of freshwaterbivalves. Eridu, Hajji Muhammad, Early Ubaid(RB2-1:2, RB2-4:1); Late and Terminal Ubaid: 3.1ha (RJ3-2:3, RB3-6:5, RB3-8:3, RB3-9:1, RB3-10:2, RB3-11:4, RB3-13:1, RB3-15:1, 03-1:6, 03-2:8, 03-3:3); Uruk?: trace (crosshatch decoratedjar shoulders: 2) (Probably Safar's 'Usaila; seeSafar 1950, p. 28.)

105 180 N X 150 X 2. Large basalt pieces onsummit. Terminal Ubaid (interior scraped sherd);Late Larsa: 2.2 ha (RJ7-5:1, B7-2:1, B7-6:1, RJ8-1:2, RJ8-2:1, RB8-4:1, B8-8:3, 08-1:3); Cassite:2.2 ha (RB9-1:2, RB9-2:1, B9-1:2, B9-2:1, RB10-4:2 thick).

106 280 NE X 160 X 2. Bricks. Late Larsa: 2.4ha (RJ8-1:10, RJ8-3:3, RB8-2:3, B8-1:1, 08-1:1,08-2:1); Post-Cassite?: trace (RB9-1:2 [1 withhatched ridge], RB9-4:1).

107 - 110 WNW X 95 X 2. Post-Cassite: 0.8 ha(RJ7-5:1, B7-1:2, B7-2:4, B7-4:4, B8-3:3, B8-5:1,RJ9-1:1, RJ9-2:4, RJ9-3:1, RB9-1:1, RB9-2:5,RB9-4:1, B9-2:1, B9-3:1, RJ10-2:1).

108 - Eridu Jinub. 600 N X 500 X 5. Bricks,33 X 32 X 7 cm, 32 X 31 X 7 cm, 25.5 X25.5 X 8 cm; ceramic slag. Ur III-Early Larsa?(RJ7-3:!, B7-1:1, B7-4:1); Late Larsa: ca. 10 ha(RJ8-1:10, RB8-3:1, RB8-4:2, B8-1:1); Neo-Baby-lonian+: 22 ha (RJ10-1:7 [3 without ridge], RJ10-

342

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

2:1, RB10-4:1 thick, B10-1:1, B10-2:1, 010-8:1,010-9:).

109 Ishan Khinaitla. 250 N X 230 X 0.5. Late Larsa:3.8 ha (RB6-:1, RJ7-3:1, RJ7-5:8, RJ8-1:8, RB8-2:2, RB8-4:2 large, B8-2:2, 08-1:2); Neo-Baby-lonian+: 3.8 ha (RJ10-1:8 [5 without ridge], RJ10-2:4, RJ10-3:1, RB10-2:3, RB10-3:3, B10-1:4, B10-4:1, 010-3:3).

110 135 N X 120 X 0.5. Neo-Babylonian+:1.3 ha (06-1:1, RJ8-1:2, RJ10-1:14, [4 withoutridge], RJ10-2:1, RJ10-3:1?, RB10-1:1, B10-1:5,010-3:1, 010-5:1).

111 230 E X 110 X 0. Brick fragments. Ter-minal Ubaid+ (03-1:1, interior combed sherd);Ur III-Early Larsa: 1.9 ha (RJ7-1:1, RJ7-2:1, RJ7-3:2, RJ7-4:2, RJ7-5:1, B7-5:8, B7-6:8, 07-1:1);Cassite: trace (RB9-1:1, RJ9-3:1, 09-1:1).

112 -- 100 WNW X 50 X 1. Neo-Babylonian+:0.4 ha (B7-1:1, RJ8-1:1, RJ10-1:9 [5 withoutband], B10-2:2, 010-8:1, 010-9:1).

113 110 NW X 80 X 0.5. Neo-Babylonian+:0.7 ha (B7-1:1, RJ8-1:1, RJ10-1:6 [6 withoutridge], RJ10-2:2, BJ10-2:1, BJ10-3:1, 010-7:1,010-9:1).

114 - 70 NW X 50 X 1. Late Larsa: 0.3 ha (RJ7-5:2, RJ8-1:2, RB8-4:2, B8-1:2, 08-1:2); Neo-Babylonian+?: trace (RB9-5:2 thin, B10-1:1).

115 - 250 NW X 110 X 1. Ceramic slag. LateLarsa? (RJ7-5:1, RJ8-1:1, RJ8-3:2, B8-3:1, 08-1:1); Neo-Babylonian+: 1.9 ha (RJ10-1:9 [2 with-out band], RJ10-2:1, RB10-2:3 [2 glazed], RB10-3:1 glazed, RB10-4:1 glazed, B10-2:1, B10-4:5,010-4:1).

116 150 NW X 95 X 0.5. Very few sherds.Neo-Babylonian?: 1.3 ha (RB10-4:2).

117 180 E X 150 X 1. Neo-Babylonian+: 2.5ha (RB6-1:4 glazed, B7-1:1, RB8-1:7 small, RB8-5:1, RB9-5:1, RJ10-1:11 [3 without ridge], RJ10-2:10, RJ10-5:1, RB10-2:2 glazed, RB9-5:1, B10-1:2, B10-2:2, 010-3:1, 010-4:2, 010-9:4).

118 165 X 80 X 1. Brick, 33 X 33 X 6.5 cm.Late Larsa: 1 ha (RJ8-1:3, RJ8-2:3, RB8-3:1, RB8-5:1, B8-1:1, B8-3:1, 08-1:1); Cassite-Post-Cas-site?: trace (RB9-5:1 thin); Neo-Babylonian+:trace (RJ10-1:3, B10-2:2).

119 - 100 NW X 80 X 0.5. Neo-Babylonian+:0.7 ha (B7-1:3:1, RJ10-1:2, RJ10-2:1, B10-4:1,010-2:1, 010-4:1, 010-5:1).

120 - 145 NW X 115 X 1.5. Ceramic slag. LateLarsa: 1.3 ha (RJ8-1:3, RJ8-3:3, RB8-1:1, RB8-3:1, RB8-4:2, B8-1:3); Cassite-Post-Cassite?: trace(RB9-2); Neo-Babylonian+ (B10-2:1, B10-3:1).

121 - 60 X 60 X 0.5. Few sherds, none diagnos-tic.

122 east 135 NW X 100 X 1.5. Akkadian: 1.2ha (RJ6-2, RJ6-3 without ribbing, ribbed sherd);Ur III-Early Larsa: 1.2 ha (RJ7-1:4, RJ7-2:11,RJ7-3:2, B7-2:2, B7-4:5, B7-5:1); Cassite-Post-Cassite: trace (B9-3:2).

122 west Tell Daima. 470 NW X 225 X 3. Ur III-Early Larsa: 6.5 ha (RJ7-1:4, RJ7-3:3, RJ7-5:1,B7-1:1, B7-2:1, B7-5:12, B7-6:1, 07-1:3); LateLarsa?: trace (RJ8-1:3, RJ8-3:2); Post-Cassite:trace (RJ9-3:1, B9-1:1).

123-25 Outside survey area.126 140 N X 115 X 1. Many bricks and brick

fragments, perhaps excavated by Woolley as a"suburb" of Ur. Few sherds.

127 - 220 W X 140 X 0.5. Early Ubaid (RB2-4);Late-Terminal Ubaid: 1.5 ha (RJ3-2:1, RB3-7:1,RB3-8:1, 03-1:5); Jemdet Nasr-Early Dynastic:2.4 ha (RJ4-1:1, 04-1:1, 04-2:12, 04-4:1, RJ5-1:1, RJ5-2:7, RJ5-3:4, RJ5-4:1, RB5-1:7, RB5-2:3,RB5-3:1, RB5-6:2, B5-3:6, B5-4:6, 05-2:4, 05-3:2,05-4:2, 05-7:1); Ur III-Early Larsa?: trace (RJ7-4:3, RB8-1:1, B7-5:1, 06-1:1).

128 180 NW X 170 X 1. Bricks, 35 X 33 X7 cm; ceramic slag. Late Larsa? (RJ7-5:5, B7-1:1,B7-2:5, RB8-1:2); Cassite: 1.8 ha (RJ9-1:3, RB9-1:3, RB9-2:4, RB9-3:3, B9-1:2, 09-1:1).

129 65 X 65 X 0.5. Few sherds. Late Larsa: 0.3ha.

130 60 X 60 X 0.5. Looted cemetery. Cassite-Post-Cassite (RJ7-5:1, B7-1:1, B7-2:1, B7-3:1, RJ8-1:1, B8-1:1, RJ9-1:1, RJ9-3:1, RB9-1:2, RB9-2:1,RB9-3:1, RB9-4:1, RB9-5:1, B9-3:3, RJ10-2:1).

131 150 N X 110 X 2. Bricks, 31 X 31 X5.5 cm, 29 X 29 X 6 cm. Site on gravel ridge ofHazim. Cassite-Post-Cassite (RJ7-5:7, B7-1:1,RB8-1:1, RB8-3:2, RJ9-2:1, RB9-1:4, RB9-2:3,RB9-4:4, B9-1:2, B9-2:2, RJ10-2:4).

132 75 X 75 X 1. Also on Hazim. Few sherds.133 260 N X 190 X 5. Also on Hazim. Ceramic

slag; bricks, 30 X 29 X 6 cm, 30 X 30 X 8.5 cm.Early Dynastic? (RB5-2:3); Ur III-Early Larsa: 1ha (RJ7-5:5, RJ7-3:1, B7-1:2, B7-3:1, 07-1:1);Late Larsa: 3 ha (RJ8-1:13, RJ8-2:3, RJ8-3:1,RB8-1:2, RB8-3:6, RB8-4:4, B8-1:1?, 08-1:8, 08-2:3); Post-Cassite: trace (RJ9-2:1, RB9-1:2); Neo-Babylonian+: trace (R10-1:4 without ridge,RJ10-4:1, B10-3:1).

134 - 110 E X 65 X 0.5. Many bricks, 26 X26 X 6.5 cm. 31 X 26 X 7 cm, 32 X 30 X 8 cm,35 X 34 X 7 cm. Planoconvex brick wall footings,25 X 16 X 7.5 cm. On bank of channel throughHazim. Late Ubaid: (RB3-4:1, O3-1:1, O3-2:1);Jemdet Nasr-Early Dynastic: 0.3 ha (RJ4-2:1,RJ5-2:1, RJ5-4:2, RJ5-5:1, RB5-2:2, B5-2:1, 05-

343

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

3:1); Akkadian? (RJ6-2?:1 ribbed sherd); Ur III-Early Larsa: 0.3 ha (RJ1-1:2 with very high col-lars, B7-1:1, B7-3:1, B7-6:1, B8-1:1, B9-3:1).

135 south 100 NNE X 70 X 0.5. Brick frag-ments. Late Larsa: 0.5 ha (RJ7-3:1, RJ7-5:1, B7-1:6, RJ8-1:9, RJ8-3:1, RB8-4:5 [4 very thick],B8-1:2, B8-2:3, 08-1:9, 08-2:2); Neo-Babylo-nian+: trace (RB9-3:1, RJ10-2:1, RJ10-4:1, B10-2:1).

135 north 100 NNE X 70 X 0.5. Remains ofrectangular oven with six cylindrical side cham-bers. Neo-Babylonian+: 0.5 ha (RB6-1:3, RB8-5:1, B8-2:1, RJ10-1:5, RJ10-2:2, RB10-2:1, B10-1:4, 010-3:1, 010-4:3).

136 180 X 180 X 1. Bricks, 30 X 29 X 6 cm,26.5 X 26 X 5.5 cm, 16 X 9.5 X 6 cm. Cassite:2.2 ha (RJ7-5:2, B7-1:2, B7-2:4, RJ9-1:2, RB9-2:5, RB9-5:1, B9-1:3, 09-1:1).

137 - 115 N X 90 X 1. Late Larsa (RJ8-1:2,RB8-4:1, 08-1:2).

138 160 NE X 125 X 1.5. Bricks, 30 X 30 X5.5 cm; ceramic slag in heaps to southeast. Cassite:1.4 ha (RJ7-5:1, B7-1:1, B7-3:2, RJ9-1:1, RB9-2:4, B9-1:1, B9-2:1).

139 - 100 NNW X 90 X 0.5. Post-Cassite?: 0.7ha (RB8-1:1, RB8-3:1, RB9-2:4, RB9-4:1, RB9-5:1).

140 150 E X 140 X 1.5. Late Larsa: 1.5 ha(RB6-1:3, RJ7-5:1, B7-1:1, RJ8-1:5, B8-1:2, B8-2:2); Neo-Babylonian +: 1.5 ha (RJ10-1:6 withoutridge, B10-2:1, 010-3:1, 010-4:1, 010-5:1, 010-9:2).

141 Tell Rifa'i al-'Ubaid. 205 NE X 110 X 1.5. EarlyUbaid (RJ2-1:1, RB2-2:2, RB2-4:3); Late Ubaid:1.5 ha (RJ3-1:2, RJ3-2:6, RJ3-3:1, RB3-2:3, RB3-3:3, RB3-5:4, RB3-6:2, RB3-7:1, RB3-8:6, RB3-9:2, RB3-11:2, 03-1:10, 03-2:2, 03-3:3, interior-combed sherds: 2); Uruk: 1 ha (RJ4-1:1, RJ4-2:3,RB4-1:7, 04-1:1, 04-2:3, 04-6:2, 04-7:1 of gray-ware, crosshatched shoulder decoration).

142 190 E X 65 X 0. Ubaid cultivation (03-1:2); Jemdet Nasr-Early Dynastic: 1:5 ha (RJ5-1:3, RJ5-2:2, RJ5-4:3, RB5-1:1, RB5-2:1, RB5-6:3,B5-1:1, B5-2:2, B5-3:3, B5-4:5, 05-2:2, 05-3:1,05-7:3).

143 - 100 NW X 60 X 1. Bricks. Late Larsa: 0.4ha (RB5-2:1, B7-1:1, RJ8-1:3, RJ8-3:2, RB8-2:1,RB8-4:2, B8-1:3); Neo-Babylonian+?: trace(RJ10-1:1).

