kenyon frederic g. , books and readers in ancient greece and rome 1951

8/13/2019 Kenyon Frederic G. , Books and Readers in Ancient Greece and Rome 1951

1/80

BOOKS AND READERS

IN ANCIENT GREECE

AND ROME

Y

F R E D E R I C G KENYONLate irecUrr and Principal Librarian

o he ritish Museum

S ECO N D E D I T I O N

OXFORD

AT T H E CLARENDON PRESS

95 1


2/80

Oxford University Press Amen House London E CGLASGOW NEW YORK TORONTO MELBOURNE WELLINGTON

BOMBA Y CALCUTTA MADRAS CAPE TOWN

GeojJrey Cumberlege Publisher to the University

F I R S T E I T I O N 932

PRINTED IN GRBAT BRITAIN

AT THE UNIVERSITY PRESS OXFORD

BY CHARLES BATEY PRINTER TO THE UNlVERSITY

'

J

P R E F C E TO S E C O N D E D I T I O N

N this edition a nu mb er of additions have been

made. The most important additional informa-tion is that which relates to the recent dis-

coveries at Oxyrhynchus both British and Italianand to early Biblical papyri. Also the opportunityhas been taken to make a few corrections and tobring the bibliographical details up to date.

I have to thank Mr. C. H. Roberts for so memost useful references to recent literature on thesubject.

I have also to thank the staff of the ClarendonPress for helping with the correction of the proofof the final sheet when I was incapacitated byillness.

F.G.K.November 1950


3/80

PREF CE TO F I R S T E D I T I O N

T IS book is the outcome of a course of

three lectures which 1 was invited by theUniversity of London to deliver at King s

College in March 1932. The material has beenslightly expanded, but the general scale of treat-ment has not been alter ed. 1 does not claim toreplace the standard works on ancient book-production, but to supplement them, and thatespecially with regard to the period during whichpapyrus was the principal material in use. I t isin respect of this period t,hat our knowledge hasbeen chiefly increas ed in the course of the last twogenerations through the discoveries of papyri inEgypt. The object of this book is to bring togetherand make available for students the results of thesediscoveries. In particular, use has been made of heremarkable collection of papyrus codices recentlyacquired by Mr. A. Chester Beatty, which hasgreatly extended our knowledge of his transitiona lform ofbook, which appears to have had a specialvogue among the Christi an community in Egypt.

Although the subject of the book is primarilybibliographical, namely, the methods of book-construction from the date of Homer wheneverthat may have been) until the supersession ofpapyrus by vellum in the fourth century of our era,

I

\t

Preface to First Edition V ll

one of its main objects has been to show the bear-ings of the material and form of books on literaryhistory and criticism, and to consider what newlight has been thrown by recent research on theorigin and growth of he habit of eading in ancient

Greece and Rome. F.G.K.ugust I93


4/80

C O N T E N T S

I T H E U S E O F B O O K S IN N C I E N T

GREECE I

11. T H E PAPYRUS R O L L . 40

111. B O O K S AND R E A D I N G A T R O M E . 7

IV. V E L L U M AND T H E C O D E X . 87

A P P E N D I X . 1 2 1

I N D E X . 1 3 5

L I S T OF ILLUSTR TIONS

A poetess with tablets and stylus. Naples Museum.Photograph, Anderson . Facing page 6

A papyrus roll open. British Museum

Papyrus roll before opening. British Museum

Teacher and students with rolls. TrevesMuseum. Photograph, Giraudon . 56

A book-box capsa) containing rolls with sillybi. Page 6

A reader holding a roll of papyrus .

Roman inkpots. British -Museum

66

Facing page 72

Roman pens and styli. British Mus eum . 80

A papyrus codex. Heidelber g UniversityBetween pages 88 and 89

I

T H E USE OF B O O K S N N C I E N T

GREECE

UNT1L within a comparatively recentperiod, which may be n1easured by thelifetime of persons stillliving, our informa

tion with regard to the physical formation and thehabitual use of books in ancient Greece and Romewas singularly scanty. Our ancestors were dependent on casual allusions in Greek and Latinauthors, intelligible enough to those for whom theywere written, but not intended for the information

of distant ages, and in no case amounting to formaldescriptions. Such results as were obtainable fromthese sources were gathered together and set out inthe well-known handbooks of Birt, Gardthausen,Maunde Thompson, and others. The position,however, has been greatly changed by the discoveries ofGreek and a few Latin) papyri in Egyptduring the last sixty years. These have not onlygiven us a large number of actual examples ofbooks,

ranging from the end of the fourth century B.C. tothe seventh century of our era, but have also throwna good deal oflight on the extent of Greek literaturesurviving in at any rate one province of the RomanEmpire, andofthe reading habitsofthepopulation.

The object ofthe present book is to prese nt briefly5294 B


5/80

2 The Use o Books in ncient Greece

the present state of our knowledge on these subjects.Same af the information, including some of greatimpartance, has been acquired only within the lastfifteen years, and has not yet been incorporated inthe existing handbooks; while some has been longfamiliar, and only needs to be reconsidered andrestated in the light of the additional ~ v i d e n c eMuch ofit will only be interesting to those who careso much for books as to wish to know something ofthe details of their construction; but some of thesedetails also have their value for those who are con-cerned with textual criticism. Here also some previous conceptions have to be revised in the light ofour fuller knowledge.

