amorium guide lightfoot

Amor›um

Chr›s and Mücahide L›ghtfoot

A Byzant ine C i ty in Anato l i a

homerkitabevi

ISBN 975-8293-..-.

Amorium

A Byzantine City in Anatolia

Chris and Mücahide Lightfoot

Editor

Betül Avunç

Book Design by

Sinan Turan

Prepared by

Homer Kitabevi

Printed by

Altan Matbaa Ltd.

First Published 2006

© Homer Kitabevi ve Yay›nc›l›k Ltd. fiti.

All rights reserved. No part of this book may be used or

reproduced in any manner without written permission

from the publisher except in the context of reviews.

Homer Kitabevi ve Yay›nc›l›k Ltd. fiti.

Yeni Çarfl› Cad. No: 28/A

Galatasaray 34433 ‹stanbul

Tel: (0212) 249 59 02 • (0212) 292 42 79

Faks: (0212) 251 39 62

e-mail: [email protected]

Manuscript illumination depicting the emperor Theophilus on horseback withimperial troops (Byzantine, 13th century AD).�

The Amorium Excavations Project has since its incep-tion received all sorts of invaluable help from manydifferent institutions and individuals. They are in facttoo numerous to mention individually by name here.However, the authors of this guidebook wish to thank,first and foremost, the Ministry of Tourism and Culture,the General Directorate of Cultural Heritage andMuseums in Ankara, the Directorate and Staff of theAfyonkarahisar Archaeological Museum, and theMayor and People of Emirda¤. In addition, the gen-erous and unfailing support of several sponsors thathas allowed the Project not only to continue but evento flourish over the years is to be acknowledged withthanks – they include the British Institute at Ankara,Dumbarton Oaks, the Metropolitan Museum of Art,and an anonymous American benefactor. We are alsograteful to those team members who have contributedphotographs and drawings to the present volume,while much sage advice has been received from thosefriends who kindly read the typsecript in draft form– notably His Grace Bishop John of Amorion,Professor Eric A. Ivison, Stuart Kettle, and HüseyinYaman. Mrs. Nanette Rodney Kelekian also gener-ously provided photographs of the gold solidus ofTheophilus, for which the excavations have yet to pro-vide an example. To all of these, and to the innumer-able Friends and Supporters of the AmoriumExcavations Project, the authors give their heartfeltthanks. Finally, it is fitting to acknowledge theProject’s enormous debt of friendship and kindnessto the villagers of Hisarköy, to whom this volume isdedicated with much love and affection.

New York21 October, 2005

5

Acknowledgements

�Making a squeeze of the St. Konon inscription in the Lower CityChurch in 1991.

Introduct›on 13

Pract›cal Informat›on 9

H›story and Archaeology 27

The Prehistoric Site 28Roman Amorium 32Amorium in Late Antiquity 43Amorium during the Dark Ages 48Middle Byzantine Amorium 59The Arrival of the Turks 61The Rediscovery of Amorium 65

Amorium

T›mel›ne 23

The Amor›an Dynasty 163

Contents

The Excavat›ons 81

The Lower City 82The Upper City 144

The Necropolis 154

The Archaeolog›cal S›te 69

The Upper City 72The Lower City 76

Select Bibl›ography 179

Pract›cal

Informat›on

9

�Survey work just after sunrise during the 2004 season.

Afyonkarahisar – the key to Anatolia.

After leaving Emirda¤ heading west, one first comesto the Emirda¤ bypass that serves the heavy trafficbetween Eskiflehir and Konya. From there the first vil-lage along the road is Suvermez, where one can seea number of ancient stones built into walls and build-ings – all of this material probably comes fromAmorium. In Suvermez there is a road junction, at whichone should take the right fork, leading towards Davulga.After another 6 km of undulating countryside one

10

Afyonkarahisar.

A morium isrelatively easy

to find. It isapproximately 12km from Emirda¤on the road thatleads towardsDavulga andYunak.

Amorium

Emirda¤.

comes to the small village of Hisarköy. The villagemosque is located to the right of the road; the UpperCity mound looms over the village to the left. On arrivalit is recommended that you make your presence known– villagers often congregate near the mosque. Amoriumis a protected archaeological site with an official siteguard. Ask the villagers to help you find the guard(bekçi). He will be able to provide you with informa-tion, directions, and even give you a guided tour.

It is hoped that you will enjoy your visit and willcome again to see how the excavations are progress-ing. For further information about the site and theAmorium Excavations Project, please write to the fol-lowing email addresses: [email protected] [email protected] see the website www.amoriumexcavations.org.

Contributions in aid of the Project are most welcome.Please contact the Excavation Director, Chris Lightfoot,at the Department of Greek and Roman Art, TheMetropolitan Museum of Art, 1000 Fifth Avenue, NewYork, NY10028-0198, USA.

Hisarköy.

11

Practical Information

Introduct›on

13

�Middle Byzantine terracotta jug from the Lower City Enclosure.

Detail of carving on a Middle Byzantine templon epistyle(10th century AD).

A new enlarged version that would be both moreinformative and more attractive was needed, not onlybecause there is a growing tide of visitors to the sitebut also because there is now a great deal more tobe seen and explained. In addition, a recent spate ofbooks on the Byzantine Empire and the related sub-ject of urban continuity or ‘decline’ in the Mediterra-nean world during late antiquity and the early mid-dle ages has largely failed to take Amorium intoaccount. It is hoped that scholars interested in suchsubjects may be tempted to read a full colour guide-book in their leisure hours and thereafter plan a tripto the site themselves. However, the guidebook isintended principally for a wide, general readership andso is written in a style that is hopefully both inform-ative and entertaining. Although the remains atHisarköy are not as impressive as those of the numer-ous Roman cities in Anatolia, it is Amorium’s signif-icance as a major settlement after the Roman peri-od that makes it so important. Indeed, from its out-set the principal aim of the Amorium ExcavationsProject has been to shed light on the nature of theByzantine city that survived there until the latter partof the 11th century AD. The guidebook, therefore,aims to fill in some of the gaps left by the archaeo-

14

The Lower City looking northeast towardsthe Upper City and the village.

I t is morethan ten

years since thefirst brief guideto Amoriumwas written.

Amorium

logical remains, bringing thecity and its history back to lifeagain.

Every excavation needs tohave a very good raisond’être, for the recovery ofburied remains inevitablyinvolves the destruction ofarchaeological contexts thathave been preserved overmany centuries, if not millen-nia. However careful theexcavators try to be in boththeir digging and their record-ing, there is always going tobe some loss of evidence andinformation. It is impossibleto restore an excavated site;the best that archaeologistscan hope to do is to preserveand explain the excavatedmaterial for future genera-tions. At Amorium the chal-lenge was to investigate a sitethat was once one of themost important, powerful, and presumably populouscities of the Byzantine empire. Before work began in

the late 1980s, themediaeval site was notthought worthy of thename ‘city’ – despitethe fact that Byzantineand Arab sources referto it as such. Also, ithad usually beenassumed that Byzanti-ne Amorium was quitesmall in size (that is,restricted principally tothe man-made moundor höyük) and of rela-tively short duration

Marble furnishings from the LowerCity Church (Early Byzantine, 6th

century AD).

Cast of a copper alloy coin of theemperor Leo VI (AD 886-912).

15

Introduction

(from the 7th to the mid-9th century AD). The exca-vations quickly disproved both of these assertions,for they produced an abundance of Middle Byzantinematerial -sculpture, coins, pottery, frescoes, and leadseals, to mention just the more important categories-scattered across the entire site. Not only is there goodevidence for a vigorous and wealthy community atAmorium in the 10th and 11th centuries AD, but ithas also become clear that after the Christian pop-ulation had left, Seljuk and later Ottoman Turks occu-pied some of the ruined city. It is equally significantthat the excavations have been able to demonstratethat the Byzantine city was not restricted to the Upper

It is equally significant that the excavationshave been able to demonstrate that theByzantine city was not restricted to the

Upper City mound but extendedthroughout the area enclosed by the Lower

City walls.

Fresco in the Lower City Church during conservation(11th century AD).

16

Amorium

City mound but extended throughout the areaenclosed by the Lower City walls. In other words,Amorium can be shown to have been a sizeable set-tlement, covering more than 50 hectares of habitableland, and lasting from the end of the 5th until the late11th century AD. Such a picture does not fit well withthe long-held views of many Byzantine historians, whohave typically characterised Byzantine sites assmall, easily defensible settlements and hill-top cas-

Lead seal of Nikephoros Melissenos from the Upper City (ca. AD 1065-1075).

Satellite photo of the ancient site showing the prehistoric mound.

17

Introduction

tles. Amorium may be a rare exception, but it is anexception, nonetheless. This alone would justify itsexcavation.

Similarly, the material wealth of Byzantium is usu-ally viewed in terms of imperial or ecclesiastical treas-ure and focussed heavily on the capital, Constanti-nople. Little is said about the provinces, especially inAnatolia. This is true even for the Early Byzantine peri-od. For example, one type of belt buckle known asthe ‘Syracuse’ type used to be regarded as a barbar-ian artefact, although it is now accepted that it wasof Byzantine manufacture; a specimen has been foundduring excavations at Amorium. Byzantine glasswareis another field where the evidence was meagre;Amorium has produced a wealth of material, show-ing not only that a good deal of utilitarian glass con-tinued to be produced and used throughout theByzantine period, but also that the city’s inhabitants

Copper alloy coin of theemperor Nicephorus II

(AD 963-969).

The coin finds fromAmorium constitute an

important newcontribution to the study

of Byzantinenumismatics andeconomic history.

18

Amorium

Byzantine glass bracelet fragment (11th century AD).

possessed a certain amount of luxury glass and, dur-ing the Middle Byzantine period, apparently investedquite heavily in glass jewellery. Likewise, there is theremarkable evidence of the coin finds from Amorium.To date roughly 650 coins –almost all bronze or cop-per alloy– have been recovered from the site. Mostwere excavated and so provide useful indications ofthe date for the contexts and layers in which they werefound. Some, however, were found on the surface, andmany of these were picked up by the villagers andhanded over to the excavation team. Not only was itpossible to identify a new and previously unknown typeof Middle Byzantine coin from amongst these finds,but the accumulated evidence of coins ranging in datefrom the mid-7th to early 9th century AD drew a pic-ture of economic activity at Amorium that is in markedcontrast to that at most other Byzantine sites in bothTurkey and Greece. The number of coins from the so-called Dark Ages (mid-7th to mid-9th century AD)found at Amorium argues strongly in favour of the con-

A new type of Byzantine anonymous follis found on the Upper City(ca. AD 1070-1080).

19

Introduction

tinuation of a monetary economy there. Its inhabitantswere not reduced to a system of barter and exchange,although many scholars persist in thinking that thismust have been the case throughout the provincialregions of the Byzantine empire after ca. AD 670.Finally, as at most archaeological sites, broken shardsof pottery make up the bulk of the excavated finds atAmorium. Until now it has been virtually impossibleto date much of Byzantine pottery accurately.Amorium thus provides a unique opportunity to studyand classify a large and varied assemblage of wares,covering a period of some five hundred years.

So, despite the lack of impressive standing remains–such as theatres, temples, and paved streets– andof stunning finds –such as classical sculptures ortreasures in precious metal– Amorium presents to thediscerning visitor a view of the past that is almostunique, not just in Turkey but world-wide, for it rep-

Amorium provides a unique opportunity tostudy and classify a large and varied

assemblage of wares, covering a period ofsome five hundred years.

20

Amorium

Terracotta vessels found in Trench LC (Dark Ages, early 9th century AD).

resents the best preserved remains of a mediaevalByzantine city that have come down to us. Moreover,Amorium is a city with an impressive and dramatichistory, the stage for major events that had a tellingeffect on the development of western civilisation, andone where important historical figures played out theirpart in world affairs.

The site has not been wellserved by the popular guidebooks to Turkey that are soldworldwide. In fact, only one ofthese, the Blue Guide toTurkey, makes any referenceto Amorium; despite a publi-cation date of 2001, thatbook claims that the siteremains unexcavated. Now, inaddition to this new guidededicated solely to Amorium,interested people can alsokeep up with the latest newsand developments by consult-ing the Amorium ExcavationsProject website at www.amo-riumexcavations.org.

Domestic and workshop area in the Lower City(11th century AD).

Middle Byzantine tombs in thenarthex of the Lower City

Church.

21

Introduction

T›mel›ne

23

�Drawing by Canan Gürel Ak of a Middle Byzantine colonettecapital from Amorium.

Drawing by Margaret Gill of a derelict village house inHisarköy.

Hellenistic 330-130 BCRoman 130 BC - AD 300Late antique 300-500Early Byzantine 500-650Dark Ages 650-850Middle Byzantine 850-1100Seljuk 1100-1300

333 BC Alexander the Great’s march across Anatolia.

133 BC Amorium incorporated into the new Roman province ofAsia.

ca. 133-27 BC City mint of Amorium established.

27 BC - AD 14 Reign of the emperor Augustus: first literary reference toAmorium.

211-217 Last known coins of Amorium minted.

330 Foundation of Constantinople as the capital of theRoman Empire.

431 Bishop Ablabios attends the Third Ecumenical Councilin Ephesus.

451 Bishop Mysterios attends the Fourth EcumenicalCouncil in Chalcedon.

ca. 480 Construction of Amorium’s fortifications under theemperor Zeno (?).

ca. 600 St. Theodore of Sykeon visits Amorium.

641 First Arab invasion of Anatolia.

foto laz›m

24

Amorium

Chronolog›cal Per›ods

T›mel›ne

644 First recorded Arab attack on Amorium.

668 Amorium captured by the Arabs but soon reoccupiedby the Byzantines.

691 Bishop Theodoros attends the Sixth Ecumenical Councilin Constantinople.

716 Arab attack on Amorium thwarted by Leo the Isaurian(emperor, AD 717-741).

741/42 The emperor Constantine V takes refuge in Amoriumduring the revolt of Artavasdos.

787 Bishop Theodosios attends the Seventh EcumenicalCouncil in Nicaea.

796 Arabs advance to Amorium but withdraw withoutgaining any success.

820-829 Reign of Michael II, founder of the Amorian dynasty ofemperors.

829-842 Reign of Theophilus, son of Michael II.

838 August 1-12: siege of Amorium by the caliph al-Mu‘tassim.

842-867 Reign of Michael III, son of Theophilus.

843 Re-establishment of icon veneration.

845 Execution of the Forty-Two Martyrs of Amorium atSamarra.

859 Bishop Theophilos participates in Patriarch Photios’embassy to Pope Nicholas I in Rome.

931 Amorium attacked by the Emir of Tarsus.

978 Battle of Pankaleia near Amorium.

1068 First Turkish raid into central Anatolia.

1071 Alp Arslan’s victory at the Battle of Manzikert.

1097 The First Crusade crosses central Anatolia.

1110 Konya established as the capital of the Seljuk Sultanate.

1116 The region around Amorium attested as being underSeljuk control.

1516 Mention of Hisarc›k (Amorium) in Ottoman archives.

1836 September 20: William Hamilton’s visit to Amorium.

1892 Establishment of the modern village of Hisarköy.

1987 Preliminary site survey conducted by Professor R.M.Harrison of Oxford University.

1988 Start of excavations at Amorium.

2006 Twentieth year of work by the Amorium ExcavationsProject.

25

Timeline

H›story

and

Archaeology

27

�A Phrygian doorstone decorated with a wheat sheaves, a sickle, andgrapevines, in the village of Tezköy (Roman, late 2nd-3rd century AD).

Marble head from a votive statuette (Roman, 2nd-3rd centuryAD).

The Preh›stor›c S›te

Certainly, the nucleus of the earliest settlement at thesite must have been in the area occupied by the largeman-made mound or höyük, which we now call theUpper City. This mound was probably inhabited dur-ing the time of the Hittite Empire in the 2nd millen-nium BC, but Late Bronze Age levels have not beenreached in the excavations, and no examples of thedistinctive red-burnished pottery of the Hittites haveyet been found. Despite this, Amorium has been iden-tified by some scholars with the Hittite site of Aura,and may have stood on the road taken by Hittite kingswhen they marched westwards from the capital atHattusas (Bo¤azköy) to deal with the land of Arzawaand the troublesome princes of ‘Ahhiyawa’ – that is,in all likelihood, the Mycenaean Greeks. If so, it wouldnot be the last time that Amorium would see greatkings and armies passing its way.

Although the heartland of the Hittite Empire layfurther east in north-central Anatolia, there are tracesof a significant Hittite presence in Phrygia. For exam-ple, the excavations at Dorylaeum (fiar Höyük,Eskiflehir) have in recent years provided importantnew evidence for Hittite occupation, while to the south

28

The Upper City mound looking north.

T he origins ofAmorium

may go back toearly prehistorictimes, although asyet only a handfulof pottery shardsbelonging to theEarly Bronze Age(3rd millenniumBC) have beenrecovered fromthe excavations.

Amorium

of Amorium some very striking finds have been made.In the mountains just to the south of Afyon a uniquebronze statuette, identified as that of a Hittite king,was found at Ahurhisar in 1990, and in the late 1990sa massive hieroglyphic stele turned up at Kocao¤uznear Sultanda¤. Closer to Amorium, the prehistoricmound of Tezköy Höyük, which lies only a few kilo-metres south of Emirda¤, provides good evidence forboth Hittite and Iron Age occupation. In addition, aNeo-Hittite relief, dated to the 8th century BC, wasfound during the 1970s at Daydal›, the village imme-diately to the north of Hisarköy. It has been suggest-ed that the stone may have been moved there alongwith several re-used Early and Middle Byzantinecarved blocks from Amorium.

Templon epistyle built into the village fountain at Daydal› (Middle Byzantine, 10th-11th century AD).

29

History and Archaeology

Closer to Amorium, the prehistoric moundof Tezköy Höyük, which lies only a few

kilometres south of Emirda¤, provides goodevidence for both Hittite and Iron Age

occupation.

In the Iron Age, too, evidence for occupation atAmorium is limited to a few stray shards of MiddlePhrygian painted pottery, dating to the 6th centuryBC. The lack of significant quantities of early mate-rial at Amorium may argue against it being a settle-ment of any size or importance in prehistoric and clas-sical times. However, the impressive size of the UpperCity mound –by far the largest tell in the whole region–speaks for a major prehistoric settlement at Amorium.The reason that so little trace can now be found –atleast on the surface– is better explained by the factthat later occupation of the site has almost complete-ly concealed or destroyed these levels. The excava-tions have demonstrated that Amorium remainedinhabited for a much longer period and was much

Middle Phrygian painted pottery fragment (7th century BC).

30

Amorium

The impressive size of the Upper Citymound –by far the largest tell in the whole

region– speaks for a major prehistoricsettlement at Amorium.

more intensively occupied than most archaeologicalsites in Anatolia. The mediaeval city of Amorium hasleft us very little trace even of its ancient predeces-sor. Aesop (620-560 BC), the legendary creator ofGreek fables, has long been claimed to be aPhrygian, and even now his name is sometimes asso-ciated with Amorium itself. Although these claims arespurious (Aesop was, if anything, a Thracian), the factthat the tradition probably originated as early as the4th century BC may be indicative of the emergenceand early development of the city. At the very leastit implies that Amorium was later seen as one of theprincipal settlements in eastern Phrygia, and one to

The Upper City mound looking west across the Lower City.

Honorific inscription to the high priestess Aelia Ammia, fromAmorium, now at Karayatak (Roman, 2nd-early 3rd century AD).

31


which it was appropriate to attach the name of thefamous storyteller.

Despite the lack of substantial archaeological evi-dence from the site itself, Amorium may well haveplayed a significant role in the history of centralAnatolia in Classical and Hellenistic times. The PersianRoyal Road, stretching from the Achaemenid capi-tal at Susa right across Anatolia to Sardis, probablyran to the south of the Emirda¤ mountains, therebymissing Amorium. Nevertheless, the site lay at animportant point on the central Anatolian plateau, onethat was to become a major crossroads in later times.For example, during his march across Asia Minor in333 BC, Alexander the Great must have passedthrough or very close to Amorium as he led his armyfrom Apameia (Dinar) to Gordion (Yass›höyük).Modern travellers trying to retrace Alexander’s routeacross Turkey have called in at Amorium and stayedwith the excavation team.