144 - 105 NX 90 X 0.5. Bricks, 29 X 21 X 7cm. Late Larsa: 0.9 ha (RB5-2:2, RB6-1:2, B7-1:2,RJ8-1:3 [1 small], RB8-2:1, B8-1:2, 08-1:3, RJ10-1:1 without ridge).

145 - 220 N X 90 X 1. Jemdet Nasr-Early Dy-nastic? (B5-4:1, O4-1:1); Ur III-Early Larsa: 1.2

ha (RJ7-2:1, RJ7-3:3, B7-1:1, B7-3:2 large, B7-5:3,RJ8-1:1, RJ8-3:4).

146 - Large holes in sandy soil, perhaps recentborrow pits; Late Larsa sherd scatter.

147 350 NNW X 210 X 2. Brick fragments;ceramic slag. Cassite: 3.7 ha (RJ7-5:7, B7-1:3, B7-2:3, RJ8-3:1, RJ9-1:1, RB9-1:3, RB9-2:6, RB9-3:7,B9-1:2, 09-1:1).

148 Tell al-Dahaila. 550 WNW X 320 X 7. Bricks,37 X 37 X 6.5 cm, 33 X 32.5 X 7 cm; ceramicslag concentration on west extremity. Ubaid culti-vation? (03-1:1); Late Larsa? (B7-3:1, RJ8-1:1,RB8-5:1); Neo-Babylonian +: 15 ha (RB6-1:1,RJ10-1:1 [2 without ridge, 1 small], RJ10-2:4,RJ10-4:1, RJ10-5:1, RB10-1:1, RB10-4:1, RB10-5:2, B10-1:5, B10-2:4, 010-3:2, 010-4:2).

149 Tell al-Skena. 95 NW X 30 X 1.5. Cassite: 0.4 ha(B7-1:3, B7-2:1, RJ9-1:1, RB9-3:1, RB9-5:1).

150-55--Outside survey area.156 - 200 NNW X 70 X 0. Pieces of carved lime-

stone and several limestone door sockets. Ubaidcultivation? (03-1:2); Jemdet Nasr-Early Dynas-tic: 0.6 ha (04-1:1, RJS-1:7, RJ5-2:2, RJ5-4:1,RB5-3:1, RB5-6:9, B5-2:4, B5-3:7, 05-2:2, 05-3:1,05-4:3, 05-5:1, 05-7:10, semicircular limestoneslab with circular perforations) (see Woolley 1956,pl. 33).

157 225 N X 150 X 3. Bricks, 25 X 17.5 X 7.5cm. Ur III-Early Larsa: 2.2 ha (RJ7-1:1, RJ7-2:1,RJ7-3:1, RJ7-4:1, B7-1:3); Late Larsa (RJ8-1:14,RB8-3:5, RB8-4:1, RB8-5:2, B8-1:10, 08-1:4, 08-4:4, RB9-1:1); Neo-Babylonian +?: trace (RB10-1:1).

158 425 NE X 290 X 5. Bricks, 39 X 39 X 7cm, 36 X 35.5 X 7 cm, 32 X 32 X 6 cm, 22.5 X22 X 6 cm; ceramic slag to east and south; kilnsor ovens to north. Late Larsa: 9.3 ha (B7-1:1, RJ8-1:3, RB8-4:1, RB8-5:2, 08-2:1); Cassite: trace (B9-1:1); Neo-Babylonian +: 9.3 ha many graves(RJ10-1:12 [3 without ridges], RJ10-2:4, B10-1:2,B10-4:2, 010-2:2, 010-3:1, 010-4:2).

159 - 260 NNW X 175 X 2. Much slag; manygraves; few sherds. Jemdet Nasr-Early Dynastic?(RJ5-1:1); Ur III-Early Larsa? (RB6-1:1, RJ7-2:1?,RJ7-5:1, B7-l:2); Late Larsa: 3.2 ha (RJ8-1:4,RJ8-3:2, B8-1:4, 08-1:2); Neo-Babylonian +:trace (RB9-2:2, B9-3:2, RJ10-1:2 without ridge,B10-2:1).

160 - 150 NW X 140 X 1.5. Ceramic slag. LateLarsa: 1.3 ha (B7-1:4, RJ8-1:3, RB8-1:1, RB8-3:2,RB8-5:3, B8-1:2); Post-Cassite: 1.3 ha (RJ9-1:1,RJ9-3:1, RB9-2:3, B9-2:1).

161 - 250 W X 200 X 2.5. Much damage bywind erosion. Bricks, 34.5 X 34.5 X 5 cm; brickand mud-brick footings visible. Ubaid cutlivation?

344

oi.uchicago.edu

Appendix: The Southern Margins of Sumer

(03-1:1); Ur III-Early Larsa: 2.5 ha (RJ7-1:4, RJ7-3:5, RJ7-4:1, RJ7-5:2, B7-1:2, B7-3:9, B7-5:5, 07-1:3); Late Larsa?: trace (08-1:3, RJ8-3:2).

162 - 125 WNW X 90 X 1. Clear small canalsand plow marks to west of site. Ubaid cultivation?(03-1:1); Neo-Babylonian +: 0.9 ha (RB6-1:1,RB8-2:1, RJ10-1:9, RJ10-2:6, RJ10-3:1, RJ10-4:1,RB10-3:1, RB10-4:2, RB10-5:1, B10-1:1, B10-4:2,010-2:1, 010-3:1, 010-4:2).

163 See map, figure 26. 155 N X 150 X 1.Bricks, 33 X 33 X 5 cm, 31.5 X 31.5 X 5 cm.Traces of rectangular wall, 130 m square, withbastions. Jemdet Nasr-Early Dynastic?: trace(RB5-2:1); Late Larsa: 1.8 ha (RJ8-1:8, RJ8-2:1,RJ8-3:1, RB8-3:2, RB8-4:2, RB8-5:1 thick, B8-1:4,08-1:4); Neo-Babylonian+: 1.8 ha (B9-3:1, RJ10-1:4, RJ10-3:1, B10-2:4, 010-3).

164 - 250 WNW X 80 X 3. Ceramic slag. Cas-site-Post-Cassite: 1.4 ha (B7-1:4, B7-2:4, RJ7-5:3,RB8-1:1, 08-1:1, RJ9-2:1, RB9-2:2, RB9-4:1,RB9-5:3, B9-1:2).

165 - 180 NW X 140 X 1. Ubaid cultivation?(03-1:1); Late Larsa? (RB6-1:1, RJ7-5:2, RB8-1:1,08-1:1); Neo-Babylonian +: 1.9 ha (RJ10-1:6,RJ10-2:4, RJ10-3:1, RJ10-4:1, 010-4:3, 010-5:1).

166 - 35 X 35 X 1.5. Traces of rectangular wall35 m square. Late Larsa: 0.2 ha (B7-3:1, RJ8-1:5,RJ8-3:1, RB8-1:1, RB8-2:1, RB8-4:1, RB8-5:2, B8-1:4, 08-2:1).

167 Small slag concentration ca. 2 m high. Neo-Babylonian +.

168 Number not assigned.169 200 X 200 X 1. Many looted graves. Late

Larsa: 1.9 ha (RJ8-1:2, RB8-4:1, B8-2:1, 08-1:1);Neo-Babylonian+: 1.9 ha (RJ10-1:7 withoutridges, B10-4:2, 010-3:1).

170 210 X 210 X 1. Post-Cassite: 3.6 ha (B7-1:2, B7-2:2, RJ8-3:1, RJ9-3:1, RB9-1:2, RB9-2:1).

171 - 350 NE X 240 X 1. Uruk: 4.2 ha (RJ4-1:11, RJ4-2:3, RB4-1:3, 04-1:1, 04-2:4, 04-3:3,04-5:2, 04-8:1); Late Larsa: trace (RB8-2:1); Cas-site-Post Cassite: trace (RB9-3:1).

172 Tell al-Lahm. 390 X 310 X 13. Air photographshows town wall. Early Dynastic: ca. 5 ha; Ak-kadian: ca. 5 ha; Ur III-Early Larsa: ca 8 ha; LateLarsa: ca. 8 ha; Cassite-Post-Cassite: ca 11 ha (seeSafar 1950).

173 Tell al-Lahm Shamal. 430 N X 410 X 2. Post-Cassite, Neo-Babylonian +: 14 ha (see Safar1950).

174 - 55 N X 45 X 0. Looted graves. Ur III-Larsa? (RJ7-5:3, B7-1:1, B7-5:1, RJ8-3:1); Cassite-Post-Cassite (RJ9-1:1, RJ9-2:1, RJ9-3:3, RB9-1:1,RB9-2:2, B9-1:1).

175 Tell al-Awaija or Maftul Shaykh Ajil. 250 N X150 X 2.5. Jemdet Nasr-Early Dynastic: 2.7 ha(RJ5-2:2, RB5-2:4, B5-1:3, B5-2:1, B5-4:5, 05-2:2,05-4:3, 05-4:1, cylindrical cup bases: 2).

176 - 135 NE X 115 X 0.5. Much salt damage.Jemdet Nasr-Early Dynastic: 1.1 ha (RJ5-2:1,RB5-1?:1, RB5-2:7, B5-1:1, B5-2:1, B5-3:2, B5-4:3, 05-2:1).

177-79- Outside survey area.180 90 X 60 X 0.5. Late Larsa: 0.6 ha (RJ7-

3:1, B7-1:2, RJ8-1:2, 08-1:2).181 Outside survey area.182 West - 250 E X 210 X 2. Ur III-Early Larsa?

(RJ7-5:2, B7-1:4, B7-2:2, B7-3:2); Late Larsa: 3.0ha (RJ8-1:3, RB8-3:3, 08-1:1, 08-2:2); Cassite:3.0 ha (RJ9-1:3, RB9-1:1, RB9-3:2, RB9-5:1, B9-1:3 small).

182 East 330 E X 150 X 2. Neo-Babylonian +:3.7 ha (RJ8-1:2 small, RJ10-1:21 [14 withoutband], RJ10-2:2, RB10-1:2, RB10-2:7, RB10-3:2,RB10-4:3, B10-1:2, 010-4:3).

183 Merejib. 430 NW X 250 X 3. Early Larsa (RJ7-3:1, RJ7-5:5, B7-1:1, B7-2:3, B7-5:1); Late Larsa?(RJ8-1:3, RB8-1:1); Cassite: 7.2 ha (RJ9-3:1, RB9-1:6, RB9-2:2, RB9-4:1, RB9-5:1, B9-1:2, B10-4:1).(Not Woolley's Merejib; see EP 29; see also Camp-bell-Thompson 1919.)

184 95 X 95 X 0.5. Few sherds. Neo-Babylo-nian +.

185 80 X 80 X 0.5. Ubaid cultivation? (03-1:1); Ur III-Early Larsa: 0.5 ha (RJ7-3:3, RJ7-5:2,B7-1:1, B7-5:3, B7-6:2); Late Larsa? (RJ8-1:2,RB8-1:2, RB8-3:1); Cassite: 0.5 ha (RB9-1:1, RB9-2:1, B9-1:1).

186 45 X 45 X 0.3. Cassite: 0.2 ha (RJ7-5:2,B7-1:5, B8-1:1, RB9-1:3, RB9-2:3, RB9-5:1, B9-1:2).

187 Tell Tuwayil. 180 WNW X 110 X 3. Ubaid culti-vation? (03-1:1); Late Larsa? (B7-1:1, B7-2:3,RJ8-1:4, RB8-1:1); Cassite: 1.5 ha (RB9-2:2, RB9-3:2, B9-1:4).

188 - 150 X 150 X 1. Cassite-Post-Cassite: 1.9ha (RJ7-5:2, B7-1:4, B7-2:4, B7-3:1, RJ8-1:1, RJ9-2:1, RB9-1:9, RB9-2:3, RB9-3:2, RB9-5:3, B9-1:4,09-1:2).

189 - 180 X 180 X 0.5. Ur III-Larsa? (RJ7-5:5,B7-1:1, B7-4:1, RB8-3:1, RB8-5:1); Cassite-Post-Cassite: 2.1 ha (RJ9-2:3, RJ9-3:2, RB9-2:1, RB9-4:4, B9-1:l, B9-3:1, RJ10-2:1).

190 - 170 E X 90 X 3. Late Larsa? (RJ7-5:6, B7-1:1, B7-2:1, RJ8-1:2, RB8-1:2, RB8-3:1); Cassite-Post-Cassite: 1.7 ha (RJ9-1:1, RB9-1:1, RB9-2:4,RB9-4:2).

345

oi.uchicago.edu

oi.uchicago.edu

Notes

CHAPTER 11. The estimate that losses amount to about half of con-

sumption requirements, or about one-third of gross diversion,may be much too low. This is a widely repeated rule of thumbin Middle Eastern agricultural planning studies, but suchstudies are not noted for the care with which they measurefeatures of the traditional agricultural regime they are seekingto replace. A recent, much more systematic and anthropo-logically oriented study of the utilization of water under simi-lar conditions in southwestern Iran comes to the radically dif-ferent estimate that there are "very high (up to 60%) lossesfrom primitive canals during conveyance to the fields" (Kirkby1977, p. 271). If something approaching three-fifths instead ofone-third of gross diversions was normally lost, the calcula-tions on maximum* extent of irrigable area that are offeredbelow may need to be reduced by almost half.

2. These general remarks fail to take into account certainminor cultivation practices that did not coincide with the pri-mary agricultural cycle. Intensive summer cultivation of fruitsand vegetables on very restricted plots was obviously one cate-gory with a different schedule. But in addition Landsbergerhas called attention to a fast-growing two-row barley (Horde-um nigrum rectum) that could be planted as late as mid-March, when water supplies were ample, and be ready forharvest by early June. In the abstract, a cereal with these quali-ties seems to have been of great potential importance as ameans of enlarging and assuring agricultural output. Inexplic-ably, however, that was not the case: "diese Art von Sommer-saat kann, da in Formularen und Wirtschaftsurkunden nichtzu belegen, stets nur untergeordnete Bedeutung gehabt haben.Schliesslich scheint sie (seit der Kassitenzeit?) ungebrauchlichgeworden und in Vergessenheit geraten zu sein" (Landsberger1949, pp. 283-84).