I propose i r ~to deal with the origins of readingand the growth of the habit in the Greek world,from the earliest times to about the third centuryafter Christ; then to describe the appearance andmethods of manufacture of books during the sameperiod; next to consider the practice of reading inthe Roman world; and finally to descrlbe the changewhich came over the methods ofbook-productionin the early centuries of the Christian era, the de-

cline of pagan literature and the growth of that ofChristianity, leading up in the fourth century tothe general adoption of vellum as the material ofbooks, and the transition froIn the ancient world tothe Middle Ages. F or the earlier part of he periodunder consideration, before about 3 B.C. , such

The Use o Books in Ancient Greece 3additional evidence as we have comes from the increase of our general knowledge of the ancientworld, due to archaealogical exploration, and notfrom concrete examples of actual books. For thelater part the discoveries of papyri in Egypt co meinto play.

Any consideration ofGreek literature necessarilybegins with the Homcric poems; and in this con-nexion we have in the first place to take into accountwhat s now known as to the origins of writing in thecountries surrounding the eastern M e d i t e r r a n ~ a nWith regard to this subject, it s not too much to saythat our knowledge has been revolutionized bymodern archaeological discoveries. In the latterhalf of the nineteenth century it was accepted doc-trine that writing was practically unknown to theHomeric age. In Grote s I-listory o Greece whichthen held the field, it s laid down in round terms andwithout qualification that neither coined money,nor the art of writing, nor painting, nor sculpture,norimaginative architecture, belong to the Homericand Hesiodic times F ew things, he sa ys, can in hisopinion be more improbable than the existence of

long written poems in the ninth century before theChristian era. He would rather suppose that a smallreading class may have come into existence aboutthe middle ofthe seventh century, about which timethe opening of Egypt to Grecian commerce would

Part I, eh. 2 (vol. ii, p. 116, of edition of 1883).


6/80


furnish increased facilities for obtaining the requisitepapyrus to write upon. 1 To that period, accordingly,he would assign the commencemen of written literature in Greece.

Here we find ourselves at once at a point onwhich much evidence is available to us which wasunknown to Grote. Recent discoveries in Egypt, inMesopotamia, in Crete, and in Asia Minor havevastly increased our knowledge of the practice ofwriting in the ancient world. I will summarize thisevidence quite briefly, beginning first with Egypt. 2

The Prisse Papyrus in the Bibliotheque Nationaleat Paris is believed to have been written during theTwelfth Dynasty of Egypt (about 2200-2000 B.C.).I t contains two ethical treatises, the Teaching ofKagemna and the Teaching of Ptah-Hetep. According to the colophon at the end ofthe former ofthese treatises, Kagemna lived in the reign ofHuni,the predecessor ofSeneferu, at the end ofthe ThirdDynasty (about 3100 B.C.), and compiled this collection ofmoral precepts for the benefit ofhis chiIdren. Ptah-Hetep lived a little later , in the reign ofKing Isesi, or Assa, of the Fifth Dynasty (about2883-2855 B.C.), and his book also was written forhis sone We have thus from Egypt an actual manuscript which was written before the end of he third

Part I, eh. 20 (vol. ii pp. 143, 1 5 0 of edition of 1883) .: The following paragraph is extraeted from a book of my O W ,

Ancient Books and Modem Discoveries issued in a limited edition bythe Caxton Club of Chicago in 1927.

The Use o Books in ncient Greece 5

millennium B.C. and the works contained in it, ifwe are to believe their own statements, were composed respectively in the fourth millennium andearly in the third. N or is there any reason to doubtthese statements; for there is confirmatory evidence. The Book of the Dead, of which we havemanuscripts on papyrus dating from the EighteenthDynasty (about IS80-I320B.C.) and portionswritten in ink on wooden coffins of the EleventhDynasty or earlier, certa inly existed many centuries earlier, since the so-called Pyramid recensionis found carved in the pyramids of U nas, the lastking ofthe Fifth Dynasty, and ofTeta and Pepi Iofthe Sixth Dynasty. I t is not unreasonable to suppose that these texts must have been written onsome more ephemeral material before being carvedon stone. Egyptian tradition would carry then1. backeven farther still. A chapter is said to have beenfound in the reign of Semti of the First Dynasty;and the same king s name is associated with a recipein a book ofmedicine which was apparently writtenor edited in his reign. Further, King Zoser, of theThird Dynasty, is said to have been a patron of

literature, and portraits and tombs of persons described as scribes exist from the -Fourth Dynasty.Certain chapters ofthe Book ofthe Dead are said tohave been composed in the reign of Men-kau-ra(Mycerinus), thefifth king ofthat Dynasty, and the

Coffins of Amamu and Mentu-Hetep in the British Museum.


7/80


medical prescriptions preserved in British MuseunlPapyrus 10059 are assigned to the Fifth Dynasty.The practice ofwriting is therefore weIl atte sted forEgypt at least as far back as the third millennium B.C.

From Mesopotamia we have evidence of the useof writing of at least equal antiquity, and a muchgreater wealth of actual specimens. The archivesdiscovered by the American excavators at Nippurin 1888-1900 include tablets bearing Iiterary texts

notably the Sumerian version of he Deluge story)which are assigned to about 2100 B.C. or earlier. I

A fragment of the same narrative previously discovered, bears an actual date equivalent to 1967B C The texts themselves, being in the Sumerian

language, must have been composed much earlier.Nor would there have been any.difliculty aboutrecording them in writing; for the ~ v i d e n eof theexistence of cuneiform writing now goes back weIlinto the fourth millennium. Thousands of tabletsdiscovered at Telloh, at Ur and at Warka showthat writing was in constant use for the preservation

I My former colleague, Mr. C. J. Gadd n confirming thisdating, suggests that though the compositions contained in thesetablets are doubtless older, they may not have been committed towriting so long as the population and language of the countrywere predominantly Sumerian. The establishment of the Isin andLarsa dynasties at ab out this time marks the definite passing ofthe land to Semitic predominance and language, and it becamenecessary that the old and difficult Sumerian literature should bewritten down. Translations into Semitic begin to appear at aboutthe same time.