Roman Amor›um

Visitors to Amorium oftenremark on its attractive name,and several have assumed,incorrectly, that it is associatedwith the Latin word for love,‘amor’. In fact, the name prob-ably derives from a form of theold Indo-European root Ma,meaning ‘mother’. It suggeststhat from early times the site wasassociated with the cult of theAnatolian mother goddess.Certainly the name predatesthe establishment of the Roman

The Persian Royal Road, stretching fromthe Achaemenid capital at Susa right across

Anatolia to Sardis, probably ran to thesouth of the Emirda¤ mountains, thereby

missing Amorium.

32

Amorium

Inscription fragment from theLower City (Roman, 2nd-early

3rd century AD)

province of Asia in 133 BC, ashas been shown by the recentdiscovery of a Hellenistic inscrip-tion at Pessinus (Ball›hisar) thatmentions the existence ofAmorion in the 3rd century BC.The city’s name first appears inancient literature in the work ofthe geographer Strabo, wholived during the reign of theemperor Augustus (27 BC–AD14). Nothing else is knownabout Amorium from the surviv-ing Roman sources, and insteadwe have to rely on the archae-ological, epigraphic, and numis-matic evidence.

It is clear that Amoriumalready had a prominent positionin eastern Phrygia by the early1st century BC, for it was one of

the first cities in theregion to be grantedthe privilege of mint-ing its own coins bythe Roman Senate.It continued to pro-duce issues sporad-ically until the reignof the emperorCaracalla (AD 198-217). As yet only afew examples ofthese coins havebeen found in theexcavations, altho-ugh several of theAmorium coins in

Votive stele to the ‘Mother ofthe Mountain’ set up by

Nicomedes, from Amorium,found at Hamzahac›l› (Roman,

2nd-early 3rd century AD).

Votive stele to the ‘Mother of theMountain’ set up by Asclepiades,

found at Amorium (Roman, 2nd-early3rd century AD).

33


the Afyonkarahisar Archaeological Museum arerecorded as coming from the site. The bronze coinsof the city mint comprise a limited number of types,most of which have the ruling emperor’s portrait onthe obverse, while one of the most interesting of the

Cast of a coin minted at Amorium and found on the Upper Citymound (Roman, after 133 BC).

Cast of a coin of the emperor Commodus (AD 180-192) minted atAmorium and said to have been found at the site (courtesy of the

Anatolian Civilizations Museum, Ankara).

34

Amorium

The Roman inscriptions of Amorium arenot numerous, and most of those that havebeen found during the excavations are very

fragmentary.

reverse types shows a recliningriver god. The relevance of thisimagery may seem puzzlinguntil it is remembered that theheadwaters of the Sangarius(Sakarya) river are located with-in the city’s territory. Anotherreverse type depicts a temple,sometimes with a statue ofZeus/Jupiter or Tyche/Fortunalocated inside. Although noremains of a temple have beenfound at the site, it is likely thatthese depictions reflect some ofthe actual public buildings ofRoman Amorium.

The Roman inscriptions of Amorium are notnumerous, and most of those that have been foundduring the excavations are very fragmentary. Onlythree Latin inscriptions have been recorded so far fromthe site. One was first recorded in the 1950s whenthe name of the city (spelt AMVRI[VM]) was visibleon the stone. Nowadays the inscription is more frag-

Cast of a coin of the city mint,depicting a temple façade from

the reign of the emperorCaracalla (AD 198-217).

Inscription of Isochrysus, a member of a guild of official Roman taxcollectors at Amorium (Roman, 1st century AD).

35


mentary, but it is still a significant document, for itrecords the presence of a college of tax collectors atAmorium. They presumably collected customs dueson goods that were shipped through Amorium

between the provinces of Asiaand Galatia. Together with theRoman procurator and tax col-lectors one would expect to finda small detachment of Romantroops. So, an inscription,recorded by William Hamiltonin 1836 but now lost, indicatesthat a detachment of legio XIIFulminata was stationed atAmorium (CIL III, 353). Anotherstone, found in a derelict villagehouse in 2002, is part of aRoman milestone dating to thereign of the emperor Philip theArab (AD 244-249). Naturally,it does not belong in the cityitself, and so it must have beenbrought to the site from one ofthe main roads that led toAmorium, probably inByzantine times when suchmarkers were no longer of rel-evance to travellers.

The rest of the inscriptionsare in Greek. Most are funeraryand are written on tombstonesthat were later taken from thecity’s extensive cemeteries tohelp build the Byzantine forti-

Tax collectors presumably collectedcustoms dues on goods that were shipped

through Amorium between the provinces ofAsia and Galatia.

36

Amorium

Fragment of a milestone of theemperor Philip the Arab (AD244-249) and his wife Marcia

Octacilia.

fications of Amorium. However, two fragments of aninscribed architrave, found in the narthex of the LowerCity church, belong to a public monument and referto an emperor who can be identified as either Tiberius(AD 14-37) or Claudius (AD 41-54). In 2004 a largeinscribed block was found re-used in the foundationsof a Byzantine wall. Although fragmentary, part of allnine lines of the inscription is preserved, indicatingthat the block served as a tomb marker, set up by aman named in large letters at the top as Athena[ios].The inscription probably dates from the first half ofthe 3rd century AD. One of the finest inscriptions from

Inscribed fragment built into the Upper City walls(Roman, 2nd-3rd century AD).

Inscribed architrave fragment of the Julio-Claudian period, reused inthe narthex of the Lower City Church (Roman, ca. AD 14-54).

37


Amorium is now to be seenin the neighbouring villageof A¤›lc›k that lies just to thesouth of the main roadbetween Emirda¤ and

Suvermez. The stone survives intact, securely builtinto the base of the minaret of the village mosque;it bears a Late Roman verse inscription in elegiac cou-plets – an epitaph for a bishop called Pientios, set upby his dutiful son, Eusebios.

The inscriptionsfound at Amoriumgive an insight into

the varied social andeconomic status of

the city’sinhabitants.

Fragment of an inscribed block froma monumental tomb, set up by a mancalled Athenaios (Roman, first half of

3rd century AD).

38

Amorium

Epitaph of bishop Pientios, now at A¤›lc›k but probably fromAmorium (Late Roman, 4th-5th century AD).

Little Roman sculpture hasbeen recovered from the site.The finer pieces include part ofa marble statue of a sphinx andan almost complete but head-less crouching lion – both mayonce have decorated funerarymonuments set up by the well-to-do inhabitants of the Romancity. Likewise, there are a fewfragments of ornate marblesarcophagi known as the‘Sidonian’ type. These wereproduced at the famous quar-ries of Docimeium and export-ed not just throughout centralAnatolia but also to Rome itself. There is, however,no evidence to suggest that Amorium played a sig-nificant part in the marble trade or that a major schoolof sculptors developed there. Despite the proximityof the quarries at Docimeium (‹scehisar), the city alsodrew on a wide range of marble sources and, as wewill see, the furnishings of the major Byzantine build-ings of Amorium comprised marbles imported notonly from other areas of Anatolia but also from Greeceand elsewhere.

Torso of a white marble sphinxfrom a funerary monument

(Roman, 1st-3rd century AD).

39


Veined marble fragments from the Lower City Church(probably 5th-early 6th century AD).

As mentioned above, the Roman cemeteries ofAmorium were extensive and contained a number ofdifferent tomb types, ranging from simple rock-cutcist graves to large and impressive temple-like tombs.The majority of the inhabitants, however, settled fora grave marker that resembled a modern tombstone.It was a slab or stele, often inscribed at the top withthe name of the deceased, and usually decorated on

the front with the representationof a closed door, symbolisingthe passage of the dead fromthis world to the next. Such ste-lae are so common in theregion that they are now knownas ‘Phrygian doorstones’.Despite their apparent unifor-mity, they in fact come in awide variety of shapes, sizes,and designs. Amorium has agood selection of such stelae;

A group of rock-cut cist graves in the ancient cemetery to thesouthwest of the site (Roman, 2nd-3rd century AD).

40

Amorium

The most common typeof Roman tombstone atAmorium is known as

the “Phrygiandoorstone”, whose

design resembles themain door of a house.

A Phrygian doorstone of unusualdesign with double doors

(Roman, 2nd-3rd century AD).

many examples can still beseen in the modern village andin the Byzantine walls thatencircle the Upper City.

An important element inappreciating these tomb-stones is to note that the doorpanels often contain symbolsand objects that indicate thegender or profession of thedeceased. So, for example, awoman is represented by aspindle and ball of thread ora weaving basket and a comb,as on the tombstone of a cer-tain Kyrilla. Other doorstonesdepict fish hanging on a line.In the past this was taken toindicate the presence of acrypto-Christian community at Amorium. In fact, itis more likely that the fish is used here not as aChristian symbol but simply as an indication of a pop-ular pastime. Indeed, fish from local rivers and lakesmust have been a welcome supplement to the dietof the inhabitants of land-locked Amorium. Moreover,fish bones have been recovered during the excava-tions, showing that fish continued to be eaten atAmorium well into Byzantine times. Even today thefish restaurants at the Sakarya Bafl› complex just out-side Çifteler remain very popular with the people ofEmirda¤.

A Phrygian doorstone builtinto a village house (Roman,

1st-2nd century AD).

Murex shells from the Mediterranean coast found during theexcavations (probably Byzantine, 6th-11th century AD).

41


These scattered fragmentsare all that remain of RomanAmorium. Yet the city must havebeen furnished with all the usualpublic buildings and amenities,including temples, a forum andmarket-place, baths (presum-ably together with an aqueduct),nymphaea and gymnasia, andprobably a theatre or, at least, anodeion. No trace of any of thesebuildings or of a regular streetplan has survived. RomanAmorium, however, was a city ofsome importance and size sinceit lay on the main highway thatlinked the cities of Ionia andPamphylia with central Anatoliaand the Pontus. The road, known

from ancient maps and itineraries, ran from Apameianorth-east towards Pessinus and then on to Ancyra(Ankara), the capital of Galatia. Amorium undoubt-edly benefited from its position on or very close tothis important highway.

Sard intaglio carved with arustic scene of an old man anda rabbit hanging from a tree

(Roman, 2nd-3rd century AD).

Amorium was an important regional centrein the province of Asia, located on or near amajor crossroads on the network of Roman

roads that stretched across Asia Minor.

42

Amorium

General view of Amorium from the Hamzahac›l› road, looking southeast.

Amor›um ›n Late Ant›qu›ty

In AD 330 the emperor Constantine I (the Great)refounded Byzantium as the new capital of the RomanEmpire, naming it after himself. From Constantinopletwo major military roads ledacross the Anatolian plateau tothe eastern frontier. One took anortherly route through Ancyraand from there to the Euphratesand Armenia or to Syria andMesopotamia via the CilicianGates. The other headed furthersouth through Dorylaeum(Eskiflehir) and Iconium (Konya)to the Mut valley and the CilicianPlain. Amorium lay on thissouthern road, and although itwas essentially a highway for

Map of ancient sites and roads in Phrygia.

Bronze coin of Constantine Ifrom the mint at Alexandria,

found in the Lower CityEnclosure (AD 326-30).

43


use by the army and imperial officials, it also broughtmerchants and other travellers to Amorium. Amongthese visitors pilgrims taking the overland route to theHoly Land formed a significant group and contributednot only to the fame but also to the wealth of the city.It is likely that Amorium became an important mar-ket-town in late antiquity with regular festivals andfairs that drew large numbers of people from the sur-rounding countryside. In this respect its developmentechoes that of other inland cities, notably Ancyra,

which enjoyed a period ofeconomic prosperity and cul-tural flowering between the 4th

and the 6th centuries AD.The excavations providesome evidence for the growthand prosperity of late antiqueAmorium, for they haveshown that an extensive pro-gramme of rebuilding wascarried out at this time. Ofcourse, some of this wasnecessitated by the changesthat occurred as a result of itstransformation from a pagan

Late Roman unguentaria fragments (6th-early 7th century AD).

Marble impost fragment fromHamzahac›l› (late 5th century AD).

44

Amorium

to a Christian city. Certain features of the ancienturban landscape were no longer needed –for exam-ple, the pagan temples, the gymnasia, the theatres,and assembly halls– while the construction ofchurches became a major priority. At Amorium thelocation of four main basilica churches can now beidentified with certainty, but there were probably manyother smaller churches as well as urban monaster-ies within the city. Another important feature of a LateRoman city was its fortifications. It remains uncer-tain whether Amorium had any walls during Hellenisticand Roman times, but in late antiquity the city was

Marble beam fragments from the Lower City Church referring to anarchianagnostes, a ‘chief lector’ (5th-6th century AD).

It is likely that Amorium became animportant market-town in late antiquitywith regular festivals and fairs that drew

large numbers of people from thesurrounding countryside.

The Lower City Church in winter, showing the large ashlar blocks inthe eastern apse.

45


endowed with a massive new circuit of defences. Themost striking feature of the Lower City walls is theirlength. Although today little of their size and heightcan be seen, it still takes a considerable time to walkaround the entire circuit, following the traces that arevisible on the ground. The construction of such a largecircuit of fortifications at Amorium is in marked con-trast to what was taking place at most late antiquecities, where it was the norm for defences to surrounda reduced area at the centre of the urban settlement.

The major phase of construction for these fortifi-cations may be dated to the reign of the Isaurianemperor Zeno (AD 474-491), to whom later sourcesattribute the refounding of the city. He perhaps cameto realise the strategic importance of the city duringthe civil wars in the AD 470s and 480s. As we shallsee, both the massive fortification walls that surround-ed the Lower City and the foundation of the LowerCity church can be associated with Zeno’s reign. Onthe other hand, a separate tradition records that it wasZeno’s successor, the emperor Anastasius (AD 491-518), who rebuilt Amorium. It seems clear, howev-

The most striking feature of the Lower Citywalls is their length.

46

Amorium

View of the Lower City Walls, showing the large ashlar blocks.

er, that no major construction at Amorium can beassociated with the emperor Justinian I (AD 527-565)since Procopius was unable to record any work thereas that of his imperial patron. So, it would appear thatthe golden age of Early Byzantium passed Amoriumby, and we only hear of the city again at the very endof the 6th century AD. The occasion was the visit ofSt. Theodore of Sykeon, who attended church serv-ices with the bishop of Amorium and performed a mir-acle by healing the paralysed son of one of its lead-ing citizens. The events are described in the Life ofSt. Theodore, which was probably composed soonafter the saint’s death in AD 615. It is likely, there-fore, that the description it provides of the city is accu-rate and so sheds some light on the appearance ofthe Early Byzantine city. So, for example, the Lifementions the city walls and confirms that its popu-lation was by then largely Christian. It also describeshow the wealthy inhabitants apparently lived in largesuburban villas, complete with private chapels.

Roughly contemporary with St. Theodore’s visitis an attractive early Christian marble sarcophagusthat has an inscribed lid bearing the date AD 591/2.This was found on the southern slopes of the Upper

Topographical survey work near the Lower City Walls in 2000.

47


City mound in the 1930s and is now displayed in thegrounds of the Archaeological Museum in Afyon. Bycontrast, although the city’s cemeteries probablyremained in use throughout the Early Byzantine peri-od, there is a striking lack of Christian tombstonesof the type that is frequently found elsewhere in cen-tral Anatolia. Why this should be so remains a mys-tery.

Amor›um dur›ng the Dark Ages

With the loss of Egypt and Syria to the Arabs in themid-7th century AD, Anatolia was left as the rumpof the Byzantine Empire. Amorium benefitted dramat-ically from these changes, and as a result of its posi-tion on one of the major highways across Anatolia itassumed an even greater military and administrativerole. From ca. AD 640 the city became the militaryheadquarters of the Army of the Anatolics and, sub-sequently, the capital of the Anatolikon Theme, thelargest and most important province in ByzantineAnatolia. Throughout the so-called Byzantine DarkAges (mid-7th to mid-9th century AD), Amorium wasa key stronghold in the defence of Byzantine territo-ry against repeated Arab attacks. The first Arab attack

An early Byzantine sarcophagus, dated AD 591/2, found at Amoriumin the 1930s (courtesy of the Afyonkarahisar Archaeological

Museum).

48

Amorium

on Amorium itself is recorded ashaving taken place in AD 644,only twelve years after the deathof the Prophet Mohammed.

The Byzantine period marksthe high-point of Amorium’sexistence. First and foremost, itis a place celebrated in the his-tory of the Byzantine Empire asthe home of a short-liveddynasty of emperors – MichaelII (AD 820-829), Theophilus(AD 829-842), and Michael III(AD 842-867). It is equallyfamous as the site of a majorByzantine disaster – the siege,capture, and sack of Amoriumin AD 838 by the armies of theArab caliph al-Mu‘tassim.

Limestone screen panel, found outside the Lower City Church(6th century AD).

Gold earring, strung withemerald and pearls, found

in the Lower City Enclosure(7th centruy AD).

The first Arab attack on Amorium itself isrecorded as having taken place in AD 644,

only twelve years after the death of theProphet Mohammed.

49


Indeed, the city figures prominently in the pages ofByzantine history throughout the period between the7th and the early 12th century AD. It was at Amoriumthat Leo III (AD 717-741) held command as strate-gos (commander of the provincial army) before hewas elevated to the imperial throne in AD 717; it wasto Amorium that his son Constantine V fled for safe-ty in AD 741/2 during the revolt of Artavasdos, andit was at the battle of Pankaleia near Amorium thatthe imperial army was defeated by the rebel BardasSkleros in AD 978.

The emperor Michael II and, in all probability, hisson Theophilus were born at Amorium and, althoughit is unclear how close their association with the citywas, the establishment of the Amorian dynasty atConstantinople clearly enhanced its prestige. Michaelis portrayed by one Byzantine source as a poorly-educated man who had difficulty in reading even hisown name, although he is said to have been a goodjudge of livestock and knew when to give a wide berthto a bad-tempered mule. This description, written bya scholar at Constantinople, is clearly intended as

Two copper alloy coins of Michael II and Theophilus (AD 821-829),found during the excavations.

50

Amorium

According to one Arab geographer,Amorium ranked, after Constantinople andThessalonica, as the third largest city in the

whole of the Byzantine empire.

a slur on the emperor’s rusticorigins, but nevertheless it mayaccurately reflect Amorium’simage as a remote culturalbackwater in central Anatolia.According to other Byzantinesources, the city was regardedas lying in a pleasant and fer-tile district. Certainly, agriculturemust have formed the basis ofthe local economy, although thismay have been supplementedby direct imperial funding fromtime to time and by income gen-erated from the presence of sol-diers serving in the thematicarmy, who still received pay incash from the central govern-ment. But, as well as being avital military base, Amoriumretained a sizeable civilian pop-ulation during these troubledtimes, so that it was, according to literary sources,one of a very few cities in Anatolia to survive the DarkAges. Indeed, according to one Arab geographer, itranked, after Constantinople and Thessalonica, asthe third largest city in the whole of the Byzantineempire.

Amorium’s strategicimportance meant that itfaced repeated attack.Best known of these isthe well-documentedsiege of AD 838. In thespring of that year thecaliph al-Mu‘tassim (AD833-842), the youngestson of Harun al-Rashid,set out from Samarra inIraq at the head of a largearmy. Its banners pro-claimed Amorium as his

Plinth of an engaged column,found in the Lower CityEnclosure (Dark Ages).

Copper alloy three-ounceweight, found in the Lower CityEnclosure (Dark Ages, early 9th

century AD).

51


principal target. According to the Arab historian al-Tabari, al-Mu‘tassim ‘planned the descent uponAnqira (Ankara) carefully so that if God conqueredit for him he would go on to ‘Ammuriyya (Amorium),as there was nothing in the land of the Byzantinesgreater than these two cities, nor anything more wor-thy to be his goal.’ As the Arab armies advancedacross Anatolia, they first defeated and put to flightthe forces that the emperor Theophilus had musteredto oppose their attack. They then occupied Ankarawithout a struggle, and finally the caliph arrived withall his forces before Amorium on August 1, deter-mined to sack Theophilus’ home town. The siege,although it lasted for only twelve days, is famous and

Manuscript illustration depicting the siege of Amorium in AD 838(Byzantine, 13th century AD).