3. Two factors generally permit irrigation deposits to bedistinguished from natural sedimentation: (1) uniform canalgradients lead to constant, relatively high velocity of flow, sothat bed deposits are more uniformly made up of coarser soilparticles than in the case of natural streams; and (2) with an

abrupt inflow into relatively small irrigation basins, sedimenta-tion in the fields is relatively more undifferentiated (Schilstra1962, p. 188).

4. M. G. Ionides (1954, p. 394) makes the same point inarguing that the Lees and Falcon estimate of 1,500 square milesshould be corrected to "more like ten times that on which theauthors have based their calculations," or about 38,400 squarekilometers. The further increase in the estimate made herestems largely from archaeological surveys in intervening years,attesting to periodic shifts in the Tigris and Euphrates coursesand the branching character of the ancient Euphrates in par-ticular. This allows for even greater dispersal of the silt loadthan lonides could have been aware of.

5. This statement finds some confirmation at Tell Abu Sala-bikh, within the intensively surveyed area. The excavators' re-ports and sections indicate that virgin soil underlying an EarlyDynastic occupation of the eastern part of the mound wasencountered at depths of between 1 and 2 meters below thelevel of the surrounding plain (Postgate and Moorey 1976,pp. 135, 141, fig. 5).

6. As Butzer summarizes present evidence for these changes,there were "higher world sea levels of +2 m or more ca.1200 B.C., as well as lower sea levels of -2 m or so ca. 2200 B.C.and again 300 B.C." (1976, p. 36; cf. Niitzel 1975).

7. Roux 1960. To judge from the appearance of the surfacepottery, Abu Salabikh goes back certainly to Neo-Babyloniantimes and Tell Aqram may date in part to the Cassite period.Nothing in the illustrations appears older.

8. Working with air photographs that I have been unableto examine, Buringh (1960, fig. 65) has traced well-definedmeander patterns at the upper end of the Shatt al-Hilla thatsurely must be those of a major ancient branch of the Eu-phrates.

9. The original map appears to have been based on compassbearings and on distances estimated from travel times. Thelatter are fairly irregular and are especially foreshortened to-ward the east. In revising it, stream courses and ancient siteand recent settlement locations were corrected where possible

347

oi.uchicago.edu

Notes to Pages 26-69

from the air photographs, and other contemporary settlementsand land-use boundaries were then entered by extrapolation.

10. No attempt has been made in the revised map to pre-serve Andrae's distinction between qal'as and muftuls thatwere in ruins and those that were still in use. This does notseem informative as to the contemporaneous distribution ofsettlements, since the map was made during the later stagesof a process of general abandonment. It also may be worthnothing that recent pottery was observed at a substantial num-ber of sites where Andrae fails to record a present or formersettlement (e.g., sites 003, 007, 010, 011, 026, 027, 028,029, 042, 043,047, 050, 052, 1357, 1358, 1396, 1407, 1452, and1457), even though no systematic attempt was made to recordall recent sites. This seems to argue that many of them hadbeen abandoned much earlier, so that they easily escaped hisnotice and were no longer known to his informants. If so, the"Recent" settlement pattern may be one of considerable an-tiquity.

11. The original map from which figure 5 has been adaptedwas drawn up by Major Ali Bey, of the staff of the ImperialWar College, who had accompanied General Ragib Aga toBaghdad on the latter's special assignment from the sultan.It was reproduced at the Imperial Army Engineering School inA.H. 1264 (A.D. 1847-48). I am much indebted to my colleagueProfessor Richard L. Chambers for translating the legend andrelevant geographical terms.

CHAPTER 2

1. Several instances are discussed in later chapters in whichthis double counting is particularly likely to have occurred,including the rapid growth of Uruk toward the end of theEarly Dynastic I period and the succession of the Ur III, Isin,and Larsa dynasties (see pp. 84, 143).

2. It is partly these considerations that have led to groupingtogether the settlement data for the Ur III and Isin-Larsaperiods and for the Neo-Babylonian and Achaemenian periodsin the chapters that follow, since their combined lengths morenearly meet the prevailing standard for other periods. But inboth instances it would have advanced our understanding ofeconomic and demographic processes if the separation betweenthem could have been maintained, and the decisive reason fornot doing so arose from the lack of clear 'criteria for separat-ing the surface collections. Were a principle of standardizationto be introduced into this study, its primary effect wouldprobably be a still further reduction in the already very lowpopulation levels that are attributed to the Middle Babylonianperiod.

CHAPTER 3

1. It will be noted that a form of nearest-neighbor analysiswas employed in the original study of the Warka region (Adamsand Nissen 1972, p. 27). The attempt is not being repeated forthe reasons given.

2. The original date of 2815 ±- 85 B.c. was published with-out reference to the half-life on which it was based (Lenzen1965, p. 20). The corrected date given here assumes that theHeidelberg laboratory was still employing the 5,570 half-life.If the 5,730 half-life that now has been generally adopted was

already being used, the figure of 3610 B.C. should be reducedto 3545-3295 B.C. The bristlecone dendrochronological correc-tion is derived from Ralph, Michael, and Han 1973.

3. Postgate has recently summarized the evidence for thelocation of ancient Larak, perhaps somewhat audaciouslydrawing together slender textual leads as to where it waslocated and how it might be recognized. He concludes that itshould be placed "on the eastern side of Sumer, to the northof Umma," and that the written evidence "rather indicates analmost total abandonment of the site from the Early Dynasticperiod (at the latest) until the 1st millennium. . . . We shouldlook either for a large mound with a major occupation nolater than ED III, overlaid by Neo-Babylonian remains, or elsefor two mounds of the two separate periods in the same vicin-ity" (1976, p. 82).

It should be noted that this extraordinary site exactlymatches the prescription he has provided entirely indepen-dently. To be sure, the Neo-Babylonian resettlement at site1306 was on a fairly limited scale. If Larak at that time wasa place of some importance, as mention in the Murashu archiveperhaps implies, one might speculate that little more than ashrine was erected on the ancient ruins while the name wastransferred to nearby site 1439, a major, newly founded Neo-Babylonian town.

Postgate also notes that the first millennium references speakof Larak as having been "on the bank of the old Tigris." Isthis possibly a reference to the very large watercourse dealtwith earlier that was already abandoned in the fourth millen-nium? If this is Larak-and I do not wish to minimize themany uncertain links in the chain of reasoning leading to theidentification-the site's attainment of greatest size as earlyas the Uruk period is of striking importance. It would seem toimply historicity for even the "antediluvian" portions of theSumerian Kinglist and thus to bring the Uruk period itselftantalizingly close to the threshold of "history."

4. There are only two exceptions to this pattern, the pairingof 790 with 792 and of 1165 with 1166. It may be noted in table7 of Appendix A to this chapter, however, that the surfacecollection from 790 was significantly later than the one from792, and that 1165 similarly may well be later than 1166. Inboth cases, therefore, we may be dealing with a single re-located town rather than adjoining, coexisting ones.

5. Leaving aside chronological and locus-of-discard uncer-tainties, the frequency of recovery also seems to differ sugges-tively from district to district. Of particular interest is the con-trastive distribution of clay sickle fragments and chert or flintblade segments that also may have been used in sickles, sincethis may reflect differing degrees of access to flint as a rawmaterial. Henry Wright (pers. comm.) analyzes this contrastas follows:

"An indication of differing technological contexts withinMiddle-to-Late Uruk Mesopotamia can be gained by compar-ing the densities of clay sickle fragments and chert bladesegments in ten-square-meter diagnostic surface collectionunits. The following figures suggest that clay sickles weremore common and flint blades less common on the centralEuphrates floodplain (Nippur-Warka subregion) than on thesouthern margins of Sumer (Ur-Eridu subregion) during thisspan of time:

348

oi.uchicago.edu

Notes to Pages 69-144

Nippur-Warka Ur-EriduMean No./ Standard Mean No./ Standard

10 m- Deviation 10 m2 Deviation

Clay sickle density 2.72 2.99 1.85 .64Flint blade density .26 .25 1.85 1.69Number

of collections 27 3

"Striking as this apparent difference is, however, it must beregarded as still doubtful. The large standard deviations areindicative of markedly skewed distributions. Testing with theKruskal-Wallace nonparametric one-way anlysis of variance,a test appropriate to such nonnormal statistical distributions,shows that the probability of the observed difference in claysickle density being due to chance is .91 while that for flintblades is .27. With more samples from the Ur-Eridu subregiona significant difference could probably be demonstrated. Butonly when small-scale excavations have been undertaken ina sample of Uruk sites in the two subregions will it be possibleto eliminate the possibility that such differences are a resultof different erosion processes in various parts of the alluvium."

6. Adams and Nissen (1972, p. 29). Note that the dimen-sions ordinarily recorded in the Warka survey as well as thisone specify, in effect, the minimum size of the rectangles en-closing sites rather than the area of the ordinarily circular orovoid sites themselves. Hence areas as given in hectaresthroughout this study need to be reduced by roughly 20 per-cent to indicate hectares of actual settlement, although indi-vidual cases will vary considerably according to the regularityof their site perimeters. Within existing time constraints it wasnot possible to further refine site descriptions and areas on thebasis of more numerous measured points along their perime-ters, advantageous as this would have been on other grounds.In any case, refinement in this respect alone would do little toreduce the prevailing crudity of population estimates likethese. Unavoidable as it is at present, the use of a uniformdensity standard like 125 persons per hectare introduces farlarger uncertainties than those deriving from imprecisely mea-sured areas. Subsumed within any average figure there isvery likely to be considerable variation attributable to settle-ment size, function, duration, politicomilitary circumstances,and other factors that could be specified only on the basis ofsubstantial archaeological and/or textual evidence.

It is an underlying premise of this study-and indeed, ofany synthetic reconstruction-that the existence of variabilitydoes not negate the careful use of averages in arriving at gen-eralizations. On this basis surface data, even in the absence of(hitherto unrealizable) large archaeological exposures or ex-tensive, systematic sampling, or both, does seem to justifytentative statements about aggregate regional population levelsto which many different kinds of settlements would contribute.But estimates of the population of individual sites on the basisof surface data alone, through the use of a single density con-stant, must be regarded as little more than speculative. Cf.chap. 4, n. 1.

7. It is interesting that the hierarchy of urban size, and pre-sumably political and economic dominance to some degree,bears no apparent relation to the conceptual categories of thetime. Repeatedly copied geographical lists of Jemdet Nasr date

follow the order: Ur, Nippur, Larsa, Uruk, Kesh, Zabalam....One would certainly think it was "reasonable to assume that ageographical list would be headed by the most important Su-merian cities" (Green 1977, p. 294). But the first three entrieswere unquestionably not comparable to Uruk in size duringeither Uruk or Jemdet Nasr times, not to speak of the EarlyDynastic I period. At Larsa, in fact, we were unable to findsurface material antedating the Early Dynastic I period, al-though admittedly the site was given only a fairly cursory andunsystematic inspection. The listing also follows no perceptibleorder based on geographical principles.

8. Following Postgate (1978, p. 73), the Sumerian sila (Ak-kadian qa) can now be estimated with reasonable certainty as0.8 liters, on the basis of the directly measured volume of aninscribed jar. Allowing for slight expansion or contraction ofthe measuring medium, Postgate suggests a range of from 0.79to 0.82 liters around his calculated volume of 0.80687 liters.Jacobsen's published figures were on the basis of an assumedvolumetric equivalence, 1 sila = 1 liter, and have been con-verted to the new standard by multiplying by a coefficient of0.8. The same applies to later references to Jacobsen (1958).Other authorities have frequently used an equivalence of 0.842liters, and their figures too have been adjusted on the newbasis.

9. I am much indebted to Dr. Robert G. Hassert, of the staffof the University of Chicago's Computation Center, for de-signing this program. He has made available a brief technicaldescription of its algorithm, since there may be wider interestin its application (see pp. 127-29). No attempt was made toapply it to the Jemdet Nasr period data, incidentally, because ofambiguities as to site size that are outlined at the beginningof this section.

10. Niitzel 1976, p. 23. There are substantial difficulties ingeneralizing from the limited data yet available to the rathersweeping conclusions of this article, and a number of its mete-orological and statistical premises are highly questionable.But the evidence does point toward climatic trends of the kindthe author describes, even if the magnitude of the effect of thosetrends on Mesopotamian irrigation agriculture must for thepresent be left quite uncertain.

11. Frequencies in Khuzestan apparently follow a somewhatdifferent curve of development. On the basis of sherd densitiesin separate excavated samples from two sites, it has recentlybeen suggested that in that region beveled-rim bowls reacheda slightly earlier peak of popularity and then began to decline.Specifically, the recorded proportions of beveled-rim bowlsherds were 5 percent in Early Uruk, 56 percent in MiddleUruk, and 39 percent in Late Uruk (Johnson 1980, table 1).But it may be premature to speak of regional contrasts untilthe possibility of site-to-site variance within each region canbe better evaluated on the basis of more numerous excavatedsamples.

CHAPTER 4

1. An increasingly voluminous body of data on Middle East-ern settlement density now is becoming available (Kramer1980, tables 2 and 3). However, the discussion of ancient townsize and population density continues to be based essentiallyon calculations from contemporary communities where large

349

oi.uchicago.edu

Notes to Pages 144-92

distortions in traditional patterns may have been introducedby massive improvements in public health, transport, and soforth. Some statistics on Mesopotamian towns during WorldWar I may therefore be of interest, in spite of their admittedimprecision. The population estimates given below are fromthe British Mesopotamian Expeditionary Force intelligencehandbook (Admiralty War Staff, Intelligence Division 1917,vols. 2 and 3), while town areas have been calculated from avariety of large scale (1", 3", and 6" to the mile) maps drawnup contemporaneously by the MEF Survey Party:

Size in P.Locality Hectares Population per

Baghdad 372 "about 200,000"Najaf 62.3 "over 30,000"Hilla 68.7 "(1908) about 30,000"Karbala 129 "perhaps about 50,000 (ex-

cusive of a large floatingpopulation)"

Ba'quba 20.6 "4,000-5,000"Diwaniya 19.3 "4,000 (1908), but then

decreasing" (less than)Kadhimayn 42.3 "about 8,000"Nasiriya 54.6 "10,000 (?)"Tawayrij 27.3 "about 4,000 (1908)"Hai 32.9 "(1908) 4,000"Shinafiya 30.7 "(1908) 3,500"Kufa 28.9 "(1908) ... about 3,000"Qal'at al-

SikarRumaythaSamarra

ersonsHectare

538482437

14.5 "(1908) about 1,000"38.1 "(1908) 2,500"36.1 "2,000 (?)"