The Use o ooks in ncient Greece

of accounts, contracts, business archives, foundation tablets, building records, and other purposes ofdaily life throughout the whole ofthe third millennium B.C. and probably earlier. Writing was therefore available for literary purposes as ea rly as it waswanted; but to what extent it was actually usedthere is at present no evidence to determine.

From the Hittite Empire also, which dominatedeastern Asia Minor in the second millennium, wehave ample evidence of the use of writing. Thearchives of Boghaz-keui contain the records of theHittite sovereigns, written in both Semitic andHittite dialects in Babylonian cuneiform. Thesehave only recently been deciphered and some

progress made in the interpretation of the Hittitelanguage. As will be seen presently, they have adirect bearing on the Homeric question ; but in anyease they are decisive evidence of the habitual useof writing at this period. There is also writing inHittite hieroglyphics, but i t is only quite lately

1948) that a clue to their decipherment is believedto have been discovered.

Coming yet nearer to theGreek world, we have

the Cretan tablets discovered by Sir Arthur Evansat Knossos. These are in two forms of script, pictographie and linear. They have not yet been de-ciphered, but certainly include accounts. So faras is at present known, there are no literary textsamong them; but they prove the existence and


8/80

8 The Use 1 Books in ncient Greece

free use of writing in Crete at least as far back as2 B C

Within the past few years the evidence of earlywriting has been greatly extended. At Ras Shamrain north-west Syria, near Alexandretta, M. ClaudeSchaeffer in 1929 and following years discoveredthe library of the Amorite-Canaanite kingdom ofUgarit, formed mainly in the reign ofNigmed, inthe first half of the second millennium B.C. This includes literary texts in an alphabetic cuneiform oftwenty-nine signs, embodying the Canaanite religion which became the riyal of the religion ofJ ehovah, together with quantities of non-literarydocuments in various languages and scripts. At

Byblos, also on the Syrian coast, the sarcophagusof King Ahiram (discovered in 1922) bears a fewlines of Hebrew writing assigned to the thirteenthcentury. Farther eastward, at Mari ne ar the middleEuphrates, hundreds of texts (some literary) havebeen found, datable about the beginning of thesecond millennium; and at Nuzi, eastofthe Tigris,American excavators in 1925 and later discoveredhundreds of tablets of about the same date in the

language ofthe HurriaD:s the Horites or Hivites ofthe Old Testament).

I t s thereforenowamplyproved thatwriting wasin habitual use in Mesopotamia, in Egypt, in AsiaMinor, in Syria, and by the Minoan predecessorsof the Greeks in Crete at dates far preceding the

i

I1

II

The Use o Books in ncient Greece 9

beginnings of Greek literature; and the questionnaturally arises, Is it likely that a peop le such as theGreeks, oflivelyintelligence, of ready initiative, and

\ with literary tastes, would have remained ignorant,or have made no use, of an invention currentlypractised among their neighbours, and even theirMinoan ancestors, and of such obvious utility fortheir own purposes ? The natural presumption mustclearly be to the contrary.

Two objections may possibly be made at thispoint. First it may be asked how it s that no specimens of early writing have survived in Greece, asthey have in the adjoining countries that have beenmentioned. The answer is that the Greeks did not

use baked clay tablets, as did the Sumerians, Babylonians, Hittites, Syrians, Hurrians and Cretans,while skins and papyrus, which they did use in latertimes, and which must be taken to be the materialswhich they would naturally have used in earliertimes, could not survive in the Greek climate andsoil, as they have survived in the drier soil andclimate of Egypt. I t has further been argued that,if the early Greeks had been acquainted with the

practice of writing, some trace of i t would havesurvived in the form of nscriptions on stone. This isan argument which may any day be invalidated bynew discoveries; hut in any case it s far from beingconclusive. No inscriptions have so far been discovered among the extensive remains of Minoan


9/80


Crete; yet we know from the isolated discovery ofthe archives of Knossos by Sir Arthur Evans thatthe Minoans were familiar with the use ofwriting.The absence of inscriptions s therefore not a validargument against an acquaintance with letters onthe part of the M ycenaean Greeks; and the presumption to the contrary, based on the generaluse of writing in the countries around the easternMediterranean, appears to hold good.

Let us look now at the earliest remains of Greekliterature, and consider the probabilities as to themethod of their composition and presetvation.

Fifty years ago the Homeric and Hesiodit poemsstood out by themselves as an island, separated by

a gulf o f centuries from the mainland of Greekliterature ; and the Trojan war and all the traditional early history of Greece were regarded aslegendary, down to ab out the time of8010n. Now,as the result of the discoveries of the last sixtyyears,the gaps in our knowledge are being filled up, theorigins of Greece are being brought into connexionwith the histories of he surrounding countries, andwe are beginning to form a generalpicture of the

wholecourseofdevelopmentinthecountriesaroundthe eastern Mediterranean. We know that therewas a great civilization in Crete in the third andsecondmillennia, w i ~ came to an. abrupt end,while still in great splendour, in the fourteenthcentury. We know that the civilization to which we


give the name Mycenaean was a descendant andoffshootfrom the Minoan stern, and out ofthis, afterthe break caused by the Dorian invasion, comes thefull Hellenie culture which we know. Further, wehave lately learntfrom the Hittite records, now being painfully interpre ted, that contemporaneouslywith the Hittite Empire of about I3qo-12oo B.C.there was a considerable power on both sides ofthe Aegean whose princes and dominions borenames which the decipherers identify with namesfamiliar to us in Greek his tory and legend-E eocles,Aegeus, Achaeans, Lesbos, and so on. The generaltendency in this, as in other provinces ofknowledge,is to vindicate tradition, as containing at least asubstantial modicum oftruth. Let us consider therefore the Greek tradition as to both facts and dates.