52

Amorium

has received considerable attention from both medi-aeval and modern historians.

al-Mu‘tassim’s forces, which included a large num-ber of Turks, both mercenaries and slave-soldiers,quickly surrounded the city. The Byzantine defend-ers likewise manned the walls and prepared to defendthe city, confident in their own strength (their num-bers included several detachments of the imperial fieldarmy as well as the local thematic troops) and, per-haps, in the expectation that Theophilus would cometo save his home town. One may assume that some

al-Mu‘tassim’s forces, which included alarge number of Turks, both mercenariesand slave-soldiers, quickly surrounded the

city

53


fierce exchanges took place over the next few days,resulting in heavy casualties on both sides. Certainly,later Arab sources speak as if a shrine (türbe) wasset up at Amorium, honouring the Moslems who fellduring the siege. More pertinent and, probably, moreaccurate is the reference in al-Tabari’s account of thesiege to the Arab use of engines (mangonels) to bat-ter the walls of Amorium. Time was not on al-Mu‘tas-sim’s side, for the summer was already drawing on,and he must have been fearful that Theophilus’ rout-ed troops would regroup and cut off his line of retreat.The assault was pressed vigorously and, by a strokeof luck, it could be directed against a weak point inthe defences that had been revealed by the treach-ery of a Byzantine deserter. Part of the city wall, said

to have been in a poor stateof repair, suddenly col-lapsed. Still the defendersresisted, but the tide hadturned in favour of al-Mu‘tas-sim. After a night of feastinginside the Arab camp, thefortifications were stormedon August 12. As the attack-ers entered the city, part ofthe population fled into alarge church for sanctuary.The building was set alight,and the fugitives inside wereburnt alive. Byzantinesources claim that 30,000people were either killed orcaptured during the sack.

The senior militarycommanders and other

The Arab historian al-Tabari provides themost detailed account of the siege of

Amorium in AD 838.

54

Amorium

Layer of ash and debris below theEnclosure wall (Dark Ages, early 9th

century AD).

important figures amongst the defenders were takenalive, to be used as hostages for ransom or exchange.They were separated from the remainder of theprisoners, who were marched off into captivity whenal-Mu‘tassim’s forces withdrew from Amorium soonafter its capture. Theforty-two high-rankingofficers and nobleswere subsequentlyincarcerated in adungeon at Tarsus forfive and a half years.Eventually they weretaken back to Samarra,where they wereexecuted on the banksof the Tigris for refusingto embrace Islam.Their martyrdom,which became thesubject for icons andfor a lament by thefamous 9th-centurycomposer, Joseph theHymnographer, is still

Manuscript illustration of the Forty-Two Martyrs of Amorium(Byzantine, 13th century AD).

55


Destruction layer and finds in situ in theLower City Enclosure

(Dark Ages, early 9th century AD).

celebrated in the calendar of the Orthodox Church –the Feast of the Forty-Two Martyrs of Amorium is onMarch 6. In addition to the human suffering, thecapture of Amorium caused immense damage to thefabric of the city. As well the church mentioned above,many other buildings must have been burned down,while the rest of the city was apparently pillaged.Literary sources suggest that it was not only captivesthat the Arabs carried off from Amorium – valuablesof all sorts were collected up as booty; these includedbooks and some doors (presumably highly ornate

Extensive ash layers from burning in the Lower City Enclosure (Dark Ages, early 9th century AD).

56

Amorium

In the 9th centuryAD Arab powerwas on the wane,

while that ofByzantium was

growing.Silver coin (miliaresion) of

Theophilus, stray find from theLower City (AD 830/1-c. 838).

ones made of bronze)that were later recordedas being displayed atthe Syrian city ofRaqqa.

As a military eventthe capture ofAmorium is of no greatsignificance, for itappears that the caliphwas not set on theconquest of Byzantineterritory and made noattempt to consolidatehis position atAmorium. In the 9th

century AD Arabpower was on thewane, while that ofByzantium was grow-ing. al-Mu‘tassim hadpressing reasons toreturn home, for rebel-lion had broken outthere in his absence.For the Byzantines thesack of Amorium was thus only a temporary setback,even though it involved the destruction of much prop-erty and considerable loss of life. It does seem, how-ever, that the city’s destruction had a profound affecton Theophilus himself, leaving him a sad and bro-ken man. The siege of AD 838 can also be regard-ed as a major turning-point in Byzantine history, forthe city’s fall marked the final humiliation of icono-clasm and led directly to the restoration of the ven-eration of holy icons in AD 843. In this sense it hashad a long-lasting effect on the development of theGreek Church, the history of Orthodox Christianity,and the traditions of eastern European art.

Equally important, especially from the stand-pointof the modern Republic of Turkey, is the fact that thesiege marks the first occasion on which the presence

Detail of an icon depicting theemperor Michael III with his mother

Theodora in imperial regalia(Byzantine, ca. AD 1400).

57


of Turks in Anatolia is clear-ly recorded. As mentionedabove, the caliph’s armywas made up largely ofTurkish soldiers; two of hissenior generals were alsoTurks, known as Ashinas andItakh. So, although it was tobe another 230 years beforethe Turks returned toAmorium and centralAnatolia in their own right,Turkish participation in thehistory of their future home-land can be traced back asfar as the first half of the 9th

century AD. It is also worthnoting that while a 13th-cen-tury Byzantine manuscriptillustration depicting thesiege may be anachronistic

in naming the attackers as ‘Turkoi’, the work has for-tuitously preserved a detail of historical reality.

All of the sources speak of the sack of Amoriumas inflicting massive destruction on the fabric of thecity, including the razing of its fortifications by al-Mu‘tassim’s troops. It seemed that for some time after

Detail of manuscript illustrationdepicting Turkish troops at the siege

of Amorium in AD 838 (Byzantine, 13th century AD).

58

Amorium

The Byzantine fortifications of Ankara Castle, looking north.

---

AD 838 the military headquarters of the AnatolikonTheme were located at Polybotus, near modernBolvadin, to the south of the Emirda¤ mountains. Theexcavations seem to confirm the literary evidence,for they have revealed little trace of occupation atAmorium during the second half of the 9th centuryAD. It clearly took some considerable time for thecity to recover, and it is surprising that the imperialauthorities apparently did little to initiate its immediatereconstruction and to relieve the suffering of thesurviving population. This situation is in markedcontrast to that of Ankara, which had also beencaptured in AD 838. Fragments of an inscription arestill visible in the walls of the inner citadel of AnkaraCastle that record the rebuilding of the fortificationsat Michael III’s command in AD 862. This lack ofimperial interest in Amorium has misled some modernscholars into assuming that the city never recoveredfrom the Arab sack. Archaeology has, however, beenable to fill the gap left by the historical sources.

M›ddle Byzant›ne Amor›um

After AD 838 Amorium is rarely mentioned in the sur-viving Byzantine texts. There was another attack onAmorium in AD 931, when troops of the Emir ofTarsus unexpectedly descended on the city, but thiswas an isolated incident. By the late 9th century AD

Inscription of Michael III built into the inner walls of Ankara Castle(AD 862).

All of the sources speak of the sack ofAmorium as inflicting massive destruction

on the fabric of the city, including therazing of its fortifications by al-Mu‘tassim’s

troops.

59


the Byzantines had begun to take the offensive, sothat military events no longer took place on theAnatolian plateau but on its southern and easternflanks. As part of this process Amorium regainedsome its former military and strategic importance, butinstead of being a bulwark in the defence ofByzantine territory it became a mustering- and sup-ply-point for the campaigns that were to push the fron-

Middle Byzantine housing built up against the rear of a survivingpart of the Lower City Walls, looking southeast.

Marble floor slabs in the nave of the Lower City Church (MiddleByzantine, late 9th - early 10th century AD).

After AD 838 Amorium is rarely mentionedin the surviving Byzantine texts.

60

Amorium

tier forward into eastern Anatolia, Syria, andMesopotamia.

Results from the excavations have shown thatprosperity did indeed return to Amorium in the 10th

and 11th centuries AD. New housing was construct-ed on the outskirts of the city, money was lavishedon ecclesiastical buildings, and local industries werere-established. This situation persisted until the arrivalof the Seljuk Turks. The first raids deep intoAnatolia occurred in AD 1068-1069 and took theByzantines completely by surprise. They included anattack on Amorium, indicating that the city was thenstill regarded as a valuable prize.

The Arr›val of the Turks

After Alp Arslan’s defeat of the emperor Romanus IVat the battle of Manzikert in AD 1071 a large part ofAnatolia was swiftly overrun by the Seljuks. Ankara,for example, fell to them in AD 1073, and by AD 1080Nicaea (‹znik) had been chosen as their first capital.But during the First Crusade the Turks were pushedback from the Sea of Marmara and suffered a seri-ous defeat in the plains around Dorylaeum(Eskiflehir). Consequently, Konya became the cap-ital of the Seljuk Sultanate in AD 1110. Amorium nowlay in a frontier zone between the remainder of the

A small tannery workshop with drainage sink (Middle Byzantine, 11th century AD).

61


Byzantine Empire to the north and west and theSeljuks to the south and east. These were not con-ditions favourable to the continued prosperity of thecity, and it is likely that most of the Byzantine inhab-itants of Amorium fled (or, indeed, were removed bythe imperial authorities) to more secure areas in west-ern Anatolia not long after AD 1071.

In addition, the local population had to contendwith the arrival of Norman and Frankish crusaders.The Norman warlord Roussel of Bailleul attemptedto found an independent state in the heart of Anatoliain ca. AD 1073, and the fighting that ensued betweenRoussel’s forces and the Byzantines centred aroundAmorium. According to Sir Steven Runciman, theroute taken by the First Crusade in AD 1097 led fromDorylaeum through Polybotus on route to PisidianAntioch (Yalvaç). He argued that the terrain aroundAmorium was ‘a long waterless tract of country’ and‘suitable only for swiftly moving cavalry.’ Hence it is

Final occupation layer with human and animal remains within theformer latrine of the Lower City bathhouse

(Middle Byzantine, late 11th century).

62

Amorium

The local population had to contend withthe arrival of Norman and Frankish

crusaders.

usually assumed that the First Crusade did not passthrough Amorium. It is, however, inherently more like-ly that the Crusaders followed the main Byzantinehighway that must have led to Amorium. Likewise,the German emperor, Conrad III, used the oldByzantine highway between Dorylaeum, Amorium,and Philomelium (Aksaray)in AD 1147.

The region of Amorium isknown to have been in Seljukhands by AD 1116, when theemperor Alexius I led theByzantine army on an expe-dition from Dorylaeum in thedirection of Polybotus andPhilomelium. But the factthat the lands both to thenorth and south of theEmirda¤ mountains remaineda war zone for most of the12th century AD probablymeans that most of the set-tled Byzantine populationwould have been encour-aged to leave, while theTurkish tribesmen who

A Roman doorstone recently uncovered in the northern cemetery,looking south towards the Upper City mound.

Byzantine spolia in the Seljuk tombat Hanköy.

63


replaced them were by and large tent-dwellingnomads. Certainly, the archaeological evidence sug-gests that parts of the site were reoccupied by theSeljuks, but this may not have occurred in any sub-stantial form until the 13th century AD. Likewise, datedSeljuk buildings in the Afyon region are known onlyfrom ca. AD 1210 onwards, and the fact that thereis little evidence for the conversion of existingByzantine buildings into mosques, medreses, or hansafter the Turkish occupation suggests a prolongedperiod of instability and dislocation.

At Amorium a break in occupation may also beindicated by the fact that its ancient name disappears,even though it had long been familiar to Arab histo-rians and geographers. The last reference to the site

After the fall of the Byzantine city in thelate 11th century the ancient name ofAmorium was forgotten by the local

inhabitants, suggesting that there was aradical change in population.

64

Amorium

Drawing of the village school and mosque by Margaret Gill in 1990.

as Ammuriye is found in thetravel writings of a 12th-cen-tury Syrian author, al-Harawi. Naturally, the namecontinues to be found inByzantine sources, but thelatest historical referenceoccurs in the work ofGeorge Pachymeres (AD1242-c.1310). It wouldappear that the local popu-lation retained no memoryof the ancient name, so thatby the 16th century the siteis recorded in official docu-ments under the name ofHisarc›k. By then the areahad come under Ottomancontrol, and the excavationshave shown that a small Turkish settlement, centredon the Upper City mound, persisted there well intothe 18th century. However, this too disappeared, sincethe present-day village of Hisarköy was only estab-lished in 1892 when refugees from the Balkans weresettled there.

The Red›scovery of Amor›um

Although the existence of an ancient city calledAmorium was known to European antiquarians andtravellers by the middle of the 18th century, the ruinsat Hisarköy were only positively identified as thoseof Amorium in 1836. In September of that year aBritish scholar called William Hamilton visited the siteand concluded from the surviving remains and fromits general location that it matched Amorium. It isclear that the standing ruins were then much moreprominent then than they are now, and Hamiltonnoted several inscriptions that can no longer be foundat the site. However, in general he seems to havebeen rather unimpressed by Amorium, describing itas ‘the deserted and dreary site of what was once apopulous city; and seldom have I witnessed a more

Traditional bread baking atHisarköy in 1997.

65


striking scene of solitude and desolation...’ Sadly, thesite was destined to lie neglected and almost forgot-ten for the next century and a half.

The eminent Russian Byzantinist A.A. Vassilievapparently paid a visit to Amorium at the turn of thecentury, but British and German epigraphers in the1920s and 1930s found little there to interest them.It was left largely to Cyril Mango, formerly Professorof Byzantine Studies at the University of Oxford, toarouse scholarly interest in its potential as an archae-ological site. He made his first trip to Amorium overforty years ago and so began a slow but eventuallypersuasive campaign to encourage colleagues toundertake its excavation. Fieldwork, however, onlybegan in 1987 under the direction of Professor R.M.Harrison. A full excavation permit was issued in thefollowing year, and work has continued at the site eversince, directed first by Martin Harrison (until hisuntimely death in 1992), and subsequently by ChrisLightfoot.

General view of the modern village of Hisarköy.

66

Amorium

In the 19th and early 20th century fewscholars showed any interest in Amorium,despite its importance in the pages of the

history of the Byzantine empire

The excavations have provided much informationabout the ancient and late antique city, as well as evi-dence for the later Seljuk and Ottoman occupationof the site. It is, however, the Byzantine material thatmakes Amorium unique. The archaeology ofByzantine Anatolia is still very poorly known. AtAmorium it is possible to investigate the nature of anentire site from its former state as a Roman city,through its survival in the Dark Ages, to its transfor-mation into a Middle Byzantine town. It represents arare opportunity to trace urban continuity from the6th to the 11th century AD, andthe results from Amorium havealready been able to refutesome widely-held views onByzantine social and economichistory. Likewise, the excava-tions have produced a wealth offinds, some of which are withoutparallel elsewhere, while othersprovide an archaeological con-text for material that was other-wise known only in museum col-lections.

The following chapters willexplain the salient features of theexcavations and highlight someof the finds.

Martin Harrison photographed at the Lower City gateway in 1992.

67


Glazed pottery fragments foundduring the excavations on the

Upper City (13th century).

The

Archaeolog›cal S›te

69

�Part of the Early Byzantine ambo parapet reused in the floor of thenorth aisle of the Lower City Church.

View of the Upper City mound from the Hamzahac›l› road,looking southeast.

The Upper City mound has an oval shape, covering anarea of some 5 hectares and standing some 20 metresabove the surrounding ground level. The Lower Cityencompasses a much larger area, extending around thesouth and east sides of the mound. The whole of the sitethus measures about 1,130 metres from north-east tosouth-west and about 700 metres from north-west tosouth-east, enclosing approximately 70 hectares. Onlypart of this area is now occupied by the modern villageof Hisarköy.

In addition, a wide area across the stream to the north-west of the Upper City appears to have been occupied inthe Roman period, while there are a number of quarries,rock-cut tombs and other structures along the sides of thesmall ravine to the south-west where the modern asphaltroad leads into the village from Emirda¤. Above the ravineto the south a low tumulus (burial mound) is visible onthe crest of the hill; this is usually the first indication ofthe ancient city to greet visitors as they approach Hisarköy.The ancient cemeteries of Amorium must have beenextensive to judge from the large number of Roman tomb-

70

Fragments of Roman and Byzantinespolia built into the walls of the Great

Mosque in Emirda¤.

T he ancient sitecomprises

two main sectors,namely an Upperand a Lower City.On the north sideof the modernvillage is theUpper City, theman-made moundor höyük thatforms the nucleusof the site.

Amorium

stones that have been found both at the site and in thesurrounding villages. Tombs have, for example, been notedas far away as the ridge of hills to the south-east wherethe modern road climbs towards Davulga.

The ‘fallen buildings’, so briefly and tantalisinglydescribed by Hamilton in 1836, have suffered consider-ably over the last century and a half, and little trace ofthe ancient and mediaeval city now remains above ground.The ruins of Amorium have long been used by the vil-lagers as a quarry for dressed stone. The houses, mosque,and cemetery of Hisarköy itself, as do those of many othervillages in the area surrounding Amorium, attest to thewholesale recent plundering of the site. So, many of theolder houses and buildings in the neighbouring villagesof Hamzahac›l›, Pörnek, Suvermez, and Karayatak, as wellas in Emirda¤ itself, containstones taken from Amorium.They include a number ofimportant inscriptions andarchitectural fragments. Thisstone-robbing has meantthat, before the excavationscommenced, the rubble coreof a small stretch of the for-tification wall on the north-east slope of the Upper Citymound was the only ruinthat remained standing

View of the ravine with ancient graves and tombs, looking NEtowards the Upper City mound.

71

The Archaeological Site

Phrygian doorstone reused as afountain in Hamzahac›l›

(Roman, 1st-3rd century AD).

above ground. Information provided by the village elders,however, makes it clear that stretches of the fortificationwalls and several other large structures in the Lower Citystill survived to some considerable height within livingmemory. The villagers can also remember when and wherevarious individual stones were dug up and what subse-quently happened to them. The excavations have beenable to verify several of their statements about the site.

The Upper C›ty

The best views of this large and impressive mound or tellcan be obtained from the crest of the hills to the east onthe road to Davulga. From this view-point it seems to dom-inate the whole site, but when seen from the southern sec-tor of the Lower City its true position, lying in hollowbetween two streams and flanked by higher ground, ismore apparent. Indeed, as one approaches the site fromEmirda¤, the Upper City is completely hidden from viewuntil one enters the modern village, and so most first-timevisitors are taken by surprise when the south-western cor-ner of the mound suddenly comes into view.

Standing masonry on the northeast slope of the Upper City mound(possibly Seljuk, 13th century AD).

The Upper City mound probably formedthe nucleus of the settlement at Amoriumright through from prehistoric times until

the Seljuk period.

72

Amorium

In Byzantine times the Upper City was completelyenclosed by a fortification wall, traces of which are visi-ble around the edge of the mound. The circuit wall wasstrengthened by a number of projecting square towers,and there appear to have been several gates. At the north-eastern and south-western corners of the mound the UpperCity wall joins up with the circuit of fortifications aroundthe Lower City. The mound was the subject of an inten-sive surface survey during the 1989 and 1990 field sea-sons. This immediately produced some unexpectedresults; for example, it revealed that the site had not beenentirely abandoned when the Byzantine city of Amoriumceased to exist in the late 11th century AD, as scholars

View of Amorium from the Davulga road, looking northwest.

Remains of the fortification wall (possibly early Byzantine, 7th

century AD) around the Upper City.

73


had previously assumed. Surface finds clearly indicateda Turkish presence, and the outlines of some buildingswere also recognised as the remains of a substantial Seljukoccupation. In the south-west corner of the Upper City asmall enclosure was tentatively identified as a Seljuk fort,perhaps the origin of the Turkish place name Hisarc›k(meaning ‘Little Fort’), while the surviving stretch ofmasonry wall below the north-east tip of the mound isunlike any of the excavated Byzantine fortification wallsat Amorium and so probably also belongs to the Turkishperiod.

The undulating surface of the Upper City indicates thepresence of numerous square and rectangular structures,which excavations have shown to belong in their final stagesto the Ottoman period. The lower courses of the walls ofthese buildings comprise large masonry blocks (spolia)

Line of part of the walls of the small fort in the south-west corner ofthe Upper City (probably Seljuk, 13th century AD).