Most of the figures recorded above for the smaller andmedium-sized towns accord well with those for more recentsettlements. Together with accumulating data on other regionsin contemporary Iraq and Iran, they support the conclusionthat "the figure for rural settlements' density is in most caseswell below 200 persons per hectare" (Kramer 1980; cf. Alden1979, p. 68). But the four largest towns in the series (which areranked in order of decreasing density) form a striking excep-tion at the head of the list. Indeed, they closely approximatethe density just suggested for Old Babylonian living quarters,without regard for the other types of urban space utilizationthat we have assumed in most cases accompanied the latter. Isthis assumption therefore to be discounted, or were there spe-cial considerations explaining the apparently high density ofthese particular four towns? Their role in connection with theShi'ite pilgrim traffic might be one example of such a consider-ation, save that the same should then apply also to Kadhimayn.Another might have been the disproportionate use of bakedbrick, and hence of multistory residential construction. But, aswith most of the more recent data, the real effect of these fig-ures is merely to underscore the still unexplained variability insettlement density that makes the reckoning of ancient popu-lation levels so hazardous (cf. chap. 3, n. 6).

2. Professor Gelb (pers. comm.) has indicated some doubtwhether Drehem or Umma was the source from which thistablet was obtained.

3. J. W. Turner, currently writing a dissertation at Yale Uni-versity on Ur III agriculture at Umma, indicates (pers. comm.)

that yields there were sometimes fixed at 30 and sometimes at34 gur per bur. Also newly interpreting some Umma recordsas accounts of sowing rather than of harvest, he believes that aslight upward adjustment in this estimate of productivity maybe necessary.

4. I am indebted to J. W. Turner for suggesting this possi-bility.

5. I am informed by J. W. Turner (pers. comm.) that, atleast in Umma, domestic animal diets also were supplementedwith large-scale cuttings of reeds.

6. The dependence of tribal consolidation and fragmenta-tion upon state policy is vividly portrayed in the official reportof a British officer, writing in 1923 from his station in whathad been southernmost Babylonia: "The period of absence ofGovernment in the area had resulted in rapid tribal disintegra-tion. Every lilliputian leader who could raise three or four fol-lowers refused to obey his shaikh and struck out on his own.This state of affairs is inconvenient for Government, and nowa certain number of shaikhs are being recognized officially.Such men as these will be great gainers by the re-establishmentof control by Government, which means that of themselvesover their tribes also. Shaikhs will always, for the edificationof their followers, raise loud lamentations over the question oftaxes. Actually . . . they are the gainers by them. Of all thetaxes they collect for Government they retain a share forthemselves. Consequently the more Government is known tobe pressing for taxes, the more the shaikhs can squeeze out ofthe cultivators and the more they get for themselves. It wasvery noticeable that as soon as tax-collecting began in Samawaall the shaikhs blossomed out in new clothes" (J. B. Glubb toAir Headquarters, quoted in Sluglett 1976, pp. 244-45).

CHAPTER 5

1. It should not be forgotten that the city (or merely town?)of Opis, its ruins still unidentified, probably lay somewherewithin the lower Diyala region along the left bank of the Tigris.

2. Almost a third of category 3 consists of sites that couldnot be directly investigated because of long-continuing lateroccupations. Hence their assignment to this size category isonly provisional, and some of them may even have been aban-doned for a time during the Sasanian-Islamic transition.

3. The question must also be left open for the present ofhow water was supplied to this system. The bed of the modernTigris at this point is too low to permit a gravity-fed offtakeon the suggested alignment, but it is of course uncertainwhether this was the case in the first millennium B.C. By Neo-Babylonian times the pulley was in use in connection withanimal-operated lifting devices (Laessoe 1953, pp. 5-26), buteven so the apparent scale of the new canal seems to exceedwhat could reasonably be attempted with ancient liftingtechnology.

4. Traditional medieval manuscript representations of Iraqare "cognitive maps" that beautifully illustrate the ultimateconsequence of the shift. The Tigris forms a straight ribbonlined with named towns so numerous that they are almost ad-jacent to one another, down the middle of a circular field. Onthe other hand, the Euphrates had by then been reduced inimportance to a series of enclosed loops, with only a handfulof named towns, in the upper left quadrant (Miller 1927, Bd. 3:Taf. 16).

350

oi.uchicago.edu

Notes to Pages 196-242

5. The Nahr al-Malik (Akkadian nar §arri, Aramaic narmalkha) is a source of some confusion. Meaning simply "royalcanal," the name could be-and was-applied repeatedly todifferent watercourses. In Neo-Babylonian times, for example,a canal by that name extended into the interior of Uruk (Coc-querillat 1968, p. 16). Hence the mere presence of the termcannot be taken as an indication of an antecedent of the canalthat later entered the Tigris below Seleucia already in the thirdmillennium (cf. Barnett 1963, p. 13). Comparable confusionattends the widespread Iraqi Arabic use of the designation Nil(Nile) for the remains of a major watercourse. In all probabil-ity what was once the "royal canal" at Uruk is the levee nowknown locally as the Shatt al-Nil.

6. While confirming this identification of Weh Ardashir, theexcavators note that their work has been limited to the westernperimeter of the city, just inside the walls. Hence the ruins ofCtesiphon may still be found under a portion of Weh Ardashirnearer the west bank of the present course, or under the adjoin-ing east bank.

7. The tax rate on dates provides a currently insoluble prob-lem. Cf. below, p. 217, where it is argued that implied EarlyIslamic orchard densities are unrealistically low, by referenceto earlier as well as later practices. Yet the general intent of thenew Islamic rate schedules was to increase revenues sharply.If the spacing between palms in Sasanian times really was any-thing approaching as large as that implied for Early Islamictimes, there is no way that even half as much of the kharajrevenue credited to Qubadh could have been collected. Mo-rony, grappling with this problem from a different standpoint,raises the possibility that reputed Sasanian tax receipts fromthe Sawad may have been drawn instead from the entireSasanian quarter of the west. As he goes on to point out,however, that latter administrative entity appears to have beenfirst introduced by Qubadh's successor, Khusrau Anosharwan(Morony, n.d., chap. 1, sect. 2).

8. It is tempting to think of the unlocated remains of Amghi-shaya, destroyed by Khalid ibn al-Walid, said to have beenthe largest city of the Sawad (Tabari I, 2036-37; Yaqut I, 363).

But cf. Musil's discussion (1927, pp. 293-94), which moreplausibly locates it on the Euphrates southeast of Hira, well tothe west of the intensively surveyed region.

9. The failure to find much trace of Sasanian occupation inthe part of this region that was archaeologically surveyed is atleast partly to be explained by the after-effects of a fairly densenineteenth century occupation there. It has been suggested thatthe Sasanian effort was of strictly limited duration. Hence, asnoted earlier, sites would generally have been low and easilysubmerged by recent alluviation in a region where there hasbeen no subsequent deflation of the land surface by the wind.The nineteenth century canal pattern, incidentally, is an irregu-lar, opportunistic one that is easily distinguishable from theunderlying early pattern. The fine detail of field canals thatwere part of it can be seen on the base map to form patchesaround the ancient site of Isin.

10. I am indebted to Fuad Safar (pers. comm.) for this re-construction. He reports having visited the Shatt al-Kar regionnear here in the late thirties and being told by local inhabitantsof an extractive activity they had followed, for export to Ger-man chemical firms, until the eve of World War I. It involveda plant locally called shinaf, probably the same as that else-where called shinan: "Various lahophytic Chenopodiaceaewhich are grazed by camels in summer and used by the peopleas a source of alkali (in making a crude soap for washingclothes) much as 'barilla' used to be made by the people inIndia, e.g. Haloxylon, Arthrocnemum, etc." (Guest 1933, p.92). Safar was told that the extractive procedure involvedburning, boiling of the ashes, and then concentration throughboiling of a solution made from them until only a thick cakeof residue was left.

CHAPTER 6

1. A few paragraphs in this chapter either recapitulate orsummarize parts of another, quite differently oriented paperthat necessarily covered some of the same ground (Adams,n.d.).

351

oi.uchicago.edu

oi.uchicago.edu

References

Abdullah, A. K. 1967. The paramount god and the old nameof Al-Dhiba'i. Sumer 23: 189-92.

Adams, Robert McC. 1961. Agriculture and urban life in earlysouthwestern Iran. Science 136: 109-22.

--. 1965. Land behind Baghdad: A history of settlementon the Diyala plains. Chicago: University of Chicago Press.

--. 1966. The evolution of urban society: Early Mesopo-tamia and prehispanic Mexico. Chicago: Aldine.

--. 1970. Tell Abul Sarifa. A Sasanian-Islamic ceramic se-quence from south central Iraq. Ars Orientalis 8: 87-119.

--. 1972. Settlement and irrigation patterns in ancient Ak-kad. In McG. Gibson, The city and area of Kish, pp. 182-208, Appendix V. Miami: Field Research Publications.

---. 1975a. An early prehistoric site in the Warka region.Sumer 31: 11-15.

--. 1975b. An ancient Uruk threshing sledge or harrow?Sumer 31: 17-20.

. 1975c. The emerging place of trade in civilizational^ studies. In J. A. Sabloff and C. C. Lamberg-Karlovsky,

eds., Ancient civilization and trade, pp. 451-65. Albuquer-que: University of New Mexico Press.

---- . 1975d. The Mesopotamian social landscape: A viewfrom the frontier. In C. B. Moore, ed., Reconstructingcomplex societies: An archaeological colloquium, pp. 1-20. Bulletin of the American Schools of Oriental Research,suppl. 20.

---- . 1978. Strategies of maximization, stability, and resili-ence in Mesopotamian society, settlement, and agriculture.Proceedings of the American Philosophical Society 122:329-35.

- . N.d. Die Rolle des Bewaisserungsbodenbaus bei derEntwicklung von Institutionen in der altmesopotamischeGesellschaft. In J. Herrmann, ed., Die Entwicklung derProduktivkrdfte und die gesetzmdssige Abfolge der Gesell-schaftsformationen. Berlin: Zentralinstitut fiir Alte Ge-schichte und Archaiologie, Akademie der Wissenschaftender D.D.R. Forthcoming.

Adams, R. McC., and Nissen, H. J. 172. The Uruk country-

side: The natural setting of urban societies. Chicago:University of Chicago Press.

Admiralty War Staff, Intelligence Division. 1917. A handbookof Mesopotamia (C.B. 294). 4 vols. London.

Alden, J. R. 1979. Regional economic organization in Baneshperiod Iran. Ph.D. diss., Department of Anthropology,University of Michigan, Ann Arbor.

Allen, J. R. L. 1965. A review of the origin and characteristicsof recent alluvial sediments. Sedimentology 5: 89-191.

Altheim, F. 1962. Das alte Iran. In Propylden Weltgeschichte,2: 135-235. Berlin: Propylaien Verlag.

Altheim, F., and Stiehl, R. 1954. Ein asiatischer Staat: Feudalis-mus unter den Sasaniden und ihren Nachbarn. Wiesbaden:Limes-Verlag.

•- . 1963. Die aramaische Sprache unter den Achaimeni-den. Frankfurt am Main: V. Klosterman.

Alusi, S. al-. 1959. Brief statistics and notes. Sumer 15: 43-55.Ammerman, A. J.; Cavalli-Sforza, L. L.; and Wagener, D. K.

1976. Toward the estimation of population growth in OldWorld prehistory. In E. B. W. Zubrow, ed., Demographicanthropology: Quantitative approaches, pp. 27-61. Al-buquerque: University of New Mexico Press.

Andrae, W. 1903. Die Umgebung von Fara und Abu Hatab.Deutsche Orient-Gesellschaft, Mitteilungen 16: 24-30.

Baqir, T. 1946. Tell Harmal: A preliminary report. Sumer 2:22-30.

- .. 1962. Foreword, Sumer 18: 5-14.Baramki, D. C. 1944. The pottery from Khirbat al-Mafjar.

Quarterly of the Department of Antiquities of Palestine10: 65-103.

Barnett, R. D. 1963. Xenophon and the wall of Media. Journalof Hellenic Studies 83: 1-26.

Bates, D. G. 1973. Nomads and farmers: A study of the Y6riikof southeastern Turkey. Museum of Anthropology, Uni-versity of Michigan, Anthropological Papers, 52. Ann Ar-bor: University of Michigan.

Bauer, J. 1967. Altsumerische Wirtschaftstexte aus Lagasch.Ph.D. diss., Philosophischen Fakultdit, Julius-Maximilians-Universitit zu Wiirzburg.

353

oi.uchicago.edu

References

Beale, T. W. 1978. Bevelled rim bowls and their implications forchange and economic organization in the later fourthmillennium B.C. Journal of Near Eastern Studies 37: 289-313.

Berry, B. J. L. 1961. City size distributions and economic devel-opment. Economic Development and Cultural Change 9:573-88.

--. 1964. Cities as systems within systems of cities. In J.Friedmann and W. Alonso, eds., Regional developmentand planning, pp. 116-37. Cambridge: MIT Press.

Biggs, R. D. 1965. A letter from Kassite Nippur. Journal ofCuneiform Studies 19: 95-102.

--. 1973. Pre-Sargonic riddles from Lagash. Journal ofNear Eastern Studies 32: 26-33.

Bosworth, C. E. 1969. Abui 'Abdallah al-Khwarazmi on thetechnical terms of the secretary's art: A contribution to theadministrative history of medieval Islam. Journal of theEconomic and Social History of the Orient 12: 113-64.

Bottero, Jean; Cassin, Elena; and Vercoutter, Jean. 1965. TheNear East: The early civilization. New York: DelacortePress.

Braudel, F. 1972. The Mediterranean and the Mediterraneanworld in the age of Philip II. 2 vols. New York: Harperand Row.

--. 1973. Capitalism and material life 1400-1800. NewYork: Harper and Row.

Brinkman, J. A. 1965. Ur: 721-605 B.C. Orientalia 34: 241-58.--. 1968. A political history of post-Kassite Babylonia.