Greek tradition assigned the origin ofGreek literature to the introduction ofthe alphabet by Cadmusfrom Phoenicia; and the traditional date ofCadmuss about 1350-130. The traditional date ofthe fallofTroy s 1184 or 1183 B.C. , and this accords sufficient ly weIl with the indications of he Hittite records.The traditional dates for Homer vary from 1075 toabout 875. Clinton accepts an intermediate date,ab out 975, which is that favoured by Aristotle. Thelatest writer on the subject, Dr. Bowra, after saying

Forrer s equations of Greek and Hitt i te forms of names,though possibly open to question in detail as to which I am notcompetent to judge), seem to me too striking and too numerousto justify disbelief in general.


10/80

12 The Use o Books in Ancient Greece

that the statement ofHerodatus that he lived in thelatter part of the ninth century and was a contem-porary of Hesiod may not be far from the truth ,proceeds in the next sentence to place hirn late inthe eighth century. The latter date is surely too late,and is perhaps merely a slip of the pen: for there isgood evidence for placing Arctinus, the author ofthe Aethiopis, in the first half ofthat century c. 775750 ; and if one thing is more certain than another,it is that the Iliad and Odysseypreceded the poems ofthe Epic Cycle.

If, then, we take the ninth century as the latestdate for Homer which suits the evidence,I whatsort of picture can we make of the manner of the

formation and preservation of his poems? (I amassuming that there was a personal Homer, whowas main ly responsible for both Iliad and Odyssey;but those who prefer the view, now less prevalentthan it was in the latter part ofthe nineteenth cen-tury, that they are the work ofa syndicate, or grewby themselves out of a number of detached lays,botched together by an incompetent editor, haveonly to substitute the phrase the Horneric poems

for Homer ) As I have already shown, there is no a

I It rnighthave been wiser to imitat e the prudence ofPausanias:Though I have investigated very carefully the dates of Hesiod

and Horner, I do not like to state my results, knowing as I do thecarping disposition of some people, especially of the professors ofpoetry at the present day ix. 30. 3, Frazer s trans1.). But it couldhardly be avoided.

The Use o Books in Ancient Greece

priori reason why they should not have been writtendown. W riting had been in common use for cen uriesin the lands adjoining the Aegean and Mediter-ranean on the east; and if Cadmus or any one elseintroduced writing to Greece abaut the fourteenthcentury,. that gives plenty of time for the establish-ment ofthe practice, and for the production ofthoseearl ier efforts in verse which must surely have pre-ceded the consummate technique ofHomer.

Looking at the matter from the point ofview ofinternal prob ability, the argument for a writtenHomer appears to me overwhelmingly strong. I t isdifficult even to conceive how poems on such a scalecould have been produced without the assistance of

written copies. I t is not that the feat ofmemorizingpoems of such length is incredible. On the contrary,one ofthe speakers in Xenophon s Symposium saysthat his father compelled hirn to learn the whole ofthe Homeric poems, and that he could still recite theentire Iliad and Odyssey. Paralleis are quoted frornvarious primitive peoples; and it is on record thatin the nineteenth century one young Wykeharnist( afterwards the defender of Silistria in the Crimean

War) learnt the whole of the I ~ a dand another thewhole of the Aeneid, in the days when such featsof memory were encouraged at Winchester. 2 Thepoems, once cornposed, could have been recited;

I Symp. iii 5.2 Leach, History o fVinchester College, p. 427.


11/80


but could they have been carried in the memory ofthe poet during the process of composition? And arewe to picture the poet after completing his magnumopus as assembling a corps of rhapsodists aroundhirn and reciting his work over and over to them

until they had committed it to memory? 1 is i f f i c ~ l tto believe. And if there was one original author scopy why should not each rhapsodist or at anyrate each school of rhapsodists have possessed onealso? 1 seems easier to believe this than the contrary.

Moreover even if we are prepared to believe thatHomerand the Homeridae could have composedand memorized the Homeric poems without bookwhat are we to say ofHesiod? Rhapsodists might

indeed think it worth while to learn the Catalogoiwhich contained the popular legends of the godsand heroes and for which listeners could readily befound; but can we suppose that there would havebeen much ora public for the Works and Days withits combination of a purely personal quarrel withagricultural precepts? 1 seems to me incredible thatsuch a poem should have survived unless it hadbeen written down whether onlead, as shown to

Pausanias on Helicon 01 in some other fashion. Thesame might perhaps be said of the poems of theEpic Cycle. The poets who produced them musthave been familiar with the Iliad and Odyssey notonly generally but in detail. They must ~ v beenrhapsodists themselves or habitual frequenters of


rhapsodists; and they must have acquired corps ofrhapsodists to learn their own poems in turn andrecite them. But Hesiod was either the contempo-rary of Homer, as was held by many in antiquityor not much later; and the later one brings downHomer, the nearer he comes to the earliest of theCyclic poets. And if the works of Hesiod and theCyclies were written down it is surely straining atagnat to refuse to allow the same to Homer. Nowthat the general antiquity of writing in the world ofthe Homeric age is established it is impossible tomaintain that writing was practised in the Greeklands in the seventh and eighth centuries but couldnot a v ~been known in the ninth, or even earlier.The basis for the old beliefis cut away.