74

Amorium

Survey in the Upper City has revealedsurface remains of numerous buildings,including those of the apse of a large

basilica.

from the ancient city. Several such stones are strewn acrossthe mound, while others are still visible built into theByzantine fortifications around its edge. Finally, the apseof a church can be traced on the surface near the easternedge of the mound. This building has not yet been inves-tigated by the excavationteam, although a geophysi-cal survey was undertaken in2001. The large pit at thenorth-eastern end of thechurch bears witness tosome of the stone-robbingand treasure-hunting activi-ties that plagued Amoriumbefore it became a protect-ed archaeological site.

The mound is not inhab-ited nowadays and is usedmerely for grazing sheep,although the village womenare also accustomed to con-gregate along its southernedge on summer evenings inorder to take advantage ofthe pleasant views over thevillage.

View of Roman tombstones built into the Upper City walls.

75


Geophysical survey of the church on theUpper City mound in 2001.

The Lower C›ty

The Lower City was also surrounded by a massive for-tification wall, the line of which can still be traced run-ning around the site. Only a small section of thesedefences has been excavated, revealing not only part ofthe curtain wall but also one of the city gates and a largetriangular tower. Other gates and towers presumably existat intervals around the circuit, but these at present can

Aerial view of part of the Lower City, showing outlines of buildingsand streets.

76

Amorium

only be guessed at. The length of this circuit of wallsprompts questions about how and by whom they weremanned – imperial troops, soldiers from the provincialarmy, or militia recruited from the city’s inhabitants. Butwhoever manned the fortifications, it would have required

Much of the Lower City has remainedunoccupied since Byzantine times and sotraces of the mediaeval city’s streets and

buildings are visible even today just belowthe ground surface.

77


a sizeable force to defend them effectively against attack.The future emperor Leo III is said to have placed a gar-rison of only 800 men in Amorium in AD 716, and yetthis seems wholly inadequate to protect the whole of theLower City in the face of a concerted Arab attack. As aresult, perhaps, it has been argued that by the early 8th

century AD the Lower City walls had already been aban-doned and that only the Upper City was fortified. This the-ory, however, runs counter to much of the archaeologi-cal evidence. In addition, while no great reliance can beplaced on the numbers given by the sources for the sizeof the civilian population gathered inside Amorium at thetime of the siege in AD 838, the fortress was clearlydefended by a substantial force, comprising not onlytroops of the Anatolikon Theme, under the command oftheir general (strategos), Aëtius, but also three of the fourdivisions (tagmata) of the imperial field army. It maytherefore be assumed that these troops were sufficient-ly numerous to have adequately manned the circuit ofLower City walls.

A disused well in a courtyard of one of the village houses.

Another important feature of the site is thenumber of wells that are scattered across

the area within the Lower Cityfortifications.

78

Amorium

Within the walls the Lower City can be divided into threedistinct areas:- (a) the modern village, (b) a number offields that are still under cultivation in the eastern and south-eastern sectors, and (c) a large stretch of open land run-ning across the southern part of the site from the Davulgaroad to the edge of the ravine to the south-west. A simi-lar area used merely for grazing is located to the north-east of the Lower City Church near the centre of the wholesite. These uninhabited areas reveal numerous traces ofbuildings and streets; the existence of several large pub-lic buildings can also be postulated from the lines of wallsand fragments of masonry that are visible on the surface.So, for example, it was possible to identify some remainsclose to the modern road through the village as thosebelonging to a church long before excavation of the struc-ture began in 1990. Another important feature of the siteis the number of wells that are scattered across the areawithin the Lower City fortifications. Although these wellsnow supply all the water needs of the villagers and of theexcavation team, theirexistence probably datesback to early mediaevaltimes. Although inRoman times Amoriummay have been furnishedwith an aqueduct thatsupplied running waterfor the public baths andnymphaea, the Byzantineinhabitants needed amore secure source ofwater, especially duringthe troubled times of theArab invasions. It wasfortunate for them that anample supply of waterlay immediately belowtheir feet; indeed, thismay well be one reasonfor the continued exis-tence of the city in theByzantine Dark Ages.

A Byzantine well head excavated in theLower City Enclosure

(probably 11th century AD).

79


The Excavat›ons

81

�Top view of one of the multi-handled vessels (Dark Ages, early 9th

century AD).

The largest surviving piece of fresco, depicting a female saint,on the south wall of the Lower City Church (MiddleByzantine, 11th century AD).

82

The Lower C›ty

The Church

During work that extended for some nine seasons themain body of the church was exposed – the apse atthe eastern end, the central nave and aisles, and thenarthex at the western end. Two principal buildingphases have been identified.

View of the Lower City Church in 1998, looking east.

Aerial view of the Lower City Churchduring excavation in 1994.

T he excavation ofthis ecclesiastical

building began in 1990and has revealed themost impressivemonument that cannow been seen atAmorium. The churchmust have formed partof a larger complex,flanked on either sideby subsidiary buildingsand having an atrium infront of its west end.

Amorium

The original plan of the church (Phase I) was thatof a basilica, divided into a nave and two side aisles.The main body of the church was entered from anarthex at the western end through three doorways;the grey and red-veined marble jambs of the door tothe south aisle are still standing to a considerableheight. At the east end of the nave is an apse nearly8 metres in diameter, the exterior of which compris-es three sides of a regular hexagon. Within the apsethe synthronon, the stone seating for the priests, hasbeen uncovered. The altar would have stood in front

Plan of the Lower City Church in 2004, showing the main twophases of construction.

83

The Excavations

Phase I

Phase II1

29

141211

10 9

7 3

3456

23

24 46 44 43 42

25

26 52 37

49 39

4047

1922

2

8 28

27

A6

A5

A4 A1

A1

Narthex

Naos

N

Bema

North aisle South aisle

A2

A3

A9

A8

A7

15

13

58

55

32

34

66

16

17

18

of the synthronon. The walls are built of large dressedand squared blocks and stand in places to a height of3 metres above the floor of the nave. The lower wallsurfaces were originally revetted with marble. Crampholes can be seen in the walls; some even retain tracesof the iron cramps that held the large revetment slabsin place. Two rows of columns flanked the chancel(bema) and nave (naos), while a chancel screen (tem-plon) or iconostasis, separated the sacred eastern endof the church from the nave. The floor in the chanceland nave was paved with different coloured marbleslabs in the opus sectile style, forming an attractivegeometric design.

It is likely that the Phase I aisled basilica had anupper storey, surmounted by a timber-framed roof, butthis is difficult to prove because of the extensive natureof the later reconstruction of the building. Likewise,the date of the construction of the church remainsimprecise, although the archaeological evidence sug-gests that it should be placed in the late 5th or early

The synthronon at the east end of the Lower City Church.

The floor in the chancel and nave waspaved with different coloured marble slabs

in the opus sectile style, forming anattractive geometric design.

84

Amorium

6th century AD. Such adate would fit well withthe architectural plan ofthe church and with thestyle of the survivingfragments of EarlyByzantine furnishings.Two large and impres-sive marble capitals,now displayed in thegarden of the AfyonMuseum, are said tohave been recovered from the site of the church manyyears ago; another has also made its way to Akflehir.Small fragments from a similar pilaster capital havebeen found during the excavations in the church, con-firming the provenance of these pieces. Another impor-tant find from the church was a marble column plinthwith an 11-line inscription mentioning the martyr saintKonon. Since the plinth had later been cut down toform an octagonal base, it is uncertain whether itbelongs to the original church, but it at least revealsan intriguing link with St. Konon, who was much ven-erated in Isauria and especially by the Isaurian emper-or Zeno (AD 474-491).

In all probability this building is the episcopal church(or cathedral) where St. Theodore of Sykeon attended

Part of the marble opus sectile floor in the nave in the Lower CityChurch (Middle Byzantine, late 9th - early 10th century AD).

One of the two ornate capitals ondisplay at the Afyonkarahisar

Archaeological Museum (EarlyByzantine, 6th century AD).

85

The Excavations

services with the city’sbishop at the end of the6th century AD. If so,there is a good chancethat it could be thechurch that wasreportedly burnt to theground by the Arabs in

AD 838. Certainly, the central location of the churchwithin the city suggests that it was from the start oneof the more important ecclesiastical buildings atAmorium – and it remained so until the very end ofthe Byzantine occupation of the site. Thearchaeological investigation of the church has revealedthat the building has enjoyed mixed fortunes,accurately reflecting those of Amorium itself, duringthe 1,500 years of its existence.

First of all, the aisled basilica suffered severe firedamage, so severe in fact that much of the fabric ofthe church was damaged beyond repair. The wood-en roof would have been the most vulnerable part ofthe building in a fire, and the blackened scorch marksthat can be seen today on the Phase I walls (especial-ly in the central bay of the northern aisle) may indi-cate where roof timbers collapsed and lay burninginside the building. The marble furnishings from theinterior of the church also apparently suffered consid-

Inscribed plinth mentioning St. Konon(Early Byzantine, late 5th-6th century AD).

The Lower CityChurch has

produced a wealthof carved marble

and limestonefurnishings of theEarly and MiddleByzantine periods.

86

Amorium

erable damage. Many small carved fragments, as wellas pieces of marble wall revetment, were later reusedas packing material in the Phase II walls. Likewise, partof the parapet belonging to the late antique pulpit(ambo) was found reused asa paving slab in one corner ofthe north aisle, and at leastone column, perhaps from thenave, was later carved downand made into the epistyle forthe chancel screen. The opussectile floor in the chancel andnave must also have beenbadly damaged, for it wassubsequently replaced by anentirely new floor, whichraised the height of floor sur-faces in the church. Traces ofthis earlier floor were foundduring the excavations whereparts the Phase II marblefloor slabs had been removedduring the Seljuk use of thebuilding. By contrast, the

Signs of burning on the main north wall of the Lower City Church, andshowing dowel holes for the Early Byzantine marble revetment.

87

The Excavations

Part of the Early Byzantine amboparapet reused in the of the northaisle of the Lower City Church.

stone synthronon seems to have been left relativelyundamaged by the destruction of the Phase I church,presumably because it was largely protected under themasonry semi-dome of the apse.

No irrefutable evidence has been found for whenthis disaster occurred. The rebuilding of the church,however, clearly took place no earlier than the late 9thor early 10th century AD, so the damage must haveoccurred before then. The most obvious and conven-ient historical event with which to associate the destruc-tion of the Phase I aisled basilica is, of course, the siegeand sack of Amorium in AD 838. Churches, howev-er, are also vulnerable to natural or accidental disas-ters, especially fires caused by the use of flame-litlamps and burning candles. Nevertheless, it is hard toavoid the conclusion that the church must have suf-

Bird’s-eye view of the surviving remains of the Middle Byzantinefloor in the bema, with the site of the altar at its centre.

88

Amorium

Large quantities of loose tesserae have beenrecovered from the collapsed deposits insidethe church, indicating that the semi-domein the apse, the central dome, and otherparts of the ceiling vaults were decorated

with brightly-coloured glass mosaics.

fered during the Arab attack, and the event that causedsuch massive damage to the church can be linked withdestruction layers that have been found in other exca-vated areas of the Lower City – these deposits of ashand burnt debris have produced several coins datingto the reign of the emperor Theophilus (AD 829-843).Such evidence, however, is lacking from within thechurch because all the debris was cleared away beforework on its reconstruction started.

Phase II saw a major restructuring of the wholechurch; in place of the columns along the nave fourlarge piers were erected in the centre of the building,and numerous smaller piers and buttresses were addedto the walls. It seems very likely that the piers sup-ported a central dome over the pulpit in the nave, whilethe rest of the building was probably covered withvaulted ceilings. Large quantities of loose tesserae havebeen recovered from the collapsed deposits inside thechurch, indicating that the semi-dome in the apse, thecentral dome, and other parts of the ceiling vaults weredecorated with brightly-coloured glass mosaics. Thewalls of the church were also covered with wall paint-ings at this time. Although little of these now remainin situ, it is clear from the surviving fragments that asecond and, in places, even a third layer of fresco wasapplied to the walls. The best preserved section of

painting is to befound on the

m a i nwall in

t h e

89

The Excavations

A fragment of Middle Byzantine glass mosaic, fallen from one of thedomes or vaults of the Lower City Church.

western bay of the south aisle. Although only the lowerhalf of the figure survives, it clearly depicts a stand-ing woman, dressed in a long robe with ornamentedborders imitating gold embroidery with jewels. Her richcostume indicates that she must have been a personof higher rank; she may be identified as either St.Barbara or St. Catherine of Alexandria.

The multiple layers of painted plaster show that aconsiderable time must have elapsed between thereconstruction of the church and its abandonment.Indeed, it is clear that the Phase II building underwenta number of repairs and alterations during its lifetime.Rich patrons, both church men and lay people, con-tinued to lavish attention (and, presumably, donations)on the church well into the 11th century AD. Recentresearch and scientific analysis of the excavation findsindicate that the Middle Byzantine church was not onlygenerously endowed but was also spectacularly

Three layers of Middle Byzantine fresco still attached to a piece ofmasonry (late 9th-11th century AD).

Although the surviving frescoes are veryfragmentary, they tell us a great deal about

the Middle Byzantine decoration of thePhase II domed basilica.

90

Amorium

adorned with a dazzling array of colour.In addition to the opus sectile floor,which included a roundel of glassmosaic tesserae representing afloral wreath, the walls werecovered with frescoes ofsaints and prophets, andliturgical furnishings suchas the chancel screen werebrightly painted to empha-sise the geometric designscarved on their surfaces.Above, the ceiling glis-tened with gold-glassmosaics. It would have pro-vided an appropriate settingfor the metropolitan archbish-op of Amorium and his clergy,dressed in fine silk vestments, to per-form their services. These would prob-ably have been accompa-nied by an assortment ofchurch plate in gilt silver,painted icons, and illumi-

The largest surviving piece of fresco, depicting a female saint, on the southwall of the Lower City Church (Middle Byzantine, 11th century AD).

Part of a glass mosaic roundel in situ inthe floor of the nave of the Lower CityChurch (Middle Byzantine, late 9th-11th

century AD).

91

The Excavations

nated manuscripts, while the whole church would havebeen lit with glass hanging lamps and sunlight shin-ing in through glazed windows, which consisted of plas-

ter frames set with panes ofcoloured glass.

The Middle Byzantinechurch, however, was not justa place for the living but alsofor the dead. During the 1998and 2002 excavation seasonsa total of nine tombs were

Photo and drawing by Elizabeth Hendrix of polychrome decorationon limestone cornice blocks (Middle Byzantine, late 9th-early 10th

century AD).

92

Amorium

Tombs were notarranged in a neat

row, side by side, butdug down into the lateantique foundations ofthe church wherever

space could be found.Copper alloy bracket for a hanging

lamp, found in the Lower CityChurch (Byzantine, early 6th

century AD).

uncovered in the narthex at the western end of thechurch. They were not arranged in a neat row, side byside, but dug down into the late antique foundationsof the church wherever space could be found. Theyvary, too, in size and type, which suggests that theywere not planned as a group but were constructedindividually as the need arose. It is clear, however, thatthey were all inserted into the narthex floor after thereconstruction of the Phase II church had beencompleted.

Only one of the tombs – the one that was locatedagainst the north wall of the narthex – had subsequent-ly been disturbed during the Seljuk reoccupation of thechurch. The remaining eight tombs, containing some34 burials, were found to be still intact. Several had

Middle Byzantine tombs in the narthex of the Lower City Church(late 9th-11th century AD).

93

The Excavations

seen reuse during the course of the Middle Byzantineperiod; Tomb 6, for example, had four occupants,placed one on top of the other. Only Tomb 4, whichoccupies the central and therefore most prestigiousposition in front of the steps into the nave of the church,does not seem to have been reused. It contained onlytwo bodies, identified as a man and a woman, both inearly middle age (about 35-45 years old).

It is usual to find clerics buried within the sacredprecincts of a church. Here in the Lower City Church,however, it seems that wealthy and influential fami-lies were allotted space in the narthex. Grave goodsare scarce in Byzantine burials. These tombs producedonly a hen’s egg from one tomb (Tomb 4) and an ironbracelet from another (Tomb 5). In Tomb 5, howev-er, were also found five copper alloy coins, all mint-ed during the reign of the emperor Nicephorus II Phocas(AD 963-969). This numismatic evidence providesuseful confirmation for the date of the tombs, but thecoins were not the most interesting or spectacular findsthat were discovered in the tombs. These comprise twogroups of material – one made of solid stone, the otherof delicate organic remains.

At the western end of Tomb 3 the fragment of aMiddle Byzantine closure slab had been used as part

Tomb 6 in the narthex of the Lower City Church during the removalof delicate organic material in 2002.

94

Amorium

of the tomb structure. In a similar but more lavish man-ner, Tomb 4 was lined with reused architectural ele-ments, including two Early Byzantine Ionic impost cap-itals, while its lid comprised three more MiddleByzantine slabs. Close inspection of these slabs sug-gested that originally they had probably served as para-pet slabs that enclosed the upper part of the pulpit,

Tombs 5 and 6 in the narthex of the Lower City Church, afterexcavation.

Tomb 4 at the centre of the narthex in the Lower City Church, afteropening in 2002.

Tombs found in the narthex of the LowerCity Church were used for members of the

most important families in MiddleByzantine Amorium.

95

The Excavations

or ambo, in the nave. Whythis pulpit should have beendismantled and reused asbuilding material remains amystery, but, again, it showsthat there was much activi-ty in the church during the

Middle Byzantine period.The second group of very remarkable and rare finds

comprises fragments of textiles and leather shoes. Thetextiles, both clothing and at least in one case (in Tomb6) a shroud, appear to have been made of silk andwere probably richly decorated and embroidered. InTomb 4 fragments of textile containing gold threadwere also preserved. The remains of shoes were foundin four of the tombs (Tombs 4, 6, 7, and 8). The shoeswere stitched together from several pieces of leather,and had flat soles. At least some were apparently ofthe slip-on variety without laces and reached up to theankle, where they could be folded down. Similar shoesare represented in wall paintings of the 10th-11th cen-turies AD in Cappadocia.

The tombs in thenarthex were found

to contain silk textilesand leather shoes as

well as severalsuccessive burials.

Detail of a Middle Byzantine amboslab reused as the top of Tomb 4.

96

Amorium

Fragment of a cuff of silk cloth with gold thread from Tomb 4(Middle Byzantine, late 10th-11th century AD).

The lavish refurbishment ofthe Lower City Church is all themore impressive when it isremembered that most MiddleByzantine churches are consid-erably smaller than those builtduring the Late Roman andEarly Byzantine periods. Evenwhen a Middle Byzantinechurch reused the core of anold building, it most frequent-ly only occupied a small part ofthe former space. Well-travelledvisitors might recall the smallMiddle Byzantine church builtwithin the late antique basilicaat Side on the south coast ofTurkey. This is not the casewith the Amorium church. Herethe whole of the main body ofthe church was reoccupied,making it one of the largest Middle Byzantine church-es known. Indeed, its overall floor space is greater thanthat of many contemporary churches in Constantinopleitself.

Tomb 6, after opening in 2002,showing the leather shoes at the

lower end of the stonesarcophagus.

97

The Excavations

General view of the naos and bema of the Lower City Church, looking south.

In recent years much effort has been directed bythe excavation team to recording and conserving thestanding remains. This work continues to this day. Inaddition, new excavations were begun in the 2005 sea-son immediately to the northwest of the church. Here,rather unexpectedly, a baptistery with an adjoiningnarthex were found. The baptistery is square in plan,with a single, western doorway, recessed rectangu-lar niches in the south and north walls, and a semi-circular apse at the east end. The western doorwayleads to a rectangular narthex, aligned north-south,that is entered at its southern end from the narthexof the church. The baptistery and its narthex, there-fore, constitute ancillary buildings that could only beentered from the main church. This new building isextremely well preserved and sheds further light onthe history of the entire ecclesiastical complex. It isalso one of very few examples of Early Byzantine bap-tisteries to have survived and provides important newinformation on baptismal liturgy in the early OrthodoxChurch.

General view of the baptistery of the Lower City Church, looking east.

98

Amorium

The baptistery walls match those of thePhase I aisled basilica church, indicating

that they were all built at the same time aspart of a major new construction.

The walls of the baptistery and its narthex share thesame construction as the Phase I aisled basilica, withlarge ashlar blocks for the lower courses and coursesof smaller blocks above. Four columns, of which onegray marble base survives in situ, were arranged in asquare around the centre of the baptistery. Thesecolumns had shafts of grey, red-veined marble, and werecapped with richly decorated Corinthian capitals of theso-called ‘Theodosian’ type, carved from fine, whitemarble. Several fragments of such capitals were foundin the baptistery during the course of its excavation; they

Western doorway and surviving column base in the floor of the baptistery.