Analecta Orientalia, vol. 43 (Rome).--. 1969. Ur: The Kassite period and the period of As-

syrian kings. Orientalia 38: 310-48.- . 1977. Notes on Arameans and Chaldeans in southern

Babylonia in the early seventh century B.C. Orientalia 46:304-25.

-. 1979. Babylonia under the Assyrian Empire, 745-627B.C. In M. T. Larsen, ed., Power and propaganda: A sym-posium on ancient empires, pp. 223-50. Mesopotamia, 7.Copenhagen: Akademisk Forlag.

Bulliet, R. W. 1975. The Camel and the wheel. Cambridge;Harvard University Press.

Buringh, P. 1960. Soils and soil conditions in Iraq. Baghdad:Republic of Iraq, Directorate General of Agricultural Re-search and Projects.

Buringh, P., and Edelman, C. H. 1955. Some remarks about thesoils of the alluvial plain of Iraq, south of Baghdad. Neth-erlands Journal of Agricultural Science 3: 40-49.

Burrows, Eric. 1935. The Ur archaic texts: Ur excavation texts,Vol. 2. London: British Museum; Philadelphia: UniversityMuseum.

Butzer, K. W. 1976. Early hydraulic civilization in Egypt. Chi-cago: University of Chicago Press.

Cadoux, H. W. 1906. Recent changes in the course of the lowerEuphrates. Geographical Journal 38: 266-77.

Cahen, C. 1947-48. Le service de l'irrigation en Iraq au debutdu XIe siecle. Institut Frangais de Damas, Bulletin d'EtudesOrientales 12: 117-43.

. 1971. Notes pour une histoire de l'agriculture dans lespays Musulmans medievaux, 1. Journal of the Economicand Social History of the Orient 14: 63-68.

Calvot, D. 1969. Deux documents inedits de Sellis-Dagan. Re-vue d'Assyriologie 63: 101-14.

Campbell-Thompson, R. C. 1919. Excavations of the BritishMuseum at Abu Shahrein. Archaeologia, 70. London: So-ciety of Antiquaries.

Cardascia, G. 1951. Les archives des Murasu, une famille d'-hommes d'affaires babyloniens a l'dpoque perse (455-403av. J. C.). Paris: Imprimerie Nationale.

--. 1966. Faut-il eliminer gisAPIN = nartabu (une ma-chine d'irrigation)? Revue d'Assyriologie 60: 153-64.

Carlston, C. W. 1965. The relation of free meander geometry tostream discharge and its geomorphic implications. Ameri-can Journal of Science 263: 864-85.

Cassin, E.; Bottero, J.; and Vercoutter, J., eds. 1965. Die Alt-orientalische Reiche. Vol. 1. Vom Paldolithikum bis zurMitte des 2. Jahrtausends. Fischer Weltgeschichte, 2.Frankfurt am Main: Fischer Bucherei. (Also available asThe Near East: The early civilizations. London: Weiden-feld and Nicolson, 1967.)

- , eds. 1966. Die Altorientalische Reiche. Vol. 2. DasEnde des 2. Jahrtausends. Fischer Weltgeschichte, 3.Frankfurt am Main: Fischer Bucherei.

- , eds. 1967. Die Altorientalische Reiche Vol. 3. Dieerste Hdilfte des 1. Jahrtausends. Fischer Weltgeschichte, 4.Frankfurt am Main: Fischer Bucherei.

Chesney, F. R. 1868. Narrative of the Euphrates expedition ...during the years 1835, 1836 and 1837. London: Longmans.

Chicago Assyrian Dictionary. 1956- . The Assyrian diction-ary of the Oriental Institute of the University of Chicago.Chicago: Oriental Institute.

Childe, V. G. 1942. What happened in history. Harmonds-worth: Pelican Books.

Chisholm, M. 1970. Rural settlement and land use: An essay inlocation. Chicago: Aldine Publishing Co.

Clark, R. M. 1975. A calibration curve for radiocarbon dates.Antiquity 49: 251-66.

Cobb, R. 1971. Historians in white coats. (London) Times Lit-erary Supplement, 3 December 1971, pp. 1527-28.

Cocquerillat, D. 1968. Palmeraies et cultures de l'Eanna d'Uruk(559-520). Ausgrabungen der Deutschen Forschungsge-meinschaft in Uruk-Warka, 8. Berlin: Gebr. Mann.

Cooke, R. U., and Warren, A. 1973. Geomorphology in deserts.London: B. T. Batsford.

Costa, P. 1971. Islamic shrines on the Sat al-Nil. Annali dell'Istituto Orientale di Napoli 31: 1-16.

Crawford, H. E. W. 1977. The architecture of Iraq in the thirdmillennium B.C. Mesopotamia, 5. Copenhagen: Akadem-isk Forlag.

Cressey, G. B. 1960. Crossroads: Land and life in southwestAsia. Chicago: Lippincott.

Crowe, Y. 1977. Early Islamic pottery and China. Transactionsof the Oriental Ceramic Society 41: 263-78.

Crumley, C. L. 1976. Toward a locational definition of statesystems of settlement. American Anthropologist 78: 59-73.

Dandamayev, M. A. 1969. Achaemenid Babylonia. In I. M.Diakonoff, ed., Ancient Mesopotamia and socio-economichistory: A collection of studies by Soviet scholars, pp. 296-311. Moscow: Nauka.

Deimel, A. 1925a. Die altsumerischen Baumwirtschaft. Orien-talia 16: 1-87.

354

oi.uchicago.edu

References

. 1925b. Der Gemiisebau bei den alten Sumerern. Ori-entalia 17: 1-33.

---- . 1927. Religi6se Abgaben (mas-da-ri-a). Orientalia 26:1-29.-. 1931. Sumerische Tempelwirtschaft zur Zeit Uruka-

ginas und seiner Vorgainger. Analecta Orientalia vol. 2(Rome).

Delougaz, P. 1952. Pottery from the Diyala region. Oriental In-stitute Publications, 63. Chicago: University of ChicagoPress.

Department of Agricultural Statistics. 1971. Estimates of netarea, average yield and production of wheat and barley inIraq 1969, 1970. Baghdad: Central Statistical Organiza-tion.

Diakonoff, I. M. 1954. Sale of land in pre-Sargonic Sumer. InPapers presented by the Soviet delegation at the XXIIIInternational Congress of Orientalists, Cambridge, pp.19-29. Moscow: USSR Academy of Sciences.

-. 1965. Main features of the economy in the monar-chies of ancient western Asia. In Third International Con-ference of Economic History, pp. 13-32. The Hague:Mouton.

-. 1974. Structure of society and state in early dynasticSumer. Monographs of the Ancient Near East 1, fasc. 3.Los Angeles: Undena Publications.

--. 1975. The rural community in the ancient Near EastJournal of the Economic and Social History of the Orient18: 121-33.

Dietrich, M. 1970. Die Aramder Siidbabyloniens in der Sar-gonidenzeit. Alter Orient und Altes Testament, 7. Neu-kirchen-Vluyn: Neukirchener Verlag.

Dols, M. W. 1977. The black death in the Middle East. Prince-ton: Princeton University Press.

Drennan, Robert D. 1976. Fabrica San Jose and Middle Forma-tive society in the Valley of Oaxaca. Memoirs of the Uni-versity of Michigan Museum of Anthropology, no. 8. AnnArbor.

Dumond, D. E. 1975. The limitation of human population: Anatural history. Science 187: 713-21.

Edzard, D. 0. 1957. Die "Zweite Zwischenzeit" Babyloniens.Wiesbaden: Harrassowitz.

Edzard, D. 0.; Farber, G.; and Sollberger, E. 1977. Die Orts-und Gewdssernamen der prasargonischen und sargoni-schen Zeit. Rdpertoire geographique des textes cunei-formes, 1. Wiesbaden: L. Reichert.

- . 1964. Mari und Aramaier? Zeitschrift fIr Assyriologie56: 142-49.

-. 1968. Sumerische Rechtsurkunden des III. Jahrtau-sends aus der Zeit vor der III. Dynastie von Ur. BayerischeAkademie der Wissenschaften, Phil.-Hist. Klasse, Abhand-lungen, N.S. 67 (Munich).

Edzard, D. O., and Farber, G. 1974. Die Orts- und Gewdsser-namen der Zeit der 3. Dynastie von Ur. Rdpertoire gdo-graphique des textes cundiformes, 1. Wiesbaden: L.Reichert.

El-Samarraie, H. Q. 1972. Agriculture in Iraq during the 3rdcentury A.H. Beirut: Librarie du Liban.

Eph'al, I. 1978. The western minorities in Babylonia in the6th-5th centuries B.c.: Maintenance and cohesion. Orien-talia 47: 74-90.

Fahd, T. 1977. Materiaux pour l'histoire de l'agriculture enIrak: Al-filhha n-nabatiyya. In Wirtschaftsgeschichte desvorderen Orients in islamischen Zeit, part 1, pp. 276-377.Geschichte der islamischen Lainder, 6. Leiden: Brill.

Falkenstein, A. 1936. Archaische Texte aus Uruk. Ausgrabun-gen der Deutschen Forschungsgemeinschaft in Uruk-Warka, 2. Berlin: Deutsche Forschungsgemeinshaft.

--. 1954. La cit&-temple sumerienne. Cahiers d'HistoireMondiale 1: 784-814.

--. 1963. Zu den Inschriftfunden der Grabung in Uruk-Warka 1960-1961. Baghdader Mitteilungen 2: 1-82.

--. 1966. Die Inschriften Gudeas von Lagal. I. Einleitung.Analecta Orientalia, 30. Rome: Pontifical Biblical Institute.

FAO Mediterranean Development Project. 1959. Iraq countryreport. Rome: United Nations Food and AgricultureOrganization.

Fehe&vari, G. 1973. Islamic pottery: A comprehensive studybased on the Barlow Collection. London: Faber and Faber.

Fernea, R. A. 1970. Shaykh and effendi: Changing patterns ofauthority among the El Shabana of southern Iraq. Cam-bridge: Harvard University Press.

Fiey, J. M. 1968. Assyrie Chretienne III: Bet Garmai, Bet Ara-mdye et Maisan Nestoriens. Institut de Lettres Orientalesde Beyrouth, ser. 3, Orient Chretien, vol. 42. Beirut: Darel-Machreq.

Figulla, H. H., and Martin, W. J. 1953. Letters and documentsof the Old Babylonian period: Ur excavation texts, vol. 5.London: British Museum; Philadelphia: University Mu-seum.

Finster, B., and Schmidt, J. 1976. Sasanidische und Friihisla-mische Ruinen im Iraq. Baghdader Mitteilungen, vol. 8.

Fish, T. 1953. Geme at Umma. Manchester Cuneiform Studies3: 47-55.

--. 1956. More Lagash texts dated Sulgi 46 (48). Man-chester Cuneiform Studies 6: 104-13.

Flannery, K. V., ed. 1976. The early Mesoamerican village.New York: Academic Press.

Flannery, K. V., and Wright, H. T. 1967. Faunal remains fromthe "hut sounding" at Eridu, Iraq. Sumer 23: 61-63.

Frankfort, H. 1939. Cylinder seals. London: Macmillan.Fried, M. H. 1968. On the concept of "tribe" and "tribal soci-

ety." In Proceedings of the 1967 Annual Spring Meeting,American Ethnological Society, pp. 3-20. Seattle: Univer-sity of Washington Press.

Frye, R. N. 1956. Notes on the early Sasanian church andstate. In Studi Orientalistici in Onore di Giorgio Levi DellaVida, 1, pp. 314-35. Pubblicazioni dell'Istituto perl'Oriente, 52. Rome: Istituto per l'Oriente.

Gadd, C. J., and Legrain, L. 1928. The royal inscriptions: Urexcavation texts, vol. 1. London: British Museum; Phila-delphia: University Museum.

Gailani, L. al-. 1965. Tell edh-Dhiba'i. Sumer 21: 33-40.Gelb, I. J. 1965. The Philadelphia Onion Archive. In H. G.

Giiterbock and T. Jacobsen, eds., Studies in honor ofBenno Landsberger on his seventy-fifth birthday, pp. 57-62. Assyriological Studies, 16. Chicago: University of Chi-cago Press.

- . 1969. On the alleged temple and state economies inancient Mesopotamia. In Studi in Onore di Edoardo Vol-terra, 6:137-54. Rome: Giuffre Editore.

355

oi.uchicago.edu

References

.1973. Prisoners of war in early Mesopotamia. Journalof Near Eastern Studies 32: 70-98.

--. 1976. Quantitative evaluation of slavery and serfdom.In B. L. Eichler et al., eds., Kramer anniversary volume,pp. 195-207. Neukirchen-Vluyn: Neukirchener Verlag.

---. 1977. Thoughts about Ibla: A preliminary evaluation,March 1977. In Syro-Mesopotamian Studies 1: 3-29. LosAngeles: Undena.

Genouillac, Henri de. 1910. Inventaire des tablettes de Tello,vol. 2. Paris: Leroux.

Gibson, McG. 1972. The city and area of Kish. Miami: FieldResearch Publications.

Goodblatt, D. M. 1979. The poll tax in Sasanian Babylonia:The Talmudic evidence. Journal of the Economic andSocial History of The Orient 22: 233-95.

Goetze, A. 1963. Sakkanakkus of the Ur III empire. Journal ofCuneiform Studies 17: 1-31.

Gray, B. 1977. The export of Chinese porcelain to the Islamicworld: Some reflections on its significance for Islamic artbefore 1400. Transactions of the Oriental Ceramic Society41: 231-62.

Green, M. W. 1977. A note on an Archaic period geographicallist from Warka. Journal of Near Eastern Studies 36: 293-94.

Gregoire, J.-P. 1970. Archives administratives sumeriennes.Paris: Paul Geuthner

Gregory, K. J., and Walling, D. E., 1973. Drainage basin formand process: A geomorphological approach. London: Ed-ward Arnold.

Guest, E. 1933. Notes on plants and plant products, with col-loquial names in cIraq. Bulletin 27. Baghdad: Departmentof Agriculture, Government of Iraq.

--. ed., 1966. Flora of Iraq. Vol. 1. Introduction. Baghdad:Ministry of Agriculture.

Haggett, P. 1966. Locational analysis in human geography.New York: St. Martin's Press.