1 believe therefore that sober criticism must allowthat the Iliad and Otfysseywere composed in writingand that writ ten copies ofthem existed to assist therhapsodists who recited them and to control theirvariations. I t is much more difficult however tosay what was the form of these written copies or inwhat manner they circulated. There is no evidenceof the existence of anything that can be called areading public. 1 do not attach any importance tothe fact that writing is not mentioned in Homer,except in the reference to the a ~ p a 7 avypa carriedby Bellerophon. There was little occasion for themention of t in such poems of war and adventure ;and 1 do not think it would be difficult to find


12/80

6 The Use of Books in Ancient Greece

nlodern poems describing a primitive age whichare equally devoid of references to it. But we do havereferences to the recital of poetry and if readinghad been a common practic e we might have foundsome allusion to it. At any rate, without evidence

which we certainly have not I do not think we areentitled to assume its existence. I

I imagine therefore that written copies ofpoems though they existed were rare, and werethe property of professional reciters from whomalone the general public derived their knowledgeof thema ~ what material they were written it isimpossible to say. Papyr us would have been obtainable from Egypt and we know from Herodotus that

skins were used at an early date in Asia Minor; butbeyond that we cannot go in the present state of ourknowledge. It is however fairly certain that poemsof such length whether written on skins or onpapyrus, could not have existed in single volumesbut must have occupied a number of separate rolls.

I Hesiod emphasizes the charm of literature, but it is poetryrecited not read (Theog. 98-103):

l ) dp 'TLS K a ~7Tlv(Jos XWVVEOK7]Sl;: (JvJLIi>

7] TaL KpaSt7]v aKaX7 ]JLlvos, aV TapaOI SJsMovad.wv (Jpa.7TWVKAta 7Tpo TlpwvavOpw7TwVv J L v ~ a nJLd.Kapd.s TE (Jeovs 0 tJ OAVJL7TOVXOVaLV,alt/J 0 ) Svat/>povl.wv 7 T L A ~ ( J ' T aov l 'T K7]SlwvJLI.JLvTJ Ta, Taxl.ws B ~7Tapl Tpa7TE Swpa (JEd.wV.

This is consistent with the theory that, while writing was employed for the composition and preservation of literature, thenormal method of publication was by recital.

A POETESS \VITH TABLETS AND STYLUS


13/80

Tlze Use o Books in Ancient Greece 7

Such a division into rolls might naturally lead to thedivision into books with which we are familiare Aswill be seen later a book ofThucydides correspondswith the contents of a papyrus roll of he largest sizein ordinary use; and the natural preSUlTIption is

that the twenty-four books of the Iliad represent astage in its history when it occupied twenty-fourrolls. \Vhen this division was made is unknown; butit may be worth observing that this theory of itsorigin would appear to point to a date earlier thanthe Alexandrian age. From that age we possess anumber ofspecimens ofHomeric manuscripts andit is clear that a normal roll could easily accommodate two books oftheIliad. Itwould seem therefore

that the division into twenty-four books nlay goback to aperiod when rolls were shorter or hand-writings larger; in which connexion it may be observed that the earliest extant literary papyrus (thatofTimotheus s Persae ofthe end ofthe 4th cent. B.G.)is in a much larger hand than later manuscripts.The Odyssey could of course easily have beenwritten in fewer rolls than the Iliad but the divisioninto twenty-four books was obviously made to

correspond.While on the subject of the tradition of the

Homeric poems it may be permissible to refer toa phenomenon, of which there is considerableevidence among the papyri of the third centuryB.G., namely the existence of copies containing a

5294 c


14/80

The Use of Books in ncient Greececonsiderable number of additionallines, which donot appear in our standard text. These lines are notsubstantial additions to the n,arrative ofthe poems,but are rather of the nature of verbal padding.There is no reason to regard them as authentie, andit is easy to accoun t for their existence. When copieswere scarce and means of inter-comparison almostnon-existent, it would have been easy for a rhap-sodist who fancied himse lf as an inventor ofHomericphrases to produce an edition of his own, whichmight obtain local currency. Only when copiesfrom various sources were brought together in aBingle place, as at Alexandria, was comparativecriticism possible, and then such excrescences asthese were speedily removed. They are rare inpapyri of the second century B.C. , and unknownlater.

With the beginning of the seventh century, orpossibly half a century earlier, we reach the lyricage ofGreek poetry, when the circulation oflitera-ture must be taken definitely to have passed fromthe rhapsodes to manuscript s. The recitations oftherhapsodes, at least of he Homeric poems, no doubt

continued to be a feature of the Pan-he llenic festivals; and the setting of poems to music, as in thecase of the odes of Terpander or Alcman, or laterthe epinicianodes anddithyrambs ofPindar, Simonides, and Bacchylides, provided a new form ofpublicity for the poets. But the more personal

The Use oJ ooks in ncient Greece 9

compositions, such as the satires of Archilochusthe political verses of Solon, and many of the lyricsof Sappho and Alcaeus, were quite unfitted formusical accompaniment or public performanc e,and must have circulated, so far as they circulated

at all, in manuscript. Throughout the seventh andsixth centuries the circumstances must have beenvery much the same. Poems, epic, elegiac, and lyric,were being produced in considerable quantities.The poets were acquainted with one another sworks, and enjoyed reputations anl0ng their contemporaries. Their poems must have been writtendown, and must have been accessible to those whodesired them: but we have no evidence to give

precision to our picture of the methods of publica-tion. Lyric ~ n elegiac poems, each composed fora particular purpose, may often have circulatedsingly; but whether their authors gathered themtogether into collected editions we do not know.Later, we know that they were so gathered; thatthe odes of Alcaeus formed six volumes, thatSappho s were arranged in nine,2 that the epinicia,dithyrambs, and paeans ofPindar and Bacchylides

were brought together in separate groups; but wedo not know that this arrangement was made in the

I Suidas, s.v.2 So Suidas, s.v. Lobel Xa7Tfpovs MATJ Oxford, 1925) suggests

t h ~the true number may be eight, and that the division may beAttIc, not Alexandrian. But it seems more in accord with theAlexandrian type of mind.