Marble column shaft and other architectural fragments foundabandoned in the baptistery.

99

The Excavations

match two intact specimens that are now displayed inthe Afyonkarahisar Archaeological Museum. These ele-gant capitals are said to have been found during ille-gal excavations at the church in the 1960s. Thesecolumns may have supported a central dome with anencircling vaulted ambulatory. The recessed niches onthe north and south sides must have been covered byarches forming arcosolia, and a half dome probably cov-ered the apse. Like the rest of the church, the walls ofbaptistery were originally sheathed in marble revetment,but only a few fragments now remain in place, notablyat the bottom of the wall in the eastern apse. It is alsoclear that the floors in the baptistery and its narthex werepaved with large rectangular slabs of marble. Thesewere robbed in the post-Byzantine period, leaving onlytheir but can be reconstructed from impressions in thesurviving sections of mortar bedding.

The baptistery was richly decorated withmarble revetment on the walls and largepaving slabs of marble on the floor in a

similar way to the main body of the EarlyByzantine church.

100

Amorium

Eastern apse of the baptistery.

Set into the pavement at the centre of the baptis-tery and framed by the four columns was a cruciformfont, with four steps in the eastern and western armsdescending into the central basin. In the northern andsouthern arms a broader step set at a higher levelenabled the priest to stand above the baptismal can-didate who knelt in the basin below. Water was prob-ably supplied from a source just outside the baptis-tery via a terracotta water pipe running under the floorfrom the southeast wall into the font. The drain wasprobably located in the lower basin but could not beexcavated.

The baptistery clearly did not suffer as much dam-age as the main body of the church during the greatfire that overwhelmed that building, probably in thesack of Amorium in AD 838. Nevertheless, during therestoration of the entire church complex in the late 9th-early 10th centuries AD the baptistery and its narthexunderwent certain alterations. The most radicalchange was the stripping and backfilling of the EarlyByzantine font with loose debris and earth. A large frag-ment of masonry was dumped into the central basin,presumably to support the floor and a new font above.

General view of the Early Byzantine sunken font in the centre of thebaptistery (6th century AD).

101

The Excavations

The site of the font was paved over with a square ofnew opus sectile pavement that closely resembles thatfound in the bema and naos of the main church. Thereason for the abandonment of the sunken font canbe explained by changes in liturgy and custom. In the6th century AD adults were still being baptised, andthe ritual involved emersion in holy water. By theMiddle Byzantine period, however, people were beingbaptised as babies, and water was sprinkled on themfrom a standing font. No trace of a new font was foundin the baptistery, but slab fragments and a basin foundin the main church may have belonged to such aninstallation.

Three Middle Byzantine tombs were also found. Onehad been cut through the pavement on the north sideof the baptistery, and a further two tombs were set inthe floor at the north end of the baptistery narthex. Allthree tombs had been looted and completely emptiedduring later times, but in construction and generalappearance they resemble those found in the adjacentchurch narthex, and so must date to the 10th-11th cen-turies. The baptistery tombs, like the ones in the churchnarthex, must have been reserved for persons of highstatus.

Brick rubble used to fill the font before a new Middle Byzantinefloor was laid in the centre of the baptistery.

102

Amorium

In the post-Byzantine period the baptistery sufferedthe same fate as the main church; the marble pave-ments, revetment, and furnishings were systematical-ly stripped away. A wall of spolia blocks was construct-ed on the western threshold into the baptistery, thusnarrowing the doorway. In the narthex floor just in frontof the doorway a circular bread oven or tand›r was dis-covered, built of mud brick and once capped by a bee-hive-like superstructure. Domestic refuse, includingcooking wares similar to those found in the mainchurch, was found in the area around the oven, togeth-er with a 13th-century Islamic coin. So, one can assumethat during the Seljuk period the baptistery complexserved as a work area connected to the main church,which had been converted into a farmhouse and barns

The two tombs in the baptistery narthex during excavation in 2005.

The baptistery was refurbished during theMiddle Byzantine period when the central

sunken font was filled in. Three tombs wereinserted into the floor, but they were foundto be empty when they were excavated in

2005.

103

The Excavations

for domesticated animals.However, a silver denier mint-ed at Lyon (central France) inthe first half of the 11th centu-ry was also found in the debris– the coin may attest to thepresence of pilgrims from west-ern Europe in Amorium in thedecades preceding the FirstCrusade. Further disturbancewas caused by a robber pit thatwas dug into the western side ofthe baptistery in the 1960s.These illicit activities were con-firmed by the discovery of a col-umn shaft buried in the rubblewithin the baptistery bearing agraffiti with the name of one ofthe villagers carved on it.

The Lower City Walls

The investigation of a stretch of the Lower City Wallswas a focus of attention from the very beginning ofthe Amorium Excavations Project. It was intended toclarify the date of their construction, to find out whetherthey were built as new defences or were a reworkingof older structures, and to ascertain, if possible, whenand how they were destroyed or abandoned. In addi-tion, it was hoped that excavation on the periphery ofthe urban settlement area would shed some light onthe extent of Byzantine occupation, especially as it hasbe argued in the past that the Byzantine city wasrestricted to the Upper City mound.

The excavations in the area designated as TrenchAB on the southwest perimeter of the Lower City haverevealed an impressive stretch of curtain wall whose

Detail of one of the tombs in thebaptistery narthex.

104

Amorium

The wall is pierced by a gateway, flanked tothe right by a large triangular tower,

originally standing several storeys high.

construction was clearly a major, probably imperial,enterprise. The lower part of the wall is built of mas-sive, finely-cut ashlar blocks (see p. 46), while abovethe wall comprises neat rows of smaller, roughly-dressed blockwork. These fortifications would appearto have been erected during peace-time when time,care, and new materials could be utilised in the work.They do not contain any re-used stones (spolia) fromthe Roman city or its cemeteries; this is in marked con-trast to the fortifications on the Upper City mound. Interms of masonry construction, the Amorium tower andcircuit wall would fit a date range between the mid-5th and mid-6th century AD. The style of constructionmatches that of Phase I of the Lower City Church, andso it may be postulated that the aisled basilica and thecity walls were built at roughly the same time. It fur-ther suggests that Amorium underwent a massive pro-gramme of urban renewal in late antiquity.

The wall is pierced by a gateway, flanked to theright by a large triangular tower, originally standingseveral storeys high. The tower is an integral part ofthe defences, intended to protect the vulnerable gate-way, and the existence of another matching tower must

The Lower City fortification wall and gateway, looking southeast.

105

The Excavations

be postulated to the other side of the gate. Since tracesof few other towers are visible in the lower circuit atAmorium it is impossible to say whether the presenceof a triangular tower (or pair of towers) is unusual.Certainly projecting rectangular towers are the normin the Late Roman and Early Byzantine periods,although instances do occur elsewhere in the EastMediterranean, most notably at Thessalonica, wherean excavated example is dated to the mid-5th centu-ry AD.

A secondary defensive feature was located duringthe excavation of the exterior of the triangular tower.Initially this feature was viewed as a massive buttressto the tower, but two small trenches along its presumedcourse soon allowed this to be recognised as an addi-tional defensive wall, probably designed to create anenclosed courtyard in front of the gateway. In effect

Another peculiar feature of the Lower CityWalls is that there are remains of a secondwall immediately to the rear of the curtain,

effectively doubling the width of thefortifications to over 6 metres.

106

Amorium

Middle Byzantine houses built on the foundations of the Lower Citywalls (late 10th-11th century AD).

one can envisage a pair of pincer-like walls extend-ing from the flanking triangular gate-towers. The peri-od between the erection of main curtain and the outerwall may have been quite short given the use of sim-ilar ashlar masonry. But why such an addition shouldbe felt necessary remains a mystery. Was it in responseto a perceived threat to the city? Or was it merely anextra embellishment to the fortifications?

Another peculiar feature of the Lower City Walls isthat there are remains of a second wall immediatelyto the rear of the curtain, effectively doubling the widthof the fortifications to over 6 metres. It appears thatthis strengthening of the circuit wall is confined to thestretch behind the triangular towers to either side ofthe gateway. Since the rear face of the inner wall tothe north of the gateway is constructed of large ash-lar blocks similar to the outer facing of the curtain tothe south of the gate, it is likely that the two widthsof wall form parts of the same construction, built piece-meal by different gangs of workmen. The full width ofthe walls would thus allow room for stairs to lead fromground level up to the parapet walk, which was pre-sumably at some considerable height.

Aerial view of the Lower City Walls, showing the triangular towerand fore wall in front of the gateway (late 5th-early 9th century AD).

107

The Excavations

By contrast, there does not appear to have beenany provision at Amorium for a separate fore-wall orproteichisma, which is a fairly common feature of LateRoman and Early Byzantine urban defences. Likewise,an entrance from the road outside into the courtyardin front of the gate has not been located. The road-way through the gate itself is relatively narrow, meas-uring only 4 metres wide, and it does not appear tohave been paved. This is rather surprising given themonumental nature of the gate. Outside the city, onthe other hand, traces of a road – with cuttings andwheel ruts – can be seen on an outcrop of rock thatlies a short distance to the west of the gateway. It mayonly represent the course of the old (Ottoman)Emirda¤-Davulga road before the asphalt highway wasconstructed through the modern village, but it wouldbe a strange coincidence if this carriageway passedso close to one of the city’s gates purely by chance.

During the excavation of the interior of the trian-gular tower in 1993, a deep layer of debris, presum-ably from the upper floors and roof of the tower, was

encountered. It consisted of amassive jumble of stone blocks,bricks, and loose rubble, togeth-er with a number of charredwooden beams. Dendrochrono-logists from Cornell Universityhave been able to identify thewood used for some of thesebeams as oak. The sampleswere subsequently fitted into amaster chronology for oak treerings, enabling the tree fromwhich the wood was cut to begiven a life-span starting in AD362 and ending some time afterAD 487. It is, therefore, temptingto associate this data with the lit-erary evidence ascribing the for-tification of Amorium to theemperor Zeno. This destructionlayer can still be seen in the sec-

Burnt timbers mixed with stoneand brick debris inside the

triangular tower.

108

Amorium

tion that runs from the back wall to the apex of thetower. Below the mass of collapsed building materialwere found thinner layers of debris, containing a mix-ture of material – broken pottery, glass, metalwork(including several arrow- and spear-heads), as well asa complicated iron mechanism (probably from a doorlatch, but perhaps from a catapult or other defensivesiege weapon), and even textile fragments (apparent-ly discarded rags, rather than pieces of clothing).Interestingly, no human remains were found, so itseems that all the occupants of the tower managedto escape before it collapsed. Below all this debris wasa compact earth floor strewn with straw, the upper sur-face of which had been badly scorched and charred.The archaeological deposits, therefore, clearly indicatethat the tower had been destroyed as the result of a

Dendrochronologists from Cornell University studying the burnttimbers from the triangular tower in 1993.

109

The Excavations

The interior of the triangular tower wasfilled with a massive layer of collapsedtimbers and masonry – evidence for the

violent destruction of the Lower City walls.

catastrophic fire, but can a date be assigned to the vio-lent end to the Lower City Walls on the strength of thisevidence?

Only a few coins were found inside the tower, thelatest of which is a copper alloy follis of the emperorTheophilus (AD 829-842). Although it is unwise to puttoo much emphasis on a single coin, it provides thefirst clue for the likely date of the tower’s destruction.Samples of ash taken in 1992 from the upper layerswithin the triangular tower have also provided aCarbon-14 date of ca. AD 800. So, the scientific evi-dence indicates that part, at least, of the tower’s con-struction can be dated no earlier than the end of the5th century AD, and that its destruction occurred inthe late 8th or early 9th century AD. It would seem

Iron javelin and spear points found during the excavations (probablyDark Ages, late 7th-early 9th century AD).

110

Amorium

Only a few coins were found inside thetower, the latest of which is a copper alloyfollis of the emperor Theophilus (AD 829-

842).

likely that this catastrophe took place in the summerof AD 838 when the Turkish soldiers under the com-mand of the Abbasid caliph al-Mu‘tassim stormedAmorium. Such a conclusion has significant implica-tions for our understanding of the size and nature ofthe Byzantine city during the Dark Ages. It suggeststhat the defenders of Amorium did not retreat insidethe walls of the Upper City during the siege but hadthe strength of numbers to man the full circuit of wallsaround the Lower City. This also presupposes that thefortifications had been sufficiently well maintained (and,presumably, repaired on occasion) since their con-struction to provide a viable defence against attack.

After the Lower City defences had apparently lainin ruins for some time, a series of rooms was built overthe levelled foundations of the double wall immediate-ly behind the triangular tower. A small hoard of twen-ty-two copper alloy anonymous folles and one goldcoin, a scyphate histamenon that dates to the reignof Constantine X Ducas (AD 1059-1067), was foundin one of these rooms in 1990. The discovery that thearea of the city walls was re-used for housing was tosome extent unexpected, since it had previously beenassumed that in the Middle Byzantine period most ofthe settlement was concentrated on the Upper City.However, this evidence, combined with the signs ofcontemporary activity in other excavation zones of theLower City, indicates that Amorium had gradually

111

The Excavations

Part of the small hoard of 22 coins found in the Middle Byzantinehouses built on top of the Lower City Walls (ca. AD 1065-1070).

recovered from the Arab sack so that by the mid-11th

century AD it stretched as far as the periphery of thepre-838 settlement. This area of domestic occupationin the Lower City, however, does not seem to havebeen protected by new walls, and so at least part ofthe population of Amorium lived beyond the shelterof any fortifications. The archaeological evidence thusfits well with the general political and military situa-tion during the brief period of revival in Byzantium’s(and thus also of Amorium’s) fortunes from the latterpart of the 9th to the middle of the 11th century AD.

The re-occupation of the area of the Lower Citywalls was destined to be relatively short-lived. Thehoard of coins from the house overlying the olddefences provides an important clue to the suddendemise of Amorium as a result of the Seljuk conquestof Anatolia after the victory at Manzikert in AD 1071.After the Byzantine settlement had been abandoned,

Sketch drawing of ruins as seen by William Hamilton in 1836.

A small hoard of coins was found in aroom built over the ruined Lower City

walls – it had been hidden there during thelast years of the city’s existence.

112

Amorium

there was no further occupation of this part of the site.Such an assertion is supported by the absence ofdemonstrably later pottery from the whole area ofTrench AB. Instead, during the past 100 years theByzantine ruins have been used as a ready source ofquarried stone. Sadly, the relatively recent stone-rob-bing activity has both confused the archaeological pic-ture and seriously degraded the surviving structures.

In order to clarify the nature of the MiddleByzantine occupation it was decided to enlarge the areaof excavation immediately behind the walls and gate-way. Work in this area, designated as Trench LC, wascarried out between 1994 and 1996. The first stagewas to excavate the masonry debris within the gate-way itself. At the outer end the lintel blocks wereexposed still lying where they had fallen across theentrance. A coin found on the surface layer immedi-ately below the collapse indicates that the gatewayitself was still standing until the last quarter of the 10thcentury AD. The horizontal lintel consisted of interlock-ing blocks, recalling Hamilton’s description of a stand-ing arch ‘on the northern slope of the Acropolis’, inwhich he remarked on the ‘peculiar formation or dove-

Interlocking lintel blocks fallen from above the gateway in the LowerCity Walls (late 5th-early 9th century AD).

113

The Excavations

tailing of the key-stone.’ [see drawing on p. 112]Strangely, however, no evidence has been found fora threshold or hinge-posts for any doors to the gate.If, therefore, the gateway was not left permanentlyopen, which seems unlikely in the face of repeatedArab attacks during the 7th and 8th centuries AD, wemust assume that there was some form of portcullisthat was dropped from above to bar intruders.

The excavations have also exposed a large areaimmediately behind the city wall and to the north ofthe gateway, revealing a series of rooms and struc-tures that was evidently built in this area after the for-tifications had fallen into disrepair and had been aban-doned. Numerous walls, still in a good state of preser-vation, have been uncovered, forming a building com-plex with a large central room and a courtyard. Theroom backed onto the partially preserved city wall andhad two entrances, one leading into the courtyard, theother into a smaller room to the east. To the north-west lies another group of smaller interconnectingrooms and corridors. Finds from within these roomswere sparse, as if they had been emptied before theywere abandoned, suggesting to some extent an order-ly evacuation rather than a sudden flight. All of thecoins found in this area were anonymous or signedfolles (copper alloy coins), the standard form of loosechange used in Middle Byzantine times, that had beendropped and lost haphazardly during the occupationof the area.

Immediately to the northwest of the MiddleByzantine housing, another room was partially exca-vated in 1996, while a second chamber was locatedin 1998 between it and the city wall. Unlike the otherareas, these two rooms produced a wealth of materi-al within a substantial ash layer. The most remarkablefinds comprised a group of pottery vessels that had

A group of very unusual terracotta vesselswas found behind the Lower City walls inthe destruction layer associated with the

siege of AD 838.

114

Amorium

been smashed and then left in the burning debris. Atleast half a dozen of the vessels belonged to the samepeculiar type, with only slight variations in size andshape. They all have a flat base and bulbous body,tapering upward towards either an everted rim-like topor a more unusual mask-like opening. Seven or eightvertical curved handles are arranged in two rowsaround the middle and upper part of the body. The topsand bases of the vessels are pierced with a series ofholes that are connected through the body of the ves-sel by a cylindrical compartment. This is sealed offfrom the rest of the interior of the body, and so thevessel resembles something like a samovar in con-struction. It is unlikely, however, that these pots servedto heat liquids, although ‘self-heaters’ (authepsae forwater or, more likely, spiced wine) are known from

Broken pottery in situ in destruction layer behind the Lower Citywalls (Dark Ages, early 9th century AD).

115

The Excavations

Group of restored pottery vessels from the destruction layer behindthe Lower City Walls (Dark Ages, early 9th century AD).

Roman and Early Byzantine times. It remains uncer-tain what the exact function of the Amorium finds wasand what they were intended to contain. Nevertheless,these extremely odd vessels are not unique toAmorium – two other examples are known, one in theIstanbul Archaeological Museum, the other inKastamonu.

The layer of debris in which the pottery was foundalso contained the remains of charred timbers, seedsand other food stuffs, more pieces of textile, and twoearly 9th-century AD coins – one of the emperorNicephorus I (AD 802-811), the other of Theophilus.The debris included a mass of mud-brick, some ofwhich still retained the shape of individual bricks. Itwas clear, therefore, that the room had once had mud-brick walls, built on stone foundations and footings,

Top view of three of the multi-handled vessels (Dark Ages, early 9th century AD).

116

Amorium

It is unlikely that these strange pots servedto heat liquids, although ‘self-heaters’

(authepsae for water or, more likely, spicedwine) are known from Roman and Early

Byzantine times.

and it was the collapse of these walls that had buriedand so preserved such a rich assemblage of materi-al. Mud brick has been the traditional building mate-rial par excellence in Anatolia for millennia, so it is notsurprising to find evidence for it in Byzantine Amorium.Indeed, it has been found, associated with similardestruction layers, in other parts of the site. Moreover,the presence of mud brick should enable us to distin-guish to some extent between public and private con-struction. So, while fortifications, churches, and othermajor buildings were entirely built of stone and firedbrick, it is likely that much of the minor structures–houses, shops, and small workshops– would havebeen built with mud-brick walls on stone footings.However, it is noticeable that the Middle Byzantinestructures that adjoin and partially overlie the LowerCity walls seem to have more substantial masonrywalls, and one reason for this may be that with thedestruction of the fortifications there was plenty of stoneat hand with which to build.

Stone foundations and lower courses of walls belonging to mudbrick buildings (Dark Ages, 7th-early 9th century AD).

117

The Excavations

The Enclosure

To the north of the Lower City Church and locatedalmost at the very centre of the whole site lies an areathat has been designated as the Enclosure. Beforeexcavations began here in 1996 the most prominentfeature was an earth- and grass-covered bank that sur-rounded the area on all four sides, giving theEnclosure a trapezoidal shape. Inside the ground slopesdown quite steeply from south-east to north-west.Elderly villagers reported finding ‘large walls’ and ‘mar-ble floors’ within the Enclosure during illicit excava-tions in the 1950s and 1960s, and signs of robbertrenches are still visible, especially along the easternside of the Enclosure.