Hall, H. R. H. 1930. A season's work at Ur. London: Methuen.Hall, H. R. H., and Woolley, Charles Leonard. 1927. Al 'Ubaid:

Ur excavations, vol. 1. London: Oxford University Press.Hallo, W. W. 1971. Antediluvian cities. Journal of Cuneiform

Studies 23: 57-67.- . 1976. Review of G. Pettinato and H. Waetzoldt, La

Collezione Schollmeyer (Rome, 1974). Bibliotheca Orien-talis 33, no. 1/2: 38-40.

Hansen, D. P. 1965. The relative chronology of Mesopotamia.Part II. In R. W. Ehrich, ed., Chronologies in Old Worldarchaeology, pp. 201-13. Chicago: University of ChicagoPress.

Hansman, J. F. 1978. The Mesopotamian delta in the firstmillennium, B.c., Geographical Journal 144: 49-61.

Harris, S. A., and Adams R. McC. 1957. A note on canal andmarsh stratigraphy near Zubediyah. Sumer 13: 157-63.

Heinrich, E., et al. 1969-73. Erster . . . Vierter vorlaufigerBericht liber die von der Deutscher Orient-Gesellschaft mitMitteln der Stiftung Volkswagenwerk in Habtiba Kabiraunternommen archiiologischen Untersuchungen (slightlyvariant titles). Mitteilungen der Deutschen Orient-Gesell-schaft zu Berlin 101: 27-67; 102: 27-88; 103: 5-58; 105:5-68.

Hitti, P. K. 1966. The origins of the Islamic state. Beirut: Khay-ats (reprint of the 1916 edition).

Hodder, I., and Orton, C. 1976. Spatial analysis in archaeology.Cambridge: Cambridge University Press.

Hole, Frank, and Shaw, Mary. 1967. Computer analysis ofchronological seriation. Rice University Studies, vol. 53,no. 3.

Hrouda, B., ed. 1977. Isin-Isan Bahriyat I: Die Ergebnisse derAusgrabungen 1973-1974. Bayerische Akademie der Wis-senschaften, Phil.-Hist. Klasse, Abbandlungen, N.S. 79(Munich).

Hydrological Survey of Iraq. 1959. Summary of monthly pre-cipitation at stations in Iraq 1887-1958. DevelopmentBoard, Republic of Iraq. Baghdad: Times Press.

Invernizzi, A. 1976. Ten years' research in the al-Madacin area:Seleucia and Ctesiphon. Sumer 32: 167-75.

Ionides, M. G. 1954. Correspondence: The geographical his-tory of the Mesopotamian plains. Geographical Journal120: 394-95.

Jabbar, M. A. 1973. Agricultural and irrigation labourers insocial and economic life of elraq during the Umayyad andeAbbasid caliphates. Islamic Culture 47: 15-31.

Jacobsen, T. 1953. On the textile industry at Ur under Ibbi-Sin.In Studia Orientalia Ioanni Pedersen, pp. 172-87. Hel-sinki: Hauniae. (Republished in W. L. Moran, ed., Towardthe image of Tammuz, Cambridge: Harvard UniversityPress, 1970.)

--. 1954. Mesopotamian mound survey. Archaeology 7:53-54.

--. 1957. Early political developments in Mesopotamia.Zeitschrift fuir Assyriologie 52: 91-140.

-. 1958. Salinity and irrigation agriculture in antiquity.Diyala Basin Archaeological Project, Report on EssentialResults June 1, 1957-June 1, 1958. Baghdad (mimeo-graphed).

--. 1960. The waters of Ur. Iraq 22: 174-85.--. 1969. A survey of the Girsu (Telloh) region. Sumer

25: 103-9.Jacobsen, T., and Adams, R. McC. 1958. Salt and silt in

ancient Mesopotamian agriculture. Science 128: 1251-58.Johnson, G. A. Local exchange and early state development

in southwestern Iran. Museum of Anthropology, Uni-versity of Michigan, Anthropological Papers, 51. AnnArbor: University of Michigan.

--. 1975. Locational analysis and the investigations ofUruk local exchange systems. In J. A. Sabloff and C. C.Lamberg-Karlovsky, eds., Ancient civilization and trade.pp. 285-339. Albuquerque: University of New MexicoPress.

--. 1976. Early state organization in southwestern Iran:Preliminary field report. In Proceedings of the IVth An-naul Symposium on Archaeological Research in Iran,1975, pp. 190-223. Tehran: Muzeh-e Iran-e Bastan.- . 1977. Aspects of regional analysis in archaeology,Annual Review of Anthropology 6: 497-508.

- . 1980. The changing organization of Uruk administra-tion on the Susiana plain. In F. Hole, ed., Archaeologicalperspectives on Iran: From prehistory to the Islamic con-quest. Albuquerque: University of New Mexico Press.Forthcoming.

356

oi.uchicago.edu

References

Jones, T. B. 1976. Sumerian administrative documents: Anessay. In S. J. Lieberman, ed., Sumerological studies inhonor of Thorkild Jacobsen, pp. 41-61. AssyriologicalStudies, 20. Chicago: University of Chicago Press.

Jones, T. B., and Snyder, J. W. 1961. Sumerian economic textsfrom the Third Ur Dynasty. Minneapolis: University ofMinnesota Press.

Keall, E. J. 1975. Parthian Nippur and Vologases' southernstrategy: An hypothesis. Journal of the American OrientalSociety 95: 620-32.

Kiepert, H. 1883. Ruinenfelder der Umgegend von Babylon.Zeitschrift der Gesellschaft fiir Erdkunde zu Berlin 18:Taf. 1.

Kirkby, M. J. 1977. Land and water resources of the DehLuran and Khuzistan plains. In F. Hole, Studies in the archaeo-

logical history of the Deh Luran plain, Appendix 1, pp.251-88. Museum of Anthropology, University of Michi-gan, Memoir 9. Ann Arbor: University of Michigan.

Kister, M. J. 1968. Al-Hira: Some notes on its relations withArabia. Arabica 15: 143-69.

Klengel, H. 1976. Hammurapi von Babylon und seine Zeit.Berlin: Akademie der Wissenschaften.

Koschaker, P. 1942. Zur staatlichen Wirtschaftsverwaltung inaltbabylonischer Zeit, insbesondere nach Urkunden ausLarsa. Zeitschrift fir Assyriologie 47: 135-80.

Kramer, C. 1980. Estimating prehistoric populations: Anethnoarchaeological approach. In Colloque InternationalC.N.R.S. no. 580 (13-15 June 1978), L'Archeologie del'Iraq du debut de l'epoque neolithique a 333 avant notreere: Perspectives et limites de l'interpretation anthropolo-gique des documents. Paris: C.N.R.S. Forthcoming.

Kraus, F. R. 1954. Le role des temples depuis la troisemedynastie d'Ur jusqu'a la premiere dynastie de Babylone.Cahiers d'Histoire Mondiale 1: 522-36.

-. 1955. Provinzen des neusumerischen Reiches von Ur.Zeitschrift fuir Assyriologie 51: 45-75.

Laessoe, J. 1953. Reflections on modern and ancient Orientalwater works. Journal of Cuneiform Studies 7: 5-26.

Landsberger, B. 1949. Jahreszeiten im Sumerisch-Akkadischen.Journal of Near Eastern Studies 8: 248-97.

--. 1967. The date palm and its by-products according tothe cuneiform sources. Archiv fur Orientforschung, suppl.17.

Lane, A. 1947. Early Islamic pottery. London: Faber and Faber.

Larsen, C. E. 1975. The Mesopotamian delta region: A recon-sideration of Lees and Falcon. Journal of the AmericanOriental Society 95: 43-57.

Larsen, C. E., and Evans, G. 1978. The Holocene geologicalhistory of the Tigris-Euphrates-Karun delta. In W. C.Brice, ed., The environmental history of the Near andMiddle East since the last ice age, pp. 227-44. London:Academic Press.

Layard, A. H. 1897. Nineveh and Babylon. London: J. Murray.Le Breton, L. 1957. The early periods at Susa, Mesopotamian

relations. Iraq 19: 79-124.Leemans, W. F. 1975. The r61e of landlease in Mesopotamia in

the early second millenium B.C. Journal of the Economicand Social History of the Orient 18: 134-45.

---- . 1975-76. Die Arten der Zuverfiigungsverstellung von

Boden ffir landwirtschaftliche Zwecke in der altbablyonischen Zeit. Welt des Orients 8: 241-53.

Lees, G. M., and Falcon, N. L. 1952. The geographical historof the Mesopotamian plains. Geographical Journal 11824-39.

Lenzen, H. J. 1965. XXI. Vorldufiger Bericht uiber die von derDeutschen Archdologischen Institut und der Deutsche:Orient-Gesellschaft aus Mitteln der Deutschen Forschungsgemeinschaft unternommen Ausgrabungen in UrukWarka. Berlin: Gebr. Mann.

Le Strange, G. 1895. Description of Mesopotamia and Baghdad, written about the year 900 A.D. by Ibn SerapiorJournal of the Royal Asiatic Society 1895: 1-76, 255-315

--. 1905. The lands of the Eastern Caliphate. CambridgeCambridge University Press (reprinted 1966, LondonFrank Cass).

Lieberman, S. J. 1968-69. An Ur III text from Drehem recording "Booty from the land of Mardu." Journal of Cuneiform Studies 22: 53-62.

Limet, H. 1972. L'etranger dans la soci6te sumerienne. In D. 0Edzard, ed., Gesellschaftsklassen im Alten Zweistromlan<und in den angrenzenden Gebieten. Bayerische Akademider Wissenschaften, Phil-Hist. Klasse, AbhandlungeriN.S., 75: 123-38.

Lloyd, S., and Safar, F. 1943. Tell Uqair: Excavations by thIraq Government Directorate of Antiquities in 1940 an<1941. Journal of Near Eastern Studies 2: 131-58.

--. 1948. Eridu: Preliminary report on the second season's excavations. Sumer 4: 115-25.

Loftus, W. K. 1856. Notes of a journey from Baghdad tiBusrah, with descriptions of several Chaldean remainsJournal of the Royal Geographical Society of London 26131-53.

-. 1857. Travels and researches in Chaldea and SusianaLondon: J. Nisbet.

Lokkegaard, F. 1950. Islamic taxation in the Classic perioaCopenhagen: Folkeuniversitetets Bibliotek.

Longrigg, S. H. 1925. Four centuries of modern Iraq. OxfordClarendon Press.

Luckenbill, D. D. 1924. The annals of Sennacherib. OrientaInstitute Publications, 2. Chicago: University of ChicagoPress.

McCown, D. E., and Haines, R. C. 1967. Nippur I. Temple oEnlil, scribal quarter, and soundings. Oriental InstitutPublications, 78. Chicago: University of Chicago Press.

McCown, D. E.; Haines, R. C.; and Biggs, R. D. 1978. NippuII. The north temple and sounding E. Oriental Institut,Publications, 97. Chicago: University of Chicago Press.

Maekawa, K. 1973-74. The development of the 4-mi in Lagaslduring Early Dynastic III. Mesopotamia 8-9: 77-144.

- . 1974. Agricultural production in ancient Sumer, chieflfrom Lagash materials. Zinbun (Memoirs of the ResearclInstitute for Humanistic Studies, Kyoto University) 131-60.

- . 1976. The Erin-People in Lagash of Ur III times. Revu,d'Assyriologie 70: 9-44.

Malecki, E. J. 1975. Examining change in rank-size systems ocities. Professional Geographer 27: 43-47.

Maricq, A. 1959. Vologesias, l'emporium de Ctisiphon. Syrii36:264-76.

357

oi.uchicago.edu

References

Martin, R. 1956. L'urbanisme dans la Grece antique. Paris:Editions A. & J. Picard.

Miller, K. 1926-31. Mappae Arabicae: Arabische Welt- undLdnderkarte. 6 vols. Stuttgart: Privately printed.

Ministry of Economics, Republic of Iraq. 1958. Statistical ab-stract 1957. Baghdad: Zahrac Press.

Mitchell, R. C. 1958. Instability of the Mesopotamian plains.Bulletin Societe Geographique d'Egypte 31: 127-40.

Moritz, B. 1888. Zur Geographie und Ethnographie von Siid-Mesopotamien. Gesellschaft fuir Erdkunde zu Berlin, Ver-handlungen 15: 185-201.

Morony, M. G. 1972. Seventh century Iraq. Ph.D. diss., De-partment of History, University of California, Los Angeles.

--. 1976. The effects of the Muslim conquest on the Per-sian population of Iraq. Iran 14: 41-59.

- . N.d. Seventh century Iraq. Revisions to Ph.D. diss.,Department of History, University of California, LosAngeles.

Miinnich, K. 0. 1957. Heidelberg natural radiocarbon measure-ments I. Science 126: 194-99.

Musil, A. 1927. The Middle Euphrates. Oriental Explorationsand Studies, 3. New York: American Geographical So-ciety.

Mustafa, M. A. 1949. Soundings at Tell Al-Dhiba'i. Sumer 5:173-97.

Naquib, K. A. al-. 1967. Geology of the Arabian Peninsula:Southwestern Iraq. United States Geological Survey Pro-fessional Papers no. 560-G. Washington, D.C.

Naumann, R., and Huff, D. 1975. Takht-i Suleiman: Berichtiiber die Ausgrabungen 1965-1973. Archdologischer An-zeiger 90: 109-204.

Neusner, J. 1966. A history of the Jews in Babylonia. Vol. 2.The Early Sasanian period. Studia Post-Biblica, 11. Leiden:Brill.

. 1969. A history of the Jews in Babylonia. Vol. 4. Theage of Shapur II. Studia Post-Biblica, 14. Leiden: Brill.

Newman, J. 1932. The agricultural life of the Jews in Babyloniabetween the years 200 C.E. and 500 C.E. London: OxfordUniversity Press.

Niebuhr, C. 1968. Reisebeschreibung nach Arabien. 3 vols.Graz: Akandemische Druck- und Verlagsanstalt (reprintof the 1774-78 edition).

Nissen, H. J. 1968. Survey of an abandoned modern village.Sumer 24: 107-14.

-. 1970. Grabung in den Quadraten K/L XII in Uruk-Warka. Baghdader Mitteilungen 5: 101-91.

7 - . 1972. The city wall of Uruk. In Man, settlementand urbanism, ed. P. J. Ucko. R. Tringham, and G. W.Dimbleby, pp. 703-98. London: Duckworth.