15/80


lifetimes ofthe respective poets, and itis more likelyto be attributable to the scholars of Alexandria. Anorganized book-trade at this time is highly improbable: at the same time it is evident that copiesof he works of all these poets must have existed and

have circulated in sufficient numbers to secure theircontinued preservation, and to make it possible forthem to be gathered into libraries when librariescame into being.

In the fifth century we reach the culminatingpoint of Greek literaturc, with Pindar, Simonides,and Aeschylus in its earlier portions, followed bySophocles and Euripides, Herodotus and Thucydides, Aristophanes and his rivals in comedy, and

all the great band ofpoets and prose writers whosurvive for us only in quotations and allusions. I tis aperiod of intense literary creativeness on thch i h ~ s tscale, and yet, so far as we canjudge, ofverylimited book-production. Oral methods of publicitycontinued. The odes of the ceremoniallyrists wereno doubt produced with musical accompanimenton the occasions for which they were written; thetragedies and comedies were performed on the

stage; even the works of the historians may havebeen read at the great festivals, as that ofHerodotusis said to have been. I t stands to reason that even forthese purposes a certain amount of production inmanuscript form was necessary. The performersmust have had copies from which they learnt their

The Use o Books in Ancient Greece 2 I

parts; the authors and reciters must have had theircopies to read from. t\That is to some extent doubtfulis the circulation of copies of books among thegeneral public, and the growth of ahabit of eading.

Contemporary references t o the reading ofbooksare very rare during the golden age ofGreek literature. In Plato s Phaedo Socrates is represented asreferring to a volume of Anaxagoras, which heheard read and subsequently procured; and in theApolog) he says that copies of Anaxagoras could bebought by any one for a drachma. 1 In the TheaetetusEucleides of Megara recalls a conversat ion betweenSocrates and Theaetetus which he wrote down atthe time, and which he now causes a slave to read

aloud to himselfand his companion. In the Phaedruson the other hand, Socrates speaks contemptuouslyof a dependence upon books in comparison withmemory;3 and his attitude is the same in his conversation with Euthydemus recorded by Xenophon. 4 More valuable for our present purpose is thestatement of Xenophon s Socrates that he was accustomed to unroll the treasures of the sages of oldtime which they had left in books written by them,

I Phaedo 97 b, 98 b in the latter passage the plural, ras lAOVSis used), Apol. 26 d.

2 Theaet. 43 a, b. Cf. Xen., Symp. iv. 27.3 Phaedr. 274 e Ir. Books may be useful to refresh the memory,

but are greatly inferior to the spoken word as a means of education (ibid. 276 d). Similarly Isocrates Phil. 25-7) admits theinferiority of the written to the spoken word.

4 Xen. Mem. IV ii.


16/80


and to study and make extracts from them with hisfriends. This proves the existence of the practice ofconsuIting books in a study or library; but it mustbe admitted that the general picture which wehave, b6th in Plato and in Xenophon, is of oralinstruction and conversation, not of reading andprivate study.

I t woul d be amistake, however, to overstress thisscantiness of evidence as an argument against theexistence and even the abundant existence and freeuse of books in the latter part of the fifth century.The very casualness of hese allusions is a proof hatthere was nothing extraordinary about them, andthat the accessibility of books might be taken forgranted. Aminute acquaintance with Homer wasassumed as part ofthe equipment of every educatedman, and allusions to Hesiod, to the Cyclic poets, orto the lyrists are made with an assurance which implies that they would be understood. Aristophaneshas a verbal knowledge of the works of Aeschylusand Euripides which could not have been derivedfrom stage representation alone. Thucydides knewand refers to the works ofhis predecessors in his tory ;

and the works of the physical philosophers and ofthe medical schools that followed Hippocrates couldonly have been known through circulation in manu-

I Xen. Mem. I vi. 14: -rovs (J 1]aavpovs -rWV 1TaAat aocfowv dvSpwvoiJs ~ r V O tKa-rl.At1TOV EV t /\I.OLS ypaoyaVTES, a V ~ / \ L T T W VKO VT/ vv .1. ' ,

~ , . . . ~ , (J " , 8TOrS cfolAO SoLEpX0p.aL, KaI. av Tl. opwp.V y OV E K / \ ~ Y 0 J . l Ea.

The Use o Books in Ancient Greece 3

scripts. Euthydemus, t he younger contemporary ofSocrates, possessed while still quite young a collec-ti on of he works of he best poets and philosophers:and the cheapness and ready accessibility of theworks of Anaxagoras, referred to above, cannot

have been confined to that philosopher.More illuminating, perhaps, is a line in the Frogsof Aristophanes, in which the chorus, inciting theriyal poets to bring their wares to the test, assuresthem that they need have no fear lest the audienceshould be unable to follow and appreciate them(as had apparently been the case at the first per-formance ofthe play); for they are now all men whohave seen the world in the course of their military

service, and each of them has his own copy of theplay in his hand and can understand the points. 2

This seems to imply that a certain amount ofbookknowledge of literature could now be presumed,though formerly it was not the case. And this is thegeneral conclusion to which all the evidence seemsto point.