At the centre of the Lower City is an areaof some 12,370 sq. metres surrounded by asubstantial defensive wall, built during the

Middle Byzantine revival of Amorium.

118

Amorium

View of the Lower City Enclosure, looking south from the Upper City mound.

The investigation of the Enclosure has been themain focus of the Excavation Project in the years since1998. Initially efforts concentrated on the bank, andexcavations soon revealed that it contained the remainsof a massive stone wall. From the amount of loose rub-ble that was found to either side of the wall, it was clearnot only that the wall was of considerable width butalso that it originally rose to a substantial height. Todaya stretch of some 25 metres of the southern perime-ter wall of the Enclosure has been exposed; in placesit still stands to a height of 3.5 metres. The wall is facedwith blockwork arranged in irregular horizontal rows,and has a mixed core of rubble, brick, and mortar. Inplaces both the facing and the core contain spolia –re-used architectural fragments from earlier buildings.Another indication of the relatively late date for thewall’s construction is that it overlies other features andbuildings. This dating also seems to have receivedadditional confirmation from the discovery in 2001 ofa group of six copper alloy coins in the rubble core ofthe wall; the coins, all of the same class of anonymous

Exterior face of the Enclosure wall, looking west (Middle Byzantine, late 10th-11th century AD).

119

The Excavations

folles, date to the last quarter of the 10th or the firstquarter of the 11th century AD.

Despite its size and evident strength, no towers orgates have been found along the exposed length ofthe Enclosure’s south wall. Moreover, until 2003 nosubstantial buildings were associated directly with it.So the exact nature and purpose of the wall remainunclear, although it must have been defensive or pro-tective in some way. It is, however, now apparent thatthe wall was not an isolated construction, for recentexcavations in the south-east corner of the Enclosurehave revealed a series of rooms clustered around theinner face of the wall. These rooms, too, contain re-used material – one low barrier included an attractive

closure slab, decoratedwith a cross, and steps anddoorway thresholds madeuse of conveniently-sizedcolumn capitals, while theinterior walls were found toinclude blocks still deco-rated with scraps of wallpainting, presumablytaken from a larger fresco-covered building. Likewise,part of an open courtyardwas uncovered; this waspaved with bricks and a

Interior face of the Enclosure wall, looking south (Middle Byzantine, late 10th-11th century AD).

Middle Byzantine buildings in thecorner of the Enclosure during

excavation in 2003 (11th century AD).

120

Amorium

large circular grinding stone. The rooms that flankedthe courtyard and were built up against the Enclosurewall appear to have been living quarters, but the dearthof artefacts from within the rooms makes it difficultto speculate on the exact nature of their occupancy.The corner room contained a central raised brickhearth and adjoining fire pit, and so may have beenused as a kitchen, while in the room immediately toits west the remains of along-shafted iron javelinwere found. Whether theserooms should be seen aspart of a barrack buildingor a secure storage areabehind the Enclosure wallremains uncertain.

However, it is clear thatthere was considerableactivity both within andimmediately outside theEnclosure during theMiddle Byzantine period. A

Middle Byzantine closure slab reused as part of a later Byzantinebuilding in the Lower City Enclosure (11th century AD) .

So far only a small gateway providingaccess to the Enclosure has been found in

the west wall.

Courtyard floor with step to doorwayinto the corner room at top (Middle

Byzantine, 11th century AD).

121

The Excavations

number of other structures of a residential and com-mercial nature have been excavated, some of whichhad been inserted into the ruins of earlier buildings.Although most of the buildings are small and poorlyconstructed, occupation is surprisingly dense, indicat-ing that this part of the Lower City retained a signifi-cant role in the economic and social life of theByzantine city in the 10th and 11th centuries AD. Coinfinds appear to suggest that much of this resurgencemay be associated with the reign of the emperorNicephorus II Phocas (AD 963-969). The purpose towhich the Enclosure was put has not as yet been fullyascertained, but it was obviously not a normal resi-dential area, and it remains an attractive hypothesis

Although most of the buildings are smalland poorly constructed, occupation is

surprisingly dense, indicating that this partof the Lower City retained a significant role

in the economic and social life of theByzantine city in the 10th and 11th

centuries AD.

122

Amorium

A complex of small rooms, probably workshops and storerooms tothe east of the bathhouse in the Lower City Enclosure (Middle

Byzantine, late 10th-11th century AD) .

that it was built to serve some specifi-cally military role. Perhaps it servedas a mustering-point and supply-base for the Middle Byzantinearmy during the offensivecampaigns in south-eastAnatolia and northernSyria.

Below the MiddleByzantine levels nu-merous buildings andwidespread destructi-on layers have beenfound. In large part the-se can be dated to theByzantine Dark Ages andgive a good indication of thelevel of occupancy and activity

in the centre of thecity during the

7th, 8th, andearly 9th centu-ries AD. The

excavations have thus provided us with a glimp-se of what life was like in an urban

community after the end of an-tiquity. Most ancient cities in

Anatolia had by then decli-ned into small and rathersqualid settlements, oftendefined as ‘squatter occu-pation;’ others had beenreduced to small pockets

of occupation in defen-sible sites such as hill-top forts. The eviden-ce from the Enclosu-re area shows thatAmorium does not fit

either of these traditio-nal views of Byzantine Anato-lia. Public buildings continu-

123

The Excavations

Glazed terracotta lamp from the Lower CityEnclosure (Middle Byzantine).

Terracotta lamp found in the LowerCity Enclosure in 2003 (Early

Byzantine, late 6th-early 7th centuryAD).

ed to function, streets we-re maintained, and newconstruction was under-taken on a sizeable sca-le, even if many of thenew buildings were nowmade of mud-brick onstone foundations. At thesame time it is clear fromthe excavations in theEnclosure that food pro-cessing and storage we-re now regularly carriedout within the city. In2005, for example, two

In 2005, two wellpreserved wine-

pressinginstallations werefound in different

parts of excavationarea.

Treading floor of a wine press found inthe Lower City Enclosure in 2005(probably Dark Ages, 7th-early 9th

century AD).

A deep, plaster-lined tank, originally part of a wine press, filled withcarbonised grain and debris, Lower City Enclosure

(Dark Ages, early 9th century AD).

124

Amorium

well preserved wine-pressing installations were foundin different parts of excavation area. One, near the so-uth-west corner of the Enclosure, preserves the who-le of the pressing floor with a central drain hole thatleads via a stone spout into a smaller collecting well.The other next to the street at the centre of the southside of the Enclosure was converted in its latest pha-se into a storage silo, and large quantities of carbo-nised grain were found inside it below a destructionlayer.

The street itself was another important find in the2005 season, showing that some elements of the lay-out of the Late Roman and Early Byzantine city werepreserved in the subsequent Dark Ages. Despite thefact that the surviving road surface is unpaved (it ismade up of compacted earth mixed with a generousscatter of broken bone, pottery, and marble), the streetclearly remained a major thoroughfare running in astraight line across the centre of the Lower City. It wasburied under a layer of ash and collapsed debris, prob-ably part of the general destruction caused by the sackof Amorium in AD 838.

In addition to the street, two major structures arevisible in the excavated area today. One is a rectan-

Unpaved Byzantine street found in the Lower City Enclosure in 2005(Dark Ages, 7th-early 9th century AD).

125

The Excavations

gular building (Structure 2) that extends under theEnclosure wall so that its southern end wall can betraced immediately outside the Enclosure. The pre-cise date and function of the building remain uncle-ar, although it was obviously built sometime beforethe Enclosure wall itself. One part of the interior wasfound to have hydraulic plaster on its walls and sto-ne-slab floor, while a large stone screw press weight,carved with a cross on its upper surface, was exca-vated near the centre of the building. Structure 2 may

Buildings inside the area of the Lower City Enclosure, includingStructure 2 incorporating a stone screw press weight in one of its

interior division walls.

126

Amorium

Some buildings within the Enclosure wereused for a long time but others were

apparently demolished and their buildingmaterials incorporated into new structures.

thus have served origi-nally as a winery.The other buil-ding is nowcompletelyp ro tec tedby a mo-dern roofso that itis difficultto see pre-cisely itsshape andsize, and thenumber of ro-oms it contai-ned. However,this building (Struc-tures 1 and 3) isone of the most in-teresting and im-portant public buil-dings so far unco-vered at Amorium,and as such it deserves its own heading.

Structure 2 in the Lower City Enclosure with stone screw pressweight (Dark Ages or Middle Byzantine, 8th-11th century AD).

127

The Excavations

Base of a terracotta wine cup inscribedwith the name of a monk called Thomas

(Middle Byzantine, 10th-11th centuryAD).

The BathhouseStructures 1 and 3 were excavated between 1998 and2002. The former is essentially a rectangular block,divided into six rooms, while the latter is a single largehall. Together they form a complex built as a singleunit, probably surrounded originally by an open spaceor courtyard. The exterior walls are composed of smallmasonry blockwork, interspersed with brick courses– a technique that is typical of Late Roman and EarlyByzantine construction. Further clues to the date ofthe complex were provided by the plan and design ofthe large hall (Structure 3). This was polygonal inshape; its exterior walls are laid out on the basis of ithaving eighteen sides, although the actual foundationsappear to be circular. The interior walls are divided intosix semicircular apses, three of which were pierced withdoorways. The ones to the northeast and southwest

Excavation of Byzantine structures to the east of the bathhouse in 2002.

128

Amorium

Polygonal buildings are a common featureof Late Roman and Early Byzantine

architecture in the East, where they serveda number of different purposes – as

mausolea or martyria, baptisteries, and thelike.

led to the outside of the building, while the southerndoor gave access to the rectangular building (Structure1). Polygonal buildings are a common feature of LateRoman and Early Byzantine architecture in the East,where they served a number of different purposes –as mausolea or martyria, baptisteries, and the like.Here, however, the hall must have functioned as agrand entrance hall or changing room that led to thebathing rooms in Structure 1.

At the centre of the polygonal hall there is a circu-lar stone stylobate, which originally supported six mar-

Plan of the bathhouse complex.

129

The Excavations

A Polygonal hallV VestibuleF FrigidariumC CaldariumT Tepidariumf PraefurniumLa Latrine (?)L Sudatorium (?)

or heated basins------ Unexcavated

Blocking

N

ble columns. Traces of mortar still adhere to its sur-face, indicating the location of each column base oppo-site one of the projecting brick piers that stand betweenthe six apsidal bays in the walls. Brick arches linkedthe columns to the main core of the building, and it isclear from the surviving architectural elements that thesuperstructure included an impressive vaulted roof. Theinterior of the hall was lavishly furnished. The walls andfloors were clad in marble, and the six columns were

The interior of the polygonal hall (6th century AD).

The central stylobate of the polygonal hall, during excavation in 2001.

130

Amorium

provided with a matching set of capitals decorated withraised crosses and blank round bosses. At the centreof the hall there is now a deep circular hole. It is unlike-ly that it was once filled with water, forming some sortof pool, but it may have marked the location of a largebasin or fountain, for a drainage channel has beenfound running through and under the exterior wall inthe southeast bay of the hall.

Pottery evidence from the earth fill below the levelof the floors within the polygonal hall supports the dat-ing offered by the architectural evidence. In all likeli-hood, therefore, the bathhouse complex was construct-ed at some point during the 6th century AD. However,the polygonal hall was subsequently abandoned, andall three doorways (and even the drainage channel)were blocked up, effectively sealing the structure offfrom the outside. The reason for this has been a hotlydisputed subject amongst the excavators. Some con-tend that the building must have become structural-ly unsafe, possibly as the result of an earthquake, butothers would attribute its abandonment to social or eco-

One of the six capitals found in the rubble inside the polygonal hall(6th century AD).

The polygonal hall was richly appointedwith marble on the walls and floor. Six

coloured marble columns stood in a circlearound the centre, each topped with a

limestone capital decorated with crosses.

131

The Excavations

nomic pressures in the 8th century AD. The latter inter-pretation is more likely, given that much of the inte-rior décor was stripped from the walls and floors beforethe doorways were blocked up. Indeed this, combinedwith the fact that the interior columns and their cap-itals were left behind, suggests that the fabric of thebuilding remained sound. Thereafter the derelict hallgradually filled up with debris and was never reused.Fortunately, too, it evaded the attentions of local stonerobbers during the 20th century, and so has survivedto provide us with an insight into the grand buildingsthat once adorned Early Byzantine Amorium.

The rectangular building that housed the suite ofbathing rooms had a different fate to that of the polyg-onal hall. Structure 1 continued in use throughout theDark Ages and even saw reoccupation in MiddleByzantine times. The multiple layers of occupationhave confused the archaeological record, but through

Fallen columns and capital as found inside the polygonal hall in2002 (mid to late 6th century AD).

The polygonal hall was stripped of much ofits marble furnishings before it was sealedup and left as a derelict building, while the

bathhouse continued to function.

132

Amorium

careful excavation these have been largely unravelled.The history of the bathhouse can thus be told, and itseems highly probable that the various rooms retainedtheir original function until the mid-9th century AD.From the polygonal hall bathers would originally havepassed through the southern doorway into a squareroom augmented with apses to either side. These semi-circular bays may have held small pools of runningwater that drained into the latrine that was located tothe northeast. However, although a doorway was latermade through the eastern apse to give access direct-ly from this room to the latrine, it seems likely that inthe original layout of Structure 1 one proceeded firstsouthwards into the cold room or frigidarium. Herethere was a marble-clad bench built up against thesouth wall next to a small closet-like plunge bath inthe south-east corner. One could also enter this roomfrom the east, where there was a small vestibule witha well and a doorway to the outside. Perhaps this doorwas originally a service entrance, but after the polyg-onal hall was abandoned it became the only way intothe bathhouse. Alternatively, as with later Islamic baths,the 6th-century complex may have been furnished withtwo entrances, one large and grand for the summer,

Water channels, drains, and pits in the floor of the square roomleading from the polygonal hall to the frigidarium in the bathhouse.

133

The Excavations

the other small and practical for the cold winter months.The latrine, as mentioned above, was convenientlylocated next to this entrance room. It seems to havehad a large window in the east wall, providing light andventilation. It had a brick barrel-vaulted roof, tracesof which still remain. Likewise, there is a deep drainagechannel running the length of the north wall, in front

The latrine in the northeast corner of the bathhouse (EarlyByzantine-Dark Ages, 6th-early 9th century AD).

Hypocaust pilae in the tepidarium of the bathhouse (probably DarkAges, late 7th-early 9th century AD).

134

Amorium

of which there are remains of a sill; this probably heldseats in place over the channel. If so, there would beroom only for three or four occupants at a time in thelatrine.

To the west of the cold room there was the warmroom (tepidarium), and beyond that the hot room (cal-darium). Both of these rooms were heated by meansof hypocausts and flues running up the walls behindthe marble revetment. At the western end of the build-ing were two smaller units, also furnished withhypocausts, flanking alarge vent that broughthot air from the furnaceroom. It is likely thatthese were individualbaths, filled with hotwater brought by handbut kept comfortablywarm by the proximityof the furnace and thehot air passing by theminto the caldarium.Although the floors inboth the hot and warmrooms have collapsed,the hypocausts them-selves are well pre-served. The hypocaustpillars (pilae) com-prised a number of dif-ferent types; as well asthe usual ones made ofbricks or tiles laid hor-izontally one on top of

One room in the bathhouse formed a smalllatrine, with seating arranged along one

wall above a drainage channel.

The hypocaust floor and surviving pilae inthe caldarium of the bathhouse (probablyDark Ages, late 7th-early 9th century AD).

135

The Excavations

another, there were pilae with bricks laid vertically.Others, again, were made out of stone (such as small,re-used columns) or large terracotta pipes. Nor wasthe arrangement of the pillars as regular as is usual-ly found in Roman baths; instead, they are placed ina somewhat haphazard fashion, sometimes groupedin pairs. All this suggests that in their present state thehypocausts may represent a refurbishment of the bath-house. Likewise, a number of terracotta spacers wererecovered from the debris in the hypocausts. Thesewould have been placed on the walls to provide a gapfor hot air to circulate between the brick inner surfaceof the walls and the marble revetment that faced them.However, a number of crudely-shaped brick fragmentswere also used as ‘replacement’ spacers. Finally, thefabric of the bathhouse itself also contains spolia.

Different types of pilae used in the tepidarium (probably Dark Ages,late 7th-early 9th century AD).

Strenuous efforts were clearly made to repairand maintain the bathhouse during the Dark

Ages, but it is unknown who provided thefunds and workmen for its upkeep.

Likewise, there is no evidence for how thebaths functioned on a day-to-day basis.

136

Amorium

Although some of the re-used material can beassigned to a later phase in the use of the building(when it was no longer a bathhouse), significantly anumber of spolia blocks were incorporated into theheating system when it was still functioning. For exam-ple, a re-used lintel, decorated with a christogram andgrape vines, was turned on its back and made into asupporting block above the hot air channel betweenthe furnace and the caldarium.

The bathhouse rooms, however, were lavishly fit-ted out with marble floor slabs and wall revetment,some of which remains in situ, while the impressionsof other slabs can be seen in the bedding plaster thatstill covers the brick walls. Although much of the mar-ble can be identified as local (that is, from Anatolia –although not necessar-ily from the nearbyquarries of Docimeium,modern ‹scehisar), anumber of fragmentsand small intact slabsof imported colouredmarble have beenfound. These includeantico verde and lapis

A finely decorated lintel block reused in the Dark Age bathhouse.

137

The Excavations

Fragments of imported red marbleand porphyry.

lacedaimonicus, both green-veined marbles fromGreece. Even though the pieces may have been re-used in the bathhouse, it is clear that such exotic andcolourful stones were chosen in order to enhance theappearance of the building, and in any event it showsthat at some point someone was rich enough to haveGreek marble brought to Amorium.

It cannot be proved that the repairs to Structure 1took place as a result of damage, natural or otherwise.It is, however, very likely that the refurbishment of the

bathhouse occurred when the polygonalhall (Structure 3) was aban-

doned and the passagewaybetween the two halves ofthe complex was blockedup. The evidence for whenthis took place is limited,but the presence of re-usedarchitectural elementssuch as the early Christian

lintel block in the buildingprovides some indication of

date – and this can now beplaced in the mid-8th

century AD. Likewise,the abandonment of

Fragment of antico verde marble imported from Greece, reused inthe bathhouse.

138

Amorium

Copper alloy coin of the emperorNicephorus II (AD 963-969), found in the

Lower City Enclosure.

the use of the bathhouse occurred some time beforethe mid-10th century AD, since coins of the emperorNicephorus II (AD 960-969) have been found in thefoundations of the rubble and brick floors that wereconstructed over the ruins of the hypocausts in thewarm and hot rooms. Other finds from both inside andthe immediate vicinity of the bathhouse include coinsof the Dark Ages (mid-7th to early 9th century) – thelatest of these was a copper alloy follis of Michael II(AD 820-829) recovered from the floor of the frigidar-ium. The fill within the hypocaust cellars of the cal-darium and tepidarium was rich in ash and containeda number of intact and broken pottery vessels. Onesuch vessel was a fragmentary large coarse ware jugthat has an incised Greek inscription running aroundthe shoulder. Although not yet fully deciphered, it clear-ly begins with a quotation from Psalm 29, verse 3: +fvnÓ K(yrºo)y ®pÓ tvn ... (Êdåtvn) “the voice of the Lordupon the waters”. These and other finds, including partof a human skull, strongly suggest that the bathhousewas not just abandoned but was violently destroyed,and it is tempting to associate it with the destructionof other parts of the city that have been linked to theevents in AD 838.

The Forty-Two Martyrs of Amorium

The story of the siege and sack of Amorium concludeswith and leads on to that of the fate of the captives.When the forces of the caliph al-Mu’tassim broke intothe city on August 12, AD 838, the inhabitants ofAmorium, together with the soldiers who had tried invain to defend them, were left at the mercy of the vic-

The bathhouse probably continued to beused until the mid-9th century AD.

Thereafter it lay derelict for over a centurybefore the shell of the building was

converted for other uses during the MiddleByzantine period.