- . 1973. Siidbabylonien in parthischer und sasanidischerZeit. Baghdader Mittielungen 6: 79-86.

- . 1974. Zur Frage der Arbeitsorganisation in Babylonienwihrend der Spaituruk-Zeit. Acta Antiqua AcademiaeScientiarum Hungaricae 22: 5-14.

- . 1976. Geographie. In S. J. Lieberman, ed., Sumerologi-cal studies in honor of Thorkild Jacobsen, pp. 9-40. As-syriological Studies, 20. Chicago: University of ChicagoPress.

Nodelman, S. A. 1960. A preliminary history of Characene.Berytus 13: 83-121.

N61ldeke, T. 1973. Geschichte der Perser und Araber zur Zeitder Sasaniden, aus der arabischen Chronik des Tabari.Graz: Akademische Druck- und Verlagsanstalt (reprint ofthe 1879 edition).

Niitzel, W. 1975. The formation of the Arabian Gulf from14,000 B.C. Sumer 31: 101-10.

--. 1976. The climate changes of Mesopotamia and bor-dering areas 14,000-2000 B.C. Sumer 32: 11-24.

Oates, D., and Oates, J. 1977. Early irrigation agriculture inMesopotamia. In G. Sieveking, I. H. Longworth, and K. E.Wilson, eds., Problems in economic and social archaeol-ogy, pp. 109-35. London: Duckworth.

Oates, J. L. 1960. Ur and Eridu: The prehistory. Iraq 22: 32-50.

-. 1968. Prehistoric investigations near Mandali, Iraq.Iraq 30: 1-20.

-. 1973. The background and development of earlyfarming communities in Mesopotamia and the Zagros.Proceedings of the Prehistoric Society 39: 147-81.

--. 1976. Prehistory in northeastern Arabia. Antiquity 50:20-31.

Oates, J. L.; Davidson, T. E.; Kamilli, D.; and McKerrell, H.1977. Seafaring merchants of Ur? Antiquity 51: 221-34.

Obermeyer, J. 1929. Die Landschaft Babyloniens im Zeitalterdes Talmuds und des Gaonats. Gesellschaft zur Forderungder Wissenschaft des Judentums, Schriften 30. Frankfurtam Main: I. Kauffmann.

Oelsner, J. 1975. Randbemerkungen zur arsakidischen Ge-schichte anhand von babylonischen Keilschrifttexten. Alt-orientalische Forschungen 3: 25-45.

--. 1978. Kontinuitdit und Wandel in Gesellschaft undKultur Babyloniens in hellenistischer Zeit. Klio-Beitragezur Alten Geschichte 60: 101-16.

Oppenheim, A. L. 1950. Assyro-Babylonian religion. In V. T. A.Ferm, ed., Forgotten religions, pp. 63-79. New York:Philosophical Library.

--. 1964. Ancient Mesopotamia: Portrait of a dead civili-zation. Chicago: University of Chicago Press.

- .. 1967. Essay on overland trade in the first millenniumB.C. Journal of Cuneiform Studies 21: 236-54.

--. 1969. Mesopotamia-Land of many cities. In I. M.Lapidus, ed., Middle Eastern cities, pp. 3-18. Berkeley:University of California Press.

--. N.d. The Babylonian evidence of Achaemenian rule inMesopotamia. Manuscript prepared (ca. 1969) for theCambridge History of Iran, 2.

Paepe, R. 1971. Geological approach of the Tell ed-Der area.In L. de Meyer, H. Gasche, and R. Paepe, Tell ed-Dir I,pp. 9-27. Louvain: Editions Peeters.

Paepe, R., and Baeteman, C. 1978. The fluvial system betweenTell ed-Der and Tell Abu Habbah. In L. de Meyer, ed.,Tell ed-Der II, pp. 37-56. Louvain: Editions Peeters.

Palerm, A., and Wolf, E. R. 1957. Ecological potential andcultural development in Mesoamerica. In Studies inhuman ecology, Social Science Monographs, 3. Washing-ton, D. C.: Pan American Union.

Parsons, J. R. 1972. Archaeological settlement patterns. An-nual Review of Anthropology 1: 127-50.

Perier, Jr. 1904. Vie d'Al-Hadjdjadj ibn Yousof (41-95 AH =

358

oi.uchicago.edu

References

661-714 AD) d'apres les sources arabes. Paris: LibrarieEmile Bouillon.

Perkins, A. L. 1949. The comparative archaeology of earlyMesopotamia. Studies in Ancient Oriental Civilization, 25.Chicago: University of Chicago Press.

Pettinato, G. 1970-71. IlIdigna-ta 17-Nun-Se: I1 conflitto trasLagas ed Umma per la "Frontiera Divina" e la sua solu-zione durante la Terza Dinastia di Ur. Mesopotamia 5-6:281-320.

Pettinato, G., and Waetzoldt, H. 1975. Saatgut und Furchenab-stand beim Getreideanbau. Studia Orientalia 46: 259-90.

Pfeiffer, R. H. 1935. State letters of Assyria. American Orien-tal Series, 6. New Haven: American Oriental Society.

Pigulevskaja, N. 1963. Les villes de l'dtat Iranien aux dpoquesparthe et sassanide. Paris: Mouton.

Popovic, A. 1976. La revolte des esclaves en Iraq au IIeP/IXesiecle. Paris: Geuthner.

Postgate, J. N. 1976. Inscriptions from Tell al-Wilayah. Sumer32: 77-100.

---- . 1977. Excavations at Abu Salabikh, 1976. Iraq 39:233-63.

---- . 1978. An inscribed jar from Tell al Rimah. Iraq 40:71-75.

Postgate, J. N., and Moorey, P. R. S. 1976. Excavations at TellAbu Salabikh, 1975. Iraq 38: 133-69.

Poyck, A. P. G. 1962. Farm studies in Iraq: An agro-economicstudy of the agriculture in the Hilla-Diwaniya area inIraq. Mededelingen van de Landbouwhogeschool teWageningen, vol. 62, no. 1.

Principal Bureau of Statistics. 1954. Report on the agricultureand livestock census of Iraq, 1952-53. Vol. 1. Baghdad:Al-Akhbar House.

Ralph, E. K.; Michael, H. N.; and Han, M. C. 1973. Radio-carbon dates and reality. MASCA Newsletter 9: 1-20.

Redman, C. L. 1971. Controlled surface collection as an inte-gral element in archaeological research strategies. Ph.D.diss., Department of Anthropology, University of Chicago.

Reitlinger, G. 1935. Islamic pottery from Kish. Ars Islamica 2:198-318.

Renfrew, A. C., and Dixon, J. 1976. Obsidian in western Asia:A review. In G. Sieveking, I. H. Longworth, and K. E.Wilson, eds., Problems in economic and Social archaeol-ogy, pp. 137-50. London: Duckworth.

Renger, J. 1970. Zur Lokalisierung von Karkar. Archiv furOrientforschung 23: 73-78.

-- . 1972. Flucht als soziales Problem in der altbabyloni-schen Gesellschaft. In D. 0. Edzard, ed,. Gesellschafts-klassen im Alten Zweistromland und in den angrenzendenGebieten, pp. 167-82. Bayerische Akademie der Wissen-schaften, Phil.- Hist. Klasse, Abhandlungen, N.S., 75.

Ricciardi, R. V. 1967. Pottery from Choche. Mesopotamia 2:93-104.

Rice, D. T. 1934. The Oxford excavations at Hira. Ars Islamica1:51-73.

Rosen-Ayalon, M. 1974. La poterie Islamique. Memoires de laD6lkgation Arch6ologique en Iran, 50. Paris: Paul Geuth-ner.

Roux, G. 1960. Recently discovered ancient sites in the Ham-mar Lake district (southern Iraq). Sumer 16: 20-31.

Rowton, M. B. 1969. The role of watercourses in the growthof Mesopotamian civilization. In W. R611ig, ed., LianMithurti: Festschrift fur Wolfram Freiherr von Soden, pp.307-16. Neukirchen-Vluyn: Neukirchener Verlag.

--. 1973. Autonomy and nomadism in western Asia.Orientalia 42: 247-58.

--. 1976. Dimorphic structure and topology. Oriens An-tiquus 15: 17-31.

Russel, J. C. 1957. Tillage practices in Iraq. Iraq College ofAgriculture, Abu Ghraib (mimeographed).

Russell, J. C. 1958. Late ancient and medieval populations.Transactions of the American Philosophical Society, vol.48, no. 3.

--. 1972. Population in Europe 500-1500. In C. M. Ci-polla, ed., The Middle Ages, pp. 25-70. Fontana economichistory of Europe, vol. 1. Glasgow: Collins-Fontana.

Russell, R. J. 1967. River plains and sea coasts. Berkeley: Uni-versity of California Press.

Safar, F. 1945. Wasit: The sixth season's excavations. Cairo:Imprimerie de l'Institut Franqais d'Archeologie Orientale.

-. 1947. Eridu: The excavations. Sumer 3: 100-111.-. 1950. A preliminary report on the third season's exca-

vations. Sumer 6: 27-38.Said, E. W. 1978. Orientalism. New York: Pantheon.Salonen, A. 1968. Agricultura Mesopotamia nach sumerisch-

akkadischen Quellen. Annales Academiae ScientiarumFennicae, ser. B., vol. 149 (Helsinki).

-. 1970. Die Fischerei im alten Mesopotamien nachsumerisch-akkadischen Quellen. Annales Academiae Sci-entiarum Fennicae, ser. B., vol. 166, (Helsinki).

San Nicolo, M. 1948. Materialien zur Viehwirtschaft in denneubabylonischen Tempeln, I. Orientalia, n.s., 17: 273-93.

Sarre, F. 1925. Die Keramik von Samarra. Die Ausgrabungenvon Samarra, vol. 2. Berlin: Dietrich Reimer.

Sauren, H. 1966. Topographie der Provinz Umma nach denUrkunden der Zeit der III. Dynastie von Ur. Teil I. Kanaileund Bewasserungsanlagen. Ph.D. diss., Hohen Philoso-phischen Fakultait, Ruprecht-Karl Universitait zu Heidel-berg.

Schilstra, J. 1962. Irrigation as a soil- and relief-forming factorin the lower Mesopotamian plain. Netherlands Journalof Agricultural Science 10: 179-93.

Schmandt-Besserat, D. 1977. An archaic recording system andthe origin of writing. Syro-Mesopotamian Studies 1, no.2. Los Angeles: Undena.

Schmidt, E. F. 1931. Excavations at Fara, 1931. MuseumJournal 22: 193-217.

Schmidt, J. 1978. Tell Mismar: Ein praihistorischer Fundort imSiidiraq. Baghdader Mitteilungen 9: 10-17.

Schneider, N. 1927. Der Viehbestand des &-gal in Lagas. Archivfuiir Orientforschung 4: 206-8.

Schnyder, R. 1974. Medieval incised and carved wares fromnorth west Iran. In W. Watson, ed., The art of Iran andAnatolia, from the 11th to the 13th centuries A.D., pp. 85-94. Percival David Foundation of Chinese Art, Colloquieson Art and Archaeology in Asia, 4. London: School ofOriental and African Studies, University of London.

Schumm, S. A. 1977. The fluvial system. New York: Wiley-Interscience.

359

oi.uchicago.edu

References

Selby, W. B.; Collingwood, W., and Bewsher, J. B. 1885. Sur-veys of ancient Babylon and the surrounding ruins, withpart of the rivers Tigris and Euphrates. . . . 6 sheets.London: H. M. Secretary of State for India in Council.

Skinner, G. W., ed., 1977. The city in late imperial China. Stan-ford: Stanford University Press.

Sluglett, P. 1976. Britain in Iraq 1914-1932. St. Anthony'sMiddle East Monographs, 4. London: Ithaca Press.

Smith, S. 1932. Notes on the Gutian period. Journal of theRoyal Asiatic Society 1932: 295-308.

- . 1954. Correspondence: The geographical history ofthe Mesopotamian plains. Geographical Journal 120: 395-96.

Sollberger, E., and Kupper, J.-R. 1971. Inscriptions royalessumeriennes et akkadiennes. Paris: Editions du Cerf.

Stolper, M. 1974. Management and politics in late Achae-menid Babylonia: New texts from the Murasu archive. 2vols. London: University Microfilms.

Stone, E. C. 1977. Economic crisis and social upheaval in OldBabylonian Nippur. In L. D. Levine and T. C. Young, Jr.,eds., Mountains and lowlands, pp. 267-89. BibliothecaMesopotamica, 7. Los Angeles: Undena.

Streck, M. 1917. Seleukia und Ktesiphon. Der Alte Orient, vol.16, no. 3-4.

Stronach, David. 1961. The excavations at Ras al 'Amiyah. Iraq23: 95-137.

Taylor, J. E. 1855. Notes on Abu Shahrein and Tell el Lahm.Journal of the Royal Asiatic Society 15: 404-15.

Ubell, K. 1971. Iraq's water resources. Nature and Resources7, no. 2: 3-15.

Vita-Finzi, C. 1978. Recent alluvial history in the catchment ofthe Arabo-Persian Gulf. In W. C. Brice, ed., The environ-mental history of the Near and Middle East since the lastice age, pp. 255-61. London: Academic Press.

von Haller, A. 1932. Die Keramik der archaischen Schichtenvon Uruk. In A. No1deke et al., Vierter vorldufiger Berichtiiber die von der Notgemeinschaft der Deutschen Wissen-schaft in Uruk unternommen Ausgrabungen, pp. 31-47.Berlin: Walter de Gruyter.

Voute, Caesar. 1957. A prehistoric find near Razzaza. Sumer13: 135-56.

Waetzoldt, H. 1972. Untersuchungen zur neusumerischen Tex-tilindustrie. Centro per le Antichita e la Storia dell'Artedel Vicino Oriente, Studi Economici e Tecnologici, 1.Rome: Centro per le Antichita e la Storia dell'Arte delVicino Oriente.

Waines, D. 1977. The third century internal crisis of the Abba-sids. Journal of the Economic and Social History of theOrient 20: 282-306.

Walters, S. D. 1970. Water for Larsa: An Old Babylonian ar-chive dealing with irrigation. New Haven: Yale Univer-sity Press.