A final reference may be made to a passage inXenophon s Anabasis, where among the cargoes of

I Ibid. IV. ii. I" , . , . , (J 8 t .2 Aris . Ran. I I 14, /\I.OV T EXWV E KaaTOS I av vn Ta E , La

The use of the singular seems to imply a single roll which thespectator could have with hirn in the theatre, not the collectedworks of Aeschylus and Euripides, which would of course occupymany rolls. But whatever be the exact explanation, it is irnplied thatthe younger generation is accustorned to the use of books. I haveto thank Mr. F. R. Earp for calling my attention to this passage.


17/80


ships wrecked near Salmydessus, on the north coastof Asia Minor, many books (1ToMat lAOL) are saidto have been included.

O f the formation of libraries there is practicallyno evidence. Athenaeus, 2 at a much later date,

does indeed refer to tradition al libraries formedby Pisistratus and Polycrates of Samos in the sixthcentury, but these are separated by two and a halfcenturies from the next collections that he canmention, and may be litde Inorc than mythical.His next example is Eucleides, who may be identical with the Megarian philosopher already referredto, though Athenaeus calls hirn ap Athenian. Hislist also includes the name of Euripides. These

libraries, however, like that owned by Euthydemus,as mentioned above, would have been small privatecollections of books, amounting at most to a fewscore-rolls; and even they seem to have been exceptional.

The general conclusion would therefore seem tobe that at the end of the fifth century and in theearly part of the fourth, books existed in Athens inconsiderable quantity, and were cheap and easily

accessible. Ahabit of reading was growing up, butwas not yet very firmly established. The general

Xen. Anab. vii. 5, 14 (quoted by Sandys, History o ClassicalScholarship, i. 84). There are references also (though rather indefinite) totheuseofbooks in Arist. Birds, 1288, andFrogs, 743 ;andapparently to book-shops in Eupolis, Fr 304 (OV a t At ~ V t a .

2. Deipllosophistae, i. 4.

I


opinion did not rate reading highly as a means ofmental training, in comparison with the play ofmind upon mind in oral discussion. Thc livelyAthenian mind accepted Bacon s distinction, andpreferred the ready man to the full man. The age of

the fuH man was, however, approaching.When we pass on another stage, from the genera

tion of Plato to that of Aristotle, a very distinctchange is marked. Whereas in the earlier period,while books must have been produced in considerable numbers, a readingpublic could hardly be saidto exist, we have now reached aperiod of readersand libraries. Even if it were not actually relatedthat Aristotle possessed a li brary , the fate of which

after his death is on record, it would be obviousfrom the me e list ofhis works that it must have beenso. His great compilations, whether of physicalscience or of political constitutions, could not havebeen produced without a reference library; and hispractice set an example which was followed by hisdisciples, such as Theophrastus and Menon, andwhich profoundly influenced the course of Greekliterary history. 1t is not too much to say that with

Aristotle the Greek world passed from oral instruction to the habit ofreading. The his tory oflibrariesin the Greek and Graeco-Roman world is rightly

Strabo, xi I I . i. 54, where Aristotle is described as TTPWTOS ~iul- VuvvayayJw t Ala, K a ~~ 3 t S a ~ a sTOVS JI AlYV1TTC[Jaut Alas t Ato-8 ~ K 1 J SU V V T a ~ l . v .


18/80

The Use o Books in ncient Greece

taken to start with the foundation of the Museumat Alexandria; but the foundation ofthe Museumand of the great Alexandrian Library was madepossible by the change of habit which took formin the time, and largely under the influence, of

Aristotle.From the date ofthe foundation ofthe Museum

and library of Alexandria we are at last on firmground in dealing with the book-world of Greekcivilization. We have no Ion ger to depend on deductions from ca sual allusions orfrom abstract prob-abilities. We have records on a fairly ample scale;and more than that, we have actual specimens ofthe books of that period, and know how they were

manufactured, and whatthey ookedlike. The creditof the foundation of these institutions is variouslyassigned to Ptolemy I - (Soter) and Ptolemy(Philadelphus). The truth would appear to be thatthe deliberate collection of books, to form a libraryand a centre of study was begun by Ptolemy I, as astep in the hellenization of Egypt, while the complete establishment of both Library and Museumwas accomplished by Philadelphus. Ptolemy I washimself an author and the friend of authors, and heentrusted the formation ofthe library to Demetriusof Phalerum, a disciple of Theophrastus and anencyclopaedic writer, who for ten years had givenAthens experience of the rule of a philosoph ertyrant. Expelled from Athens, he was glad to find

The Use o ooks in ncient Greece 7

an asyIutTI with Ptolemy 2 9 0 B.C.) and to confinehimselfto t he more harmless task ofcollecting books.

That books were by this time plentiful is shownby the size which Ptolemy s library almost inlmediately attained. According to one account,2 0 0 , 0 0 0 volumes had been collected by the end ofhis reign, i.e. within about five years. Such figuresare, however, totally unreIiabIe, and another storyspeaks of 1 0 0 , 0 0 0 at the death ofPhiIadelphus, andyet another of 7 0 0 , 0 0 0 when the Library was burntin the time of Caesar; but in any case it is clear thata substa ntial collection was formed by Soter, whichwas transferred by his son to the Museum ofwhichhe was the founder. Thi s Temple ofthe Muses was

the first great library after those formed by the kingsofNineveh and U garit; and besides being a libr ary,it was an Academy of Letters and Learning. Emi-nent men of letters and scholars, such as Callimachus, Apollonius Rhodius, and Aristarchus, wereplaced in succession at its head; students gatheredround it; a corps ofcopyists was employed to multi-ply manuscripts ; and Alexandria became the centreof the literary life of the Hellenistic world.