139

The Excavations

tors. Some were doubtless cutdown immediately or otherwisesuffered death, others probablymanaged to flee, but a largenumber of people were captured.The sources tell us that theseprisoners were then forced toaccompany the Arab army whenit retreated from Amorium, head-ing back towards Cilicia andMesopotamia. Amongst the cap-tives forty-two high-ranking menwere singled out for special treat-ment. They, it seems, wereimprisoned in Tarsus (then partof Abbasid territory) to awaitexchange or ransom by the

Byzantines, but they waited in vain. Eventually theywere taken to Samarra where, since they refused toconvert to Islam, they were put to death in AD 845.Their story is recorded in a Byzantine Martyrion text,probably written soon after the mid-9th century AD.As a result their memory is preserved to this day, andthe Orthodox Church still celebrates their martyrdomas part of the ecclesiastical calendar on March 6. Whilethe story is itself fascinating, the contemporary detailsthat it provides are of particular relevance to the set-ting at Amorium in the 9th century AD. The textdescribes the conditions that the Forty-Two Martyrshad to endure during their six-and-a-half year impris-onment. The prisoners are said to have been kept inshackles in a dark dungeon with little food and water,and no creature comforts. One passage specificallystates that ‘they were not allowed to visit a bath, tocut their hair, or to expose themselves to the sun’s

Amongst the captivesforty-two high-ranking

men were singled out forspecial treatment.

Detail of a picture of the Forty-Two Martyrs of Amorium

(Byzantine, 13th century AD).

140

Amorium

rays.’ Whether or not suchbathing facilities existed in Tarsusat the time, it is clearly impliedthat it was still the custom forByzantines to go to the baths ona regular basis and that, whendeprived of the opportunity, itwas considered a severe hard-ship. The recently excavatedbathhouse at Amorium wouldseem to add archaeological proofto this assertion.

Byzantine Baths

There is relatively little literary orarchaeological evidence for bathsand bathing in Byzantine times,so the discoveries at Amoriumare particularly significant. Bycontrast, a great deal is knownabout Roman baths (thermae);numerous examples have beenexcavated throughout the empireand form some of the mostimpressive Roman remains thattourists can now see and visit. The majority of suchbuildings, however, fell into disuse during late antiq-uity, when changing economic and social conditionsled to a decline in public bathing. For one thing it wasvery expensive to run a large public bathhouse – aswell as the cost of staff to maintain the building andserve the bathers, a constant supply of fuel would beneeded to heat the rooms and to provide hot water.Then there was the water supply itself. In Roman timesmost public baths were provided with a plentiful sup-ply of fresh running water by means of an aqueduct.Even in the largest metropolitan cities, however, aque-ducts fell into disrepair during late antiquity; in Istanbul,for example, it is known that the aqueduct of Valenswas cut by the Avars during the siege of Constantinoplein AD 626 and was not repaired until AD 766. Finally,the grandest Roman thermae benefited from imperi-

Marble column shaft withcross graffito (Dark Ages orMiddle Byzantine, 8th-11th

century AD).

141

The Excavations

al patronage, a practice thatseems to have ceased by the5th century AD. It is notknown who paid for the con-struction and maintenance oflater baths, but it is clear fromboth literary and archaeolog-ical evidence that small baths(balneae) were built and usedin Byzantine times.

It was, of course, mucheasier to maintain a smallbathhouse than an imperial-sized baths. Such establish-ments did not require a con-stant supply of water from anaqueduct, and the well nearthe entrance at the south-eastend of the Amorium bath-house clearly provided at

least some of the water that was needed there.Likewise, the cost of fuel may have been kept low byrestricted opening times, and in any case it was muchcheaper to heat a small set of bathing rooms than asuite of grand halls with large pools and recreationareas. The change in the function and appearance ofRoman baths can, in fact, be traced in the Amoriumbathhouse. In its original configuration it included thepolygonal hall, which not only had some architectur-al pretensions with its columns and alcoves but alsomust have served as an area where bathers could sitand gather socially. It is, therefore, one step removedfrom the typical Roman baths, where citizens met insurroundings that were not unlike those of other pub-

Small arched service conduit intothe bathhouse’s tepidarium, later

blocked up.

The Amorium bathhouse shows that duringthe Dark Ages the practice of public

bathing was preserved, even though its usewas probably severely restricted.

142

Amorium

lic spaces – such as the forum, the odeion, or the gym-nasium – where grand architecture and honorific stat-ues provided the backdrop to their daily lives.However, when the Amorium bathhouse was renovat-ed and the polygonal hall abandoned, these socialaspects of bathing were clearly much reduced.Indeed, the size of the individual rooms indicates thatthe Amorium bathhouse was used by only a few clientsat a time. Perhaps only certain privileged members ofthe community –the seniormilitary, civil andecclesiastical offi-cials (people, inother words, like theForty-Two Martyrs)–were allowed entran-ce. At any rate, thebaths clearly no longerserved as a meetingplace for the generalmass of citizens.

Bone gaming pieces found in the destruction layer near thebathhouse (probably early 9th century AD).

Fragmentary amber pendant, foundin 2004 (probably Byzantine).

143

The Excavations

The Upper C›ty

The Upper City is, as previously stated, a man-mademound or tell, which in ancient times must have servedas the acropolis for the city. Traces of a fortificationwall can be seen running around the edge of themound, and in places there are outlines of square orrectangular projecting towers. The excavations thatwere carried out on the Upper City between 1990 and1996 have shown that these defences are in fact twodistinct constructions. The earlier circuit is composedof large reused blocks (spolia) taken from buildingsof the Roman city and, most especially, from the exten-sive cemeteries that surrounded it. Around much ofthe mound this wall now lies exposed on the surfaceor is buried below the remains of a later wall. On thenorth slope, however, it is clear that large-scale rob-bing occurred many centuries ago; the stratigraphy ina trench excavated here clearly indicates that the firstwall was robbed out and backfilled before the second-phase defensive wall was built behind and inside it. Thissecond wall was first exposed along the south side ofthe Upper City in 1992. Its construction is of a differ-ent type, comprising small, irregular blockwork andpieces of brick. The interior face includes a series ofsmall buttresses, and similar features were observedat several other points around the perimeter of themound during the preliminary survey conducted in1990. In the places where the first-phase spolia con-

144

Amorium

View of the ravine cutting across the ancient cemeteries and quarries,looking south.

struction is visible on the surface it must be assumedthat all trace of the later curtain wall has disappeared.But elsewhere the second wall survives to a consid-erable height; for example, the 1993-95 excavationson the north side of the höyük have uncovered astretch of wall that still stands to a height of 2 metres.

The Upper City fortifications are linked to those ofthe Lower City at the northeast and southwest corners,so it is clear that at some point the two circuits formedpart of the same defensive system. However, the factthat the walls run round the whole of the mound, notjust the north and west sides, suggests that the UpperCity was a distinct and separate zone within the city.

Part of the Middle Byzantine wall around the Upper City mound(late 9th-11 century AD).

Excavated area in the southern sector of the Upper City mound.

145

The Excavations

One account of the siege in AD 838 speaks of howsome of the defenders fled to the Upper City when theArab forces first gained entry into Amorium – that is,presumably, into the Lower City. The archaeologicalfindings from the two trenches on the mound allow usto offer as a working hypothesis the following inter-pretation of the Upper City walls.

The use of spolia in the construction of fortificationsis often taken to indicate a period of insecurity, when

Aerial view of the southern sector of the Upper City mound,showing the line of the circuit wall.

146

Amorium

a city had to undertake suitable measures to defenditself from threat. It is unlikely, however, the work atAmorium should be assigned to the spate of fortifica-tion building that took place at various cities acrossAnatolia during the troubled times of the mid-3rd cen-tury AD. The use of spolia is also common to manyByzantine fortifications – the walls of Ankara castle pro-vide a good example, as do those that surround St.John’s Church on Ayasoluk near Ephesus. A date inthe early or mid-7th century AD would, therefore, seemmore appropriate for the first-phase wall on the UpperCity at Amorium. Its construction may have been

147

The Excavations

prompted not so much by the onset of Persian raidsin the AD 610-620s as by the new role that Amoriumassumed as the headquarters of the frontier general,the magister militum per Orientem, in the AD 640s. Themilitary authorities may perhaps have established asecure base in the more defensible Upper City, whilethe civilian inhabitants continued to look to protectionfrom the Lower City defences. The Amorium moundwould thus have fulfilled the same functions as thecitadel at Ankara when these two cities became provin-cial capitals in the latter part of the 7th century AD.

As in the case of the fortifications of the Lower City,the destruction of the first-phase Upper City defencesmay well be attributable to the disastrous siege in AD838. The large Roman tombstones that had been incor-porated into this wall were thereafter scattered acrossthe höyük in great confusion. These blocks were laterreused for a second time in the foundations and lowercourses of the houses and workshops that were builton the mound during the Middle Byzantine and Seljukperiods. The fact that these large blocks, eminentlysuitable for building work, were not incorporated intothe second-phase defensive curtain is good reason forsuggesting that the defences of the Upper City atAmorium were not rebuilt for some considerable time.At Ankara, which also fell to the Arabs in AD 838, theheadquarters of the Bucellarian theme were re-estab-lished almost at once, while the walls of the citadel wererestored twenty-one years later by the last emperorof the Amorian dynasty, Michael III (AD 842-868).There is as yet no archaeological evidence for suchreconstruction work at Amorium during his reign. Thisis, perhaps, surprising, and a convincing explanationfor the delay in refortifying Amorium is still lacking.The second-phase wall resembles the Lower City

148

Amorium

During the Dark Ages the Upper City musthave served as the citadel and stronghold of

Amorium – like the ‘kale’ at Ankara.

Enclosure wall in its construction, and it is possiblethat both were erected only in the mid-to-late 10th

century AD. Whether the Upper City then took on therole of a citadel again remains debatable, althoughsome explanation has to be offered for its refortifica-tion. It is evident that areas of the Lower City (i.e. theLower City Church) remained very much in use andplayed an important part in the life of Amorium inMiddle Byzantine times. It cannot be argued that thewall was built to protect all the major buildings andfunctionaries that were there. Rather, it would seemthat the new wall around the Upper City was both sym-bolic and defensive – a reflection of its former gloryand a reaffirmation of its status as a city of the empire.

Apart from the remains of a church visible on thesurface near the north-east corner of the mound, theUpper City does notappear to contain any sub-stantial Middle Byzantinestructures. Rather, goodevidence has been foundfor domestic and industri-al use both in MiddleByzantine times and laterduring the Turkish period.The houses were relative-ly small and simple with

Copper alloy handle of a frying pan, found on the Upper Citymound (probably Middle Byzantine, 10th-11th century AD).

Copper alloy earrings (MiddleByzantine, 10th-11th century AD).

149

The Excavations

foundations and lower walls made out of reusedmasonry, while the superstructure was of mud-brick,and the roofs comprised wooden beams, probably sup-porting layers of reed thatch and mud plaster. Theyappear to have differed little in plan and type of con-struction from the modern-day houses in the villageof Hisarköy. Those excavated on the southern side ofthe mound included one room where two large stor-age jars (pithoi) had been buried up to their rims in

the earth floor. Whenthey were first discov-ered in 1993, the jarswere still covered withtheir original stone lids,although when the lat-ter were lifted it wasfound that they werecompletely empty. Thepithoi appear to be ofByzantine manufacturesince one of them isdecorated with a crossand bears a sign inGreek letters, perhapsas an indication of its

Terracotta lid handle in the form of a stylised bird’s head (Byzantine,probably 8th-11th century AD).

150

Amorium

Large storage jar during excavation in 1993in the southern sector of the Upper City

mound (10th-11th century AD).

capacity. However, it may be that they were reusedby the Seljuks who settled at Amorium, probably inthe early 13th century AD. Certainly it appears that theTurkish newcomers adapted and reused the desertedliving quarters that they found still standing on theUpper City mound.

More remarkable was the discovery in the north-ern trench of a kiln, part of a potter’s workshop, togeth-er with a considerable amount of wasters – vessels thathad become deformed during firing and so discarded– in the surrounding earth fill. The kiln lay below theTurkish levels and so clearly belongs to the Byzantineperiod, a dating that was confirmed by the presencenearby of two open-necked storage jars, one of whichbore two maker’s stamps inscribed in Greek. Not onlyare excavated examples of Byzantine kilns extreme-ly rare, but the location of this kiln not just within thecity but actually in the ‘prestige’ area of the Upper Cityis significant. Such industrial sites are usually thoughtto have been kept well away from any major build-

The pottery kiln found in the northern sector of the Upper Citymound in 1995 (Byzantine, 8th - early 9th century AD).

The discovery in 1995 of a kiln in theUpper City provided the first concrete

evidence for Byzantine pottery productionat Amorium.

151

The Excavations

ings or living quarters because of the pollution and therisk of fire that were associated with them. Yet at

Amorium it is situated notfar the Upper City churchand the prevailing norther-ly winds would have car-ried the smoke from thekiln across most of the set-tlement on the mound. Inthe coming years it ishoped that further excava-tion of the area will bringto light more evidence forByzantine pottery produc-tion at Amorium.

The excavations in thenorthern trench have alsoshown that Turkish occupa-tion continued well into the18th century AD. Here atwo-room structure wasuncovered whose roof tim-bers had collapsed andcovered the floors of beat-en earth. One room, whichhad a stone-lined fireplacebuilt into one wall, pro-

Large stamped jar found near the kilnon the Upper City mound (Byzantine,

8th - early 9th century AD).

Pottery wasters found with the kiln on the Upper City mound(Byzantine, 8th - early 9th century AD).

152

Amorium

duced a silver coin of the Ottoman sultan Mustafa III (AD1757-1774) and the iron mechanism of a flintlock mus-ket, complete with the flint still in place. Tree-ring dat-ing of the roof timbers provided further confirmation ofthe Ottoman date for this final stage of occupation onthe Upper City. Although the Turkish settlement mayhave been quite small, the finds of glazed pottery, includ-ing fragments of both ‹znik plates and Kütahya coffeecups, and of Ottoman tobacco pipes indicate that it

achieved a certaindegree of wealth andmaterial prosperitybefore its ultimatedemise. This appearsto have occurred dur-ing the latter part ofthe 18th century,perhaps as a result ofgovernment efforts toconsolidate and con-centrate the ruralpopulation for pur-poses of taxation andsecurity. By 1836,when WilliamHamilton first identi-fied the ruins as thoseof Amorium, it seemsthat the site wasentirely deserted.

Fragment of Byzantine polychromeware (Middle Byzantine, 10th-11th

century AD).

In future seasons thenature of the

Byzantine, Seljuk, andOttoman occupationof the Upper City will

be investigated ingreater detail.

Flintlock gun mechanism found in thenorthern sector of the Upper Citymound (Ottoman, 18th century).

153

The Excavations

The Necropol›s

Roman Amorium was provided with an extensivenecropolis that almost completely surrounded the cityitself. One cemetery stretches over the broken groundto the west of the Lower City and to either side of theravine through which the modern road now approach-es Hisarköy from Emirda¤. The tombs are of varioustypes, ranging from a single example of a tumulus tonumerous rock-cut shaft graves, and from monumen-

Fragments of Kütahya ware coffee cups found at Amorium(Ottoman).

Rock-cut graves in the southwest cemetery of Amorium (Roman, 1st-3rd century AD).

154

Amorium

tal tombs to simple funerary stelae. During the 1994season the dig team was made aware of the existenceof an underground rock-cut tomb approximately 0.5km. to the south-west of the Lower City gateway(Trench AB). The tomb had been known to the localsfor some time and had been the target of some unau-thorised digging. The large limestone slabs that formed

Drawing of the ravine and rock-cut tombs by Margaret Gill.

Interior of the rock-cut tomb at the entrance to the ravine on theEmirda¤ road (Roman, 1st-3rd century AD).

Extensive cemeteries surrounded theRoman city of Amorium. Various different

types of Roman tomb can be found inthem.

155

The Excavations

the roof of the vestibule or passageway (dromos) tothe tomb had been prised apart in order to provideaccess to the interior of the tomb, which had been thor-oughly ransacked. On first inspection it was found thatthe dromos was clogged with loose rubble, while theburial chamber itself was three-quarters filled with amound of earth mixed with substantial quantities ofbone. It was, therefore, decided to carry out a rescueexcavation there in the following year in order to ensurethe recovery of whatever artefacts were left, the col-lection of the bone material, and the restoration andpreservation of the site itself.

The tomb is a typical Roman arcosolium structure,partially excavated out of a limestone outcrop, with ashort man-made dromos and a vertical entrance shaft.There are three burial couches, two along the sidesand one across the back of the chamber, carved outof the bedrock. Each couch measures approximate-ly 1.85 by 0.9 metres, sufficient to accommodate anaverage-sized adult. The couches are edged with a lowrock-cut barrier and arranged around a narrowsunken passage, which served to drain the tomb. Thefront edges of the two side couches had been dam-aged, while the barrier separating them from the couchat the rear was pierced with a central drainage chan-nel. Neither the main chamber nor the dromos retainedany trace of decoration or inscription; there was alsono sign of any wall plaster.

The excavation produced a very large quantity ofhuman skeletal material. Although no intact skeletonsor articulated groups of bones were discovered, a studyof the remains has shown that more than 90 individu-als (men, women, and adolescents) were buried in thetomb. It is not a mass grave, for it would have beenimpossible to place so many fully-fleshed bodies in thetomb at one time. Rather, the tomb must have been

A Roman tomb excavated in 1995 wasfound to contain the remains of over ninety

individuals, which must have beendeposited over a prolonged period of time.

156

Amorium

reopened on numerous occasions and bodies addedindividually or, at most, in twos or threes. So, althoughit was constructed as a family tomb for only three occu-pants, the site was frequently re-used for additional buri-als, not necessarily associated with or related to the orig-inal owners. By contrast, the amount of pottery shardsrecovered was minimal, suggesting that most of thegrave goods had been removed during the re-use and/orlooting of the tomb. Nevertheless, the profiles of two ves-sels could be reconstructed from fragments found scat-tered in the tomb. Both are examples of Late Romanware, dating between the 5th and early 7th centuries AD.In addition, a few objects in other materials had beenoverlooked by the tomb-robbers. The most attractiveand distinctive find was a bronze belt buckle (of the so-called ‘Syracuse’ type), complete with its pin. This typeof buckle, once thought to originate from central Europe,is now recognised as a Byzantine artefact and is datedto the 6th or early 7th century AD.

The precise dating of the tomb remains problem-atic, largely because of the dearth of datable materi-

Interior of the rock-cut tomb found in 1994, after cleaning andconservation (Roman, probably 2nd century AD).

157

The Excavations

al and the extensive disturbance of the burials. Thefinds, however, indicate that the tomb remained in usefor a considerable period, probably extending from the2nd century AD, through the Late Roman period, andwell into Byzantine times. After the rescue excavationhad been successfully completed, the interior was

Copper alloy belt buckle from the rock-cut tomb (Early Byzantine,6th-early 7th century AD.

Early Christian tomb with Roman doorstones as re-used walls,before excavation.

158

Amorium

planned, the finds were drawn and recorded, and thepainstaking job of washing, counting, and studying allof the human bone material was carried out. The tombwas subsequently closed and made secure. It can nowbe visited under the supervision of the site guard.

Another tomb was excavated in 2005. This is locat-ed on the northern side of the ravine and overlooksboth the Upper City mound and the modern road thatleads into the village from Emirda¤. The tomb hasbeen partially exposed for many years as a result ofillicit excavations that had uncovered several lime-stone funerary stelae of the Phrygian doorstone type.The present excavations, however, revealed that thetomb structure wasmore extensive thananticipated, compris-ing four separate butinterconnected com-partments. The divid-ing walls, comprisingre-used Roman door-stones, were arrangedso that narrow pas-sageways were leftbetween them and theexterior walls of thetomb on the south,east, and north sides.Whether the intentionwas to allow accessfrom one compart-ment to anotherremains uncertain. Early Christian tomb during excavation,

revealing additional compartments.

New tombs were constructed in earlyByzantine times with distinctive Christianfeatures but reusing earlier pagan Roman

tombstones.

159

The Excavations

Although Byzantine graves that reused earlierRoman tombstones are well attested (notably at neigh-bouring Pessinus), the peculiar arrangement of theslabs to create a communal tomb of four interlockingcompartments does appear to be exceptional. This

arrangement provides acruciform plan to thetomb as a whole, and itwould appear that it wasbuilt specifically forChristian occupants sinceit is orientated east-west.The presence of Christiancrosses and symbolsreinforces this conclu-sion, as does the fact thatthose bodies thatremained semi-articulat-

Early Christian tomb after excavation, showing four compartmentsarranged around a cross-shaped partition.