Watson, W. 1970. On Tang China. Percival David Foundationof Chinese Art, Colloquies on Art and Archaeology in

Asia, 1. London: School of Oriental and African Studies,University of London.

Weiss, H. 1977. Periodization, population and early state for-mation in Khuzistan. In L. D. Levine and T. C. Young, Jr.,eds., Mountains and lowlands, pp. 347-69. BibliothecaMesopotamica, 7. Los Angeles: Undena.

Wenke, R. J. 1975-76. Imperial investment and agriculturaldevelopment in Parthian and Sasanian Khuzestan: 150B.C. to A.D. 640. Mesopotamia 10-11: 31-221.

Wilcke, C. 1969-70. Zur Geschichte der Amurriten in der Ur-III-Zeit. Die Welt des Orients 5: 1-31.

Wilkinson, J. C. 1977. Water and tribal settlement in south-eastAra ia: A study of the Aflaj of Oman. Oxford: ClarendonPress.

Wiedemann, E. 1906. Beitriige zur Geschichte der Naturwis-senschaften, X. Physikalisch-Medizinischen Sozietdt in Er-langen, Sitzungsberichte 38: 307-57.

W61fel, W. 1962. Probleme um Wasser im Irak. Acta Hydro-physica 7: 151-83.

Woolley, Charles Leonard, 1939. The ziggurat and its sur-roundings. Ur excavations, vol. 5. London: BritishMuseum; Philadelphia: University Museum.

---- 1956. The early periods: Ur excavations, vol. 4. Lon-don: British Museum; Philadelphia: University Museum.

---- . 1962. The Neo-Babylonian and Persian periods: Urexcavations, vol. 9. London: British Museum; Philadel-phia: University Museum.

--. 1965. Excavations at Ur. Apollo ed. New York:Crowell.

---- . 1974. The Ur III period: Ur excavations, vol. 6. Lon-don: British Museum; Philadelphia: University Museum.

Wright, G. A. 1973. Bristlecone pine calibrations of radiocar-bon dates: Some examples from the Near East. AmericanJournal of Archaeology 77: 197-201.

Wright, Henry T. 1967. A Note on a Paleolithic site in thesouthern desert. Sumer 22: 101-4.

---- . 1969. The administration of rural production in anearly Mesopotamian town. Museum of Anthropology,University of Michigan, Anthropological Papers, 38. AnnArbor: University of Michigan.

--. 1977. Recent research on the origin of the state. InAnnual Review of Anthropology, 6: 379-97. Palo Alto:Annual Reviews.

Wright, H. T., and Johnson, G. A. 1975. Population, exchange,and early state formation in southwestern Iran. AmericanAnthropologist 77: 267-89.

Wright, Henry T.; Neely, J. A.; Johnson, G. A.; and Speth,John. 1975. Early fourth millennium developments insouthwestern Iran. Iran 13: 129-47.

Yoffee, N. 1977. The economic role of the crown in the OldBabylonian period. Bibliotheca Mesopotamica, 5. Los An-geles: Undena.

Zipf, G. K. 1949. Human behavior and the principle of leasteffort. Cambridge: Addison-Wesley.

360

oi.uchicago.edu

Index

'Abbasid Caliphate, 18, 183, 215-16,223, 225

Abu Salabikh, Tell, 63, 67Abu Sarifa, Tell, 38, 46Achaemenian period, 177-79, 185-92,

194, 228-30Adab, 16, 63, 88, 145, 164Administrative functions and "density,"

76-78, 80-81, 144, 151, 168, 170,200-201, 204, 213, 215, 246-47, 329

Aeolian deposition and erosion, 10, 11,22, 30-31, 36, 38, 39

Aerial photographs, 28, 29, 46, 297-98,335

Agriculture: ancient regimes, 86-87, 90-93, 142, 144-50, 151-52, 181, 186-88, 203-4, 216; decline of, 151-52,184, 213, 215, 218, 252, 334; flexi-bility or resilience of, 23-24, 53, 250;modern regime, 5-6, 13, 32; stabili-zation of, 20, 250; traditional regime,4,20-21,23-24,26

Akkad, 55, 58, 155, 251Akkadian period, 160, 162, 164, 170,

171, 310-11, 328-29Akkad Survey, 35, 55, 61, 190, 206Akshak or Upi (later Opis), 7, 158Alluvial deposition, 10, 11, 36, 38Arab, Shatt al-, 1, 9Arabo-Persian Gulf, geographic history

of, 14-17, 31, 54, 62, 251Archaeological sites: later disturbance

of, 10, 44-45, 53; recording of, 44-46,298

Archaeological survey: assumptions, ob-jectives, procedures, and limitationsof, 27, 29, 30-43, 45, 47, 49-53, 55,130-31, 143, 176, 183, 228-29, 232,

234, 236-38, 242-43, 297-98, 300,329; boundaries of, 35-37

Architecture, public, 59, 66, 77, 85, 244,329-30

Assyria, Babylonian conquests of, 152-55

Babylon, 134, 158, 190, 251Baghdad, 2, 7, 46, 184, 192, 218, 223,

225Basra, 184, 218Beveled-rim bowls, 77-79, 116, 119,

122-26, 304Borsippa, 19, 158Buranunu. See Euphrates River

Camel, 168, 170, 176, 336Cassite period, 18, 152, 159, 165, 167,

168, 171, 174, 188, 316-19, 331-34Central place theory, 53, 54, 67Charax Spasinou, 2, 192, 199City. See Urban; SettlementClimate, 11-13, 94, 130"Cognitive maps," 25-26, 192nCone mosaics, clay, 50, 77, 304, 306,

325, 327Copper, 66, 78-80, 211, 325, 330Crops, 86-87, 142, 146, 149, 186-87,

203-4Ctesiphon, 2, 7, 46, 180, 194

Dalmaj, Haur, 16, 37Dayr, Tell al-, 16Demographic trends, 61, 69, 70, 75, 90,

142, 143, 178-80, 184-85, 196, 252Dilbat, 19, 158Directorate General of Antiquities, sur-

veys by, 15, 35-37, 43Diyala plain, 18, 46-49, 55, 57, 58, 178-

361

79, 180, 182, 185, 192, 194, 224-27,251,252

Diyala Survey, 46-49Dunes, 22, 30-32, 38, 39, 300, 326Dynastic integration and fragmentation,

133, 150, 151, 168, 183, 213, 214, 247,250,328

Early Dynastic I period, 82-90, 93-94,127, 304, 306-10, 327-28

Early Dynastic II/III period, 88, 132,160-61, 170, 304, 306-10, 327-28

Early Islamic period, 7, 183-85, 214-20,223-25,234-39

Epidemic disease, 214Eridu, 58, 323-27, 334Eridu period, 55, 58, 116, 301-2, 323-25Euphrates River: anastomosing regime

and hydrology, 1, 3, 4, 7-9, 18, 19, 22,155; course and branches, 1-3, 18-21,155, 159, 197, 300, 327, 328, 330, 333,334 (see also Hilla, Shatt al-; Hindiya,Shatt al-); floods, 6, 16, 141; and ir-rigable area, 6, 210

Falaika Island (ancient Icarus), 15Fallow system, 6, 86, 146, 151, 204, 210Fara. See ShuruppakFauna, 142, 147-49, 190,203-4,323Fishing, 16, 58, 66, 86, 142, 204, 324Flood control, 6, 21Fortifications, 26, 82, 244, 330, 332, 335

Gharraf, Shatt al-, 6, 31, 32, 36Girsu, 66, 86, 87, 146, 147Glass manufacture, 211, 213, 335

Hajji Muhammad period, 55, 58, 116,301-2, 323-25

oi.uchicago.edu

Hammar, Haur al-, 1, 333Hazim, 300, 326, 327Herding, 11, 13, 21, 23, 24, 26, 59, 66,

68, 142, 147-49, 151, 190, 213, 247,250, 323-24, 327

Hilla, Shatt al-, 18, 19, 23, 209Hindiya, Shatt al-, 18, 23, 46, 210Hira, al-, 205-6, 210Hunter-gatherers, 54, 323

Ilkhanid period, 183, 225, 227, 240Irrigation: canals and facilities, 8, 10,

18-24, 140, 144, 165, 186-88, 194,196, 201, 202, 204, 210-11, 216,245, 324, 326-27, 332, 334; jointEuphrates-Tigris, 7, 22, 210-11, 216;social effects of, 243-48

Irrigation systems: large-scale, 22, 30,32, 53, 165, 188, 196, 209-11, 245;local, 22, 30, 59, 144, 245, 323

Isin, 10, 36Isin-Larsa period, 18, 31, 163, 171, 310,

312-15, 329-31

Jemdet Nasr period, 81-83, 126-27, 245,304, 306-10, 326-27

Karun River, 9Khuzestan. See SusianaKilns, ceramic, 50, 78, 122, 325-27, 330Kinglist, Sumerian, 3, 82, 132Kish, 88, 134, 251Kish Survey, 61, 190, 206Kufa, 184, 218

Lagash, 15, 134, 160, 251, 329Lahm, Tell al-, 326, 330, 332-34LANDSAT, 33-34, 160, 190, 206, 209,

210Late Islamic period, 184, 222, 225, 228,

234-41Levees, ancient, 10, 20, 22, 30, 32, 33,

244, 300Linen textiles, 186, 202

Mandali district, 55, 58Map coverage, 28, 297Marad, 19, 149, 158, 190Meanders, river, 8, 9, 16; ancient, 16, 17,

31, 61-63, 164-65; modern, 16, 17Mesopotamian alluvium, historical geo-

morphology, 16-19, 300-301Middle Babylonian period, 152-55, 168,

169, 174, 177-78, 188,316-19, 333-34Middle Islamic period, 184, 221, 223,

225, 226, 234-40Military policies, 63, 150, 175-76, 200,

247

Index

Nahrawan canal, 194, 196, 211, 213,218, 225, 246

Neo-Babylonian period, 177-78, 179,185-92, 194, 228-30, 319-22, 334,335

Nil, Shatt al-, 32, 33, 42, 206, 223,246Nippur, 16, 59, 63, 67, 88, 132, 140, 146,

147, 150, 153, 159, 168, 186-88, 190,199

Obsidian, 66, 78, 79Old Babylonian period, 18, 21, 31, 165-

66, 171,310, 314-15, 330-32

Pastoral peoples, 135, 136, 147, 151, 330,336-37

Pleistocene, 15, 54, 300Population: density, 69,,85, 140, 143,

144, 146, 149, 168, 180, 251, 325, 327;movements, 63, 70, 84, 88, 130, 135,150, 154, 160, 177-78, 184, 196, 218,223, 225, 331. See also Demographictrends

Purattu. See Euphrates River

Radiocarbon dates, 60, 82Regional boundary problems, 133, 134,

175, 200, 250Religious communities, 176, 202-4Revenues, state, 175, 186, 201-2, 215,

216, 217, 218, 247

Salinity, 4, 5, 151-52, 213,245, 247, 251;adaptations to, 5, 20, 21, 149

Samarra, 183, 223Sasanian-Islamic continuity, 204-6Sasanian period, 5, 7, 179-85, 200-214,

230-34, 246,335Seleucia, 7, 192, 196Seleucid-Parthian period, 178-79, 180,

192-200, 228-30, 334-35Settlement: continuity and abandonment

of, 61, 81, 82, 84, 130, 132, 138-41,179-80, 184-85, 205, 218, 223, 225,325, 331-33, 336; regional systems of,63, 66-68, 70, 71, 75, 90, 155, 325,327; small, 23, 24, 26, 39, 42-44, 58,59, 63, 66, 67, 70-72, 75, 78, 84, 136,137, 178-79, 182, 188, 248, 249, 324,325, 327, 329, 330, 332

Shuruppak (modern Fara), 23, 84, 132,145

Sickles, clay, 40, 55, 61, 68-69, 78, 80,. 116, 121, 122-26, 304, 324, 325

Sippar, 3, 16, 190Sippar River. See Euphrates RiverState, growth of, 74, 76, 77, 81, 248

Stratification, social, 77-80, 245Subsistence requirements, annual, 86-

87, 146Susiana, 60, 63, 70Swamps, 1, 10, 15, 16, 21, 37, 54, 66,

141, 180, 205, 209, 213,218,252

Tempelwirtschaft, 66, 150Temples, 66, 77, 132, 190, 244, 323, 324,

327Textiles, 81, 148, 186, 202Tigris River, 1, 4, 16-18, 158-59, 197,

225; canals from, 134, 159, 190, 192,197; floods, 6, 7, 205, 218

Topography, human influences on, 19,20, 22

Trade, 80, 81, 176, 192, 199-200, 211Tribalism, 23, 153, 154, 250

'Ubaid, Tell al-, 324-27Ubaid period, 54, 55, 58, 59, 116, 302-4,

323-25Umma, 84, 134, 145, 160, 165'Uqair, Tell, 31, 59, 61Ur, 15, 132, 252, 324-34Ur III period (Third Dynasty of Ur),

144-52, 163, 164, 171, 180, 188, 194,246, 251,310, 312-13, 329-31

Ur-Eridu region (southern Sumer), 59,60, 69, 155,250,252,299-301,323-27

Urban: functions, 66, 67, 75, 80, 132-33,144, 244; hierarchies, 63, 66, 67, 70-76, 90, 136-39, 142, 178-79, 182, 217,249; rank-size and primacy, 22, 72-75, 85, 183; rivalries, 75,133, 141, 176,244; and rural differences, 137, 153,154, 215, 223, 249-50

Urbanism, 2, 76, 81, 94, 141, 248, 249,252; state-initiated, 179, 194, 199,200-201, 217, 247

Uruk (modern Warka), 2, 10, 42, 43, 59,71, 72, 80, 82, 86-87, 88, 94, 132, 190,199,244

Uruk period, 2, 16, 17, 31, 60-81, 94,116-26, 245, 251, 304, 305, 325-27

Vegetation, natural, 14, 300Village. See Settlement, smallVologasias, 192, 199

Warka. See UrukWarka Survey, 36Wasit, 33,184, 211,218Waterborne transport, 86, 142, 164, 192,

245,323Wilaya, Tell al-, 37, 88, 158-59Woolen textiles, 81, 148

362

oi.uchicago.edu

oi.uchicago.edu