Copper alloy pendant crossfrom the Early Christian

tomb.

The tomb has a cruciform plan and itwould appear that it was built specifically

for Christian occupants since it isorientated east-west.

160

Amorium

ed appeared tohave been laidin the tombwith their headsto the west.However, aswith the rock-cut tomb, it isdifficult to datethis Christiantomb preciselybecause of itsuse for multipleburials. Its construction clearly postdates the Romanperiod, although the fill contained fragments of earlyRoman pottery and glass. Two intact terracotta lampsfound in the northeastern compartment belong to ageneral type that is placed ca. ad 550-650, implyingthat the tomb was built before the late 6th or early 7th

century AD. But the decorated reliquary cross and pen-dant from the northwestern compartment would be bet-ter placed in the 9th century AD. If so, it would sug-gest that the population of Amorium continued to usetombs in the ancient cemeteries well into Byzantinetimes.

Two terracotta lamps found in the EarlyChristian tomb (Early Byzantine, 6th-early

7th century AD).

Byzantine reliquary cross found in the Early Christian tomb.

161

The Excavations

The Amor›an

Dynasty

163

�Detail of a copper alloy coin of Michael II and Theophilus (AD821-829), found during the excavations.

Copper alloy coin of Theophilus, showing the emperor intriumphal regalia, found in 2006.

Although the dynasty was short-lived, lasting for justunder half a century, its existence nevertheless con-stituted an unusual period of relative stability inByzantine history. It was a rare occurrence for anyemperor to die naturally and to be succeeded peace-fully by his son and heir; it was even more extraor-dinary for the throne to pass to a third generation. Asmentioned above, Michael II was a native of Amorium.His son, Theophilus, was also probably born and spentat least some of his childhood years there. The cityremained of special significance to Theophilusthroughout his reign, and this may be seen as the mainreason why it became the principal target for the Arabattack in AD 838.

The details of Michael II’s early life are obscure.His family is said to have been extremely poor, butsources also refer to him receiving some educationand to his skill at judging all forms of livestock. It was,perhaps, natural for him, given his background atAmorium, to choose a military career. Despite someform of speech impediment that later earned him thenickname of ‘the Stammerer’, Michael II quicklyshowed ability as a soldier and gained both promo-tion and recognition from his superiors. In AD 802

164

Reverse of a gold coin, minted inAD 830/1-840, depicting

Michael II and Theophilus’ firstson, Constantine (private

collection).

D uring the mid-9th century

AD Byzantium wasruled by threeemperors insuccession fromfather to son –Michael II (AD820-829),Theophilus (AD829-842), andMichael III (AD842-867).

Amorium

Michael was enrolled in thebodyguard of the general(strategos) Bardanes, alongwith Thomas the Slav and thefuture emperor Leo V (AD 813-820). It seems that Michaeland Leo became close person-al friends; their ties of friend-ship were strengthened bytheir marriages to two sisters,Thecla and Barka respective-ly, the daughters of their com-mander, Bardanes. Later, butbefore July AD 813, Leo alsobecame godfather to Michael’sson, Theophilus. Most scholarsregard Bardanes and hisdaughters as of Armenian ori-gin, but the fact that he had thenickname of Tourkos (‘theTurk’) suggests otherwise. If hewas a Turk who had becomea Christian, it would not havebeen impossible for him to takean Armenian Christian name.

Byzantine manuscriptillustration of goats and awolf (Middle Byzantine,

11th century AD).

Brick fragment with paw marks of a large dog (Late Roman or Byzantine).

165

The Amorian Dynasty

So, it may not be idle to speculate that the futureemperor Theophilus had a Turkish grandfather!

When Bardanes rebelled in AD 803, both Michaeland Leo deserted him in favour of the reigning emper-or, Nicephorus I, an act for which Michael received asignificant promotion as reward for his loyalty.Shortly after the accession of the emperor Michael IRhangabe in AD 811 Leo was appointed to commandthe Army of the Anatolics based at Amorium, andMichael was summoned by his old friend to join himthere as his confidential advisor and right-hand man.Sources indicate that Michael played a prominent rolein Leo’s bid for the throne after Michael I’s defeat atthe battle of Versinicia in Thrace. After Leo becameemperor in late AD 813, Michael was duly rewardedwith the title of ‘patrician’ (patricius) and made com-mander (comes) of the imperial guards (the excu-bitores), a position that he apparently held until AD820.

Some time in the summer or early autumn of AD820 relations between Leo V and Michael turned sour,and the emperor began to suspect his old friend ofbeing involved in a plot against him. Accusationsagainst Michael led to his imprisonment, and he onlyescaped death on Christmas Eve because of the inter-vention of Leo’s second wife, the empress Theodosia.She urged her husband to postpone the execution untilafter the Christmas festivities. Leo reluctantly accept-ed her plea, but his hesitation proved fatal. It gavethose involved in the conspiracy time to rally theirforces and so strike the emperor down while he attend-ed church services on Christmas Day. It remainsuncertain whether Michael was in fact privy of the plotto assassinate Leo, but it is clear that he was the prin-

Michael II was a career soldier who rosethrough the ranks to become emperor inAD 820. His path to the imperial throne

was not smooth but involved much intrigueand betrayal.

166

Amorium

cipal beneficiary since he was promptly acclaimed asLeo’s successor.

The sources certainly give the impression that Leo’sreign became more unpopular as it progressed, where-as Michael’s accession to the throne is said to havebeen met with general approval. Soon afterwards hisown coronation he crowned his son, Theophilus, asco-emperor, perhaps partly in reaction to news of therebellion of his other former comrade-in-arms,Thomas the Slav, but it would also seem that fromthe outset Michael was bent on establishing his poweralong dynastic lines. For example, when his wifeThecla died in ca. AD 824, he decided to remarry,despite the fact that the Greek Church frowned on sec-ond marriages. He chose as his new bride,Euphrosyne, the daughter of the emperor ConstantineVI (AD 780-797), although she was then living as anun in a convent. This could be interpreted as delib-

167

The Amorian Dynasty

Marble relief fragment of an eagle with its prey (Middle Byzantine, 11th century AD).

erate attempt to legitimise his rule. Similarly, variousstories that later grew up concerning prophecies andomens foretelling his rise to power may have beendeliberately started by Michael himself or, at least, bythose close to the throne.

The major event during the early part of Michael’sreign was the rebellion of Thomas the Slav. The strug-gle developed into a full-scale civil war, during whichMichael was besieged for a time within Constantinople.Only two themes, the Opsikion and the Armeniakon,

Manuscript illumination depicting the emperor Theophilus on histhrone, flanked by bodyguards and officials

(Byzantine, 13th century AD).

168

Amorium

remained loyal to the emperor. Thomas had beenserving as commander of the Army of the Anatolicsand was based at Amorium when news arrived Leo’sassassination and Michael’s elevation to the purple.Sources claim that Michael’s unpopularity was theresult of his associations with the heretical Athinganoi,since the influence of sect was particularly strong atAmorium, but this seems unlikely. The basis for the

169

The Amorian Dynasty

When Michael II became emperor he firsthad to contend with a serious rebellion ledby Thomas the Slav, who ironically was the

general in charge of the troops based atAmorium.

conflict should rather been seen as a simple powerstruggle between leading military figures. Michael wasa supporter of iconoclasm, but this does not appearto have been a significant factor in the struggle withThomas. Michael’s treatment of iconophiles wasapparently very lenient and certainly more so than thatof his predecessor, so that it is only in the later sourcesthat he is reviled for his religious beliefs and practices.It is surprising to find that Thomas enjoyed so muchsupport, especially at Amorium. The civil wardragged on for three years and had a serious effecton the Byzantine empire, weakening both the armyand the navy. As a result imperial control over out-lying parts of the empire diminished during Michael’sreign. The main beneficiaries were the Arabs, who inAD 826 overran the western part of Sicily and cap-tured the whole of Crete. The conflict must also haveaffected central Anatolia, although nothing is knownabout the part played by Amorium in the struggle orof the consequences the city may have suffered as aresult of backing the wrong side.

Michael died in AD 829, aged about 60, succumb-ing to a kidney infection. He was the first emperor formore than half a century to die of natural causes inhis own bed, and the first since Leo IV (AD 775-780)to leave a strong and healthy son to succeed him.

At the time of his father’s death Theophilus hadalready been co-emperor for eight years. His exactage in AD 829 is unknown, but sources speak of himas being a grown man, so he was probably born some-time in the first decade of the 9th century AD. Whatis clear is that he, like his father, was a native ofAmorium and spent his early years there. Hereceived a good education, and was at one time a pupilof John the Grammarian, the most gifted Byzantine

The civil war dragged on for three yearsand had a serious effect on the Byzantineempire, weakening both the army and the

navy.

170

Amorium

scholar of the age. Theophilus’ learning and culturedmanner are seen as being in stark contrast to the plain,unsophisticated ways of his soldier father. Since heinherited the empire without the stigma of usurpation,without serious opponents at home or enemies abroad,and without any pressing military or financial prob-lems, Theophilus’ reign began with a surge of opti-mism and energy. He was determined to win a rep-utation as a just and cultivated as well as a success-ful ruler, taking as his role model the famous Abbasidcaliph, Harun al-Rashid (AD 786-809).

Theophilus’ stepmother, Euphrosyne, had beennamed as co-ruler in Michael’s will, but her main taskwas to organise a ‘bride show’ and ensure thatTheophilus chose a suitable empress. It was imper-ative that the young emperor should marry and fatheran heir – another clear indication of the dynasticambitions of Michael and his family. Theophiluspicked a lady called Theodora to be his bride. Oncetheir marriage and joint coronation had been com-

171

The Amorian Dynasty

Gold coin of the emperor Theophilus, minted in AD 830/1-840(private collection).

pleted, Euphrosyne retired from public life. The mar-riage quickly produced three daughters before a sonand heir, Constantine, was born, probably in ca. AD834. Coins were minted depicting Constantine seat-ed beside his father, but he died while still an infant.Thereafter, the succession became increasingly acause for concern until, in AD 840, Theodora unex-pectedly provided Theophilus with another son, thefuture Michael III. Meanwhile, two further daughtershad been born, and these five imperial princessesserved to provide marriage alliances that could shoreup the Amorian dynasty.

Theophilus is noted for having initiated a huge pro-gramme of construction both in the capital and else-where. Sources speak of some repair work that wascarried out at Amorium during his reign, but most ofhis attention and resources were devoted to strength-ening the defences of Constantinople. The walls alongthe shore of the Golden Horn had been damaged dur-ing Thomas’s siege; their repair had been started byMichael II, but it was Theophilus who decided toheighten the sea walls along their entire length. Manyof the inscriptions recording this work, which wereinserted into the stonework of the towers, still survive.Attention was also paid to increasing the strength ofthe army, so that it recovered from the losses sufferedduring the civil war with Thomas. The army was rein-forced with detachments of Khurramites, a group fromeastern Iran, who sought refuge from religious perse-cution in the Caliphate. Finally, Theophilus also reor-ganised the ‘Theme’ system and extended it toCherson, a distant imperial outpost on the northernshores of the Black Sea.

Although Theophilus was apparently a greatadmirer of Islamic art and culture, he clearly saw the

172

Amorium

Theophilus’ stepmother, Euphrosyne, wasnamed as co-ruler in Michael’s will in order

to ensure a smooth succession.

Abbasid caliph as his political rival and enemy. Hethus embarked on a series of minor offensive cam-paigns against the Arabs on the eastern frontier. Theseenabled him to hold two ceremonial triumphs inConstantinople in AD 831 and AD 837. His victorieswere celebrated on a new series of copper alloy coins,whose reverse legend reads: ‘You conquer, OTheophilus Augustus.’ This was the first majorreform of the base metal coinage since earlyByzantine times. The standard reverse design on basemetal coinage had been introduced by the emperorAnastasius in AD 498. Similar coins, still bearing the

Manuscript illumination depicting the emperor Theophilus onhorseback with imperial troops (Byzantine, 13th century AD).

173

The Amorian Dynasty

denomination letter ‘M’ (for 40nummi), were minted at thebeginning of Theophilus’ reign,but were soon superseded by the‘Victory’ type. Coins with bothreverse types have been found

during the excavations at Amorium.Theophilus’ second triumph was prompted by the

capture in AD 837 of Sozopetra, a frontier fortress inthe southern foothills of the Taurus mountains nearthe modern city of Kahramanmarafl. It was not a par-ticularly significant military success, but Sozopetrahappened to be the birthplace of the ruling caliph, al-Mu‘tassim. Its capture and destruction by theByzantine forces was, therefore, a serious loss of face

for al-Mu‘tassim, who wasobliged to launch a large retal-iatory raid into Byzantine terri-tory, as much to win populari-ty with his own subjects as toinflict punishment onTheophilus. In the spring of thefollowing year he marched intoAnatolia, intent on destroyingthe emperor’s home town,Amorium.

Theophilus’ inability to repulsethe Arab invasion and to saveAmorium from capture showshis abilities as a general in avery poor light. Yet, despite the

Theophilus undertook alavish programme of

rebuilding both inConstantinople and

elsewhere in the empire.

Reverse of a copper alloy coinof the emperor Theophilus,

minted in AD 830/1-842, foundin the Lower City Enclosure.

Reverse of a copper alloy coinof the emperor Theophilus,

minted in AD 829-830/1, foundduring the excavations.

174

Amorium

reverse in AD 838 – which involved a severe defeaton the battlefield for the Byzantine army as well as thecomplete destruction of Amorium – he remainedemperor for a further three and a half years, and hisposition does not seem to have been seriously chal-lenged. The only consistent opposition to Theophilus’rule came from various clergymen who advocated iconveneration. Later sources saw his persecution oficonophiles as the reason for his military failures andeventually for his death. The loss of Amorium is rep-resented as having weighed heavily on Theophilus’conscience, and for the rest of his reign he exhibitednone of the energy and confidence that had so markedhis earlier years. Theophilus died on January 20, AD842, after he had been taken ill, probably with dysen-tery; he was buried in the Church of the Holy Apostlesin Constantinople.

Theophilus was succeeded by his two-year-old son,Michael III. The extreme youth of the new emperor nat-urally led to a great deal of court intrigue, as power-ful men tried to exert their influence first overTheodora, who acted as regent, and then over Michaelwhen he grew into boyhood.However, Theodora herselfseems to have been one of themain instigators in the restitu-tion of icon veneration in AD843. The first Sunday in Lent,when the Council announcedits decision, has been celebrat-ed in the Eastern Church eversince as the Sunday ofOrthodoxy.

Little is known aboutMichael’s childhood or youth,but in ca. AD 856 when hereached the age of about 16 itappears that he fell out withhis mother over her choice ofbride for him. Against hiswishes Theodora forcedMichael to marry Eudocia

175

The Amorian Dynasty

Byzantine graffito on smallmarble slab: cross with slogan‘Jesus Christ conquers’ (Middle

Byzantine, 10th-11th century AD).

Decaploitissa, althoughhe had already fallen inlove with anotherwoman, EudociaIngerina. As a conse-quence Michael replacedhis mother with herbrother, his uncleBardas, as his chief min-ister. Bardas continued todirect the Byzantine statesuccessfully in theemperor’s name for thenext ten years. In AD866, however, Bardaswas murdered in anoth-

er flurry of court intrigue, and because of Michael’sinability to rule in his own right Basil I (AD 867-886)was proclaimed as co-emperor. The joint reign did notlast long. The able and ambitious Basil plotted to haveMichael assassinated, and achieved his end on 24September, AD 867. Michael’s ineptitude and drunk-enness may, however, have been exaggerated bysources influenced by the propaganda of the succeed-ing Macedonian dynasty. For much of his reign theByzantine empire continued to prosper, and Michael’sname is found on numerous inscriptions recording therebuilding of fortifications at important cities such asNicaea and Ancyra. By contrast, there is as yet noevidence to suggest that Michael showed any inter-est in the reconstruction of Amorium.

With Michael’s death the Amorian dynasty official-ly came to an end. It had lasted for not quite half acentury and included only three emperors, whereas theMacedonian dynasty, which began with Basil I, con-tinued for the next two centuries. However, Basil’s suc-cessor Leo VI (AD 886-912) is claimed by somesources to be Michael’s illegitimate son by EudociaIngerina, who had been married off to Basil althoughshe remained Michael’s mistress. The story may findsome support from the fact that when Leo becameemperor (after the death of his supposedly real father

Marble archivolt fragment found atEmirda¤, probably from Amorium

(Byzantine, 10th century AD)..

176

Amorium

Basil), he had Michael’s remains transferred from atomb at Chrysopolis to the Church of the Holy Apostlesin Constantinople. Perhaps, then, the Amorian dynastysurvived, just as Amorium did, to witness and partic-ipate in the Byzantine revival of the 10th and 11th cen-turies AD.What is certain is that both the Amorian emperors andthe city played a significant part in the history of theByzantine empire in the 9th century AD.

With the death of Michael III in AD 867the Amorian dynasty came to an end,

having lasted for not quite half a century. Itprovides a rare example in Byzantine

history of a family that was able to retainthe imperial throne for three generations.

177

The Amorian Dynasty

Copper alloy coin of Michael II and Theophilus, found during theexcavations.

Select

B›bl›ography

General

J. Ash, A Byzantine Journey, London 1995.

W.J. Hamilton, Researches in Asia Minor, Pontus and Armenia, Vol.

1, London 1842.

M. Harrison (ed. W. Young), Mountain and Plain. From the LycianCoast to the Phrygian Plateau in the Late Roman and EarlyByzantine Period, Ann Arbor 2001.

C. Mango (ed.), The Oxford History of Byzantium, Oxford 2002.

S. Mitchell, Anatolia: Land, Men, and Gods in Asia Minor, 2 vols.,

Oxford 1993.

M. Waelkens, Die kleinasiatischen Türsteine, 2 vols., Mainz 1986.

F›nal Reports and Survey Art›cles

M.A.V. Gill (with contributions by C.S. Lightfoot, E.A. Ivison, and

M.T. Wypyski), Amorium Reports, Finds I: The Glass (1987-1997). BAR International Series 1070, Oxford 2002.

C.S. Lightfoot (ed.), Amorium Reports II: Research Papers andTechnical Reports. BAR International Series 1170, Oxford 2003.

E.A. Ivison, “Urban Renewal and Imperial Revival in Byzantium

(730-1025),” Byzantinische Forschungen 26 (2000), esp. 13-

18, 27.

Chris S. Lightfoot, “The Public and Domestic Architecture of a

Thematic Capital, the Archaeological Evidence from Amorium,”

in Byzantine Asia Minor (6th-12th cent.), Athens 1998 (National

Hellenic Research Foundation, Institute for Byzantine Research

International Symposium 6), 303-20.

C.S. Lightfoot, “The Survival of Cities in Byzantine Anatolia, the

Case of Amorium,” Byzantion 68, Fasc. 1 (1998), 56-71.

179

�Sard intaglio carved with a rustic scene of an old man and a harehanging from a tree (Roman, 2nd-3rd century AD).

C.S. Lightfoot, “Amorium and the Afyon Region in Byzantine

Times,” in R. Matthews (ed.), Ancient Anatolia. Fifty Years’Work by the British Institute of Archaeology at Ankara, London

1998, 301-14.

C.S. Lightfoot, “Amorium: The History and Archaeology of an

Ancient City in the Turkish Period,” in A. Aktafl-Yasa (ed.),

Uluslararas› Dördüncü Türk Kültürü Kongresi (4-7 Kas›m 1997,

Ankara), vol. 2 (Atatürk Kültür Merkezi Yay›n› 229), Ankara

2000, 79-89.

C.S. Lightfoot, “Byzantine Anatolia: reassessing the numismatic

evidence,” Revue Numismatique 158 (2002), 229-239.

Prel›m›nary Reports

Kaz› Sonuçlar› Toplant›s›, vols. 11-26 (Ankara 1990-2005).

For the years 1987-1995, see Anatolian Studies vols. 38-46 (1988-

1996).

For the years since 1995, see Dumbarton Oaks Papers vols. 51-53

(1997-1999), and 55-56 (2001-2002).

Shorter reports have also appeared in Anatolian Archaeology 1-

11 (1995 -2005).

180

Amorium

amorium guide lightfoot

Documents