We have now reached astate ofthings which iscomparable with that of today. Greek culture hadbroken the bounds of the old Greek worId, hadspread over the Near East and the Mediterraneanbasin and absorbed Rome as soon as Rome hadawakened to intellectual life. The formation of


19/80

Tlze Use o Books in ncient Greece

the Alexandrian Library and of other librarieselsewhere, of which that of Pergamum to be mentioned later, is the most notable, encouraged tbeproduction of books, much as the British MuseumLibrary encourages and facilitates it to-day. The

output of books of learning, or of what desired topass as learning, was enormous. The standard ofworks of the highest literature might have fallenwoefully since the generations of Aeschylus andThucydides and Plato; but the trade of bookmaking prospered exceedingly. Commentatorscompilers, popularizers swarmed, as they do today; and it s evident that there wasa great quantityof minor literature which has disappeared with

hardly a trace.In estimating the extent of the habit of reading

in the Greek world, we have to remember that theliterature in the Greek language h ~ hhas survivedto our own day is onlya small fraction ofthat whichexisted in the three centuries on either side of theChristi an era. I t may be ofinterest to adduce evidence on this head soriIe of which is the result ofre cent discoveries.

There are two methods by which some idea couldbe obtained of the total extent of Greek literature.One is an examination of he references to lost workswhich appear in authors who still survive. I t w ouldbe a laborious, but not uninteresting or uninstructive task to compile a catalogue oflost Greek books

The Use 1 Books in ncient Greece

from tbe references to tbem in extant literature. Ican only give a few indications here, omitting theevidence ofthenewdiscoveries. VVe know that withthe exception of a substantial part of Pindarand a smaller fraction of Bacchylides, all Greek

lyric poetry has disappeared as a collected whole,and is known to us only through casual quotations.VVe know that only 7 plays of Aescbylus have survived out of at least 70, only 7 of 80phocles out of113, only 8 of Euripides out of 92, only ofAristophanes out of at least 43; and that of all theother tragic and comic poets of Greece we havenothing. In the great anthology cOlTIpiled by 8to-baeus about the end ofthe fifth century, the quota-

tions from lost works far exceed those from worksthat have survived, although thelatter are naturallythe most famous works of their respective authorsand therefore the most likely to be quoted. A roughcount shows that in the first thirty sections ofStobaeus, 314 quotations are taken from works stillextant, and I I 5 from works that are lost. Out of470 names in Photius s list of authors quoted byStobaeus, 40 at most exist in any substantial form

to-day. And this is from a collection which dra\vsfrom the best works and the best authors, and takesno account of the much larger mass of inferiorliterature, from which no quotations are takenand much of which had already disappeared atthe time when the anthology \vas made.


20/80


An earlier work which consists mainly of extractsis the Deipnosophistae of Athenaeus, which as a ragbag of quotations may be compared with Burton sAnatomy o Melancholy. have counted the quota-tions or references in a single book, and (though

cannot guarantee the absolute accuracy ofthe enu-meration) 1 find that out of 366 quotations (mainlyfrom the comic dramatists) only 23 are from worksthat have come down to uso I t is as though ofall theworks quoted in Burton, only those had survivedwhich are included in the World s Classics orEveryman s Library.

The second line ofinvestig ation into the extent ofGreek literature is to be found in an examination

ofthe fragments ofliterary works which have beenbrought to light among the papyri discovered insuch quantities in Egypt during the last sixty years.The vast majority of these fragments are derivedfrom the rubbish heaps that surrounded the townsand villages ofGraeco - Roman Egypt, and especiallythose of Oxyrhynchus. They are the debris of thebooks which the Greek-reading population ofEgyptused and possessed. They are therefore specially

valuable for our present purpose. Any scrap ofpapyrus sufficiently. large to make i t possible toascertain the character of ts text is evidence of theexistence of a complete manuscript at the timewhen it was written. I t is therefore possib le for us todetermine the proportion between the manuscripts


of works that have otherwise co me down to us andthose w i ~ have been lost. We can see, further,what authors were the most popular, and in whatcenturies there was the greatest activity in the production (and therefore presumably the study) of

books.An inventory ofliterary papyri (including under

this term the fragments of vellum manuscripts,tablets, and ostraca which have been found in thesame conditions) is that of C. H. Oldfath er, co mpiled in 1922 1 The twenty-eight years since thatdate have added appreciably to the totals, as shownbelow, but have not affected the general characterofthe results. Omitting Biblical texts and Christian

works, as forming a category apart, Oldfather Iists1,167 literary manuscripts, repre sented sometimesby the merest scraps, sometimes by substantial rollsor codices. Of this tota l, no less than 315, or morethan a quarter ofthe whole, are Homeric, 282 beingactual copies of parts of the Iliad or Odyssey while33 are commentaries, lexicons, or the like. Of theremaining 852, approxi mately 240 are from workswhichhave co me down to usotherwise, and612 are

from works wholly lost or known to us only by quotations or references. I t is fair to add tha among theseare included a number of school exercises, briefextracts, and some works which are barelyon thefringe ofliterature. Nevertheless the disproportion

I University of Wisconsin Studies, no. 9 (Madison, 1923).


21/80

3 The Use 01 Books in Ancient Greeceis rnarked, and completely confirms the conclusionsinclicatecl by the evidence ofStobaeus and Athenaeus.I t is clear tha the lost warks of Greek literature verygreatly exceeded in nurnber those which have sur-vived. Every student ofthe collections of fragments

of Greek authors will agree in this conclusion.I t is interesting also, as indicating the literary

tastes and educational practice of Graeco-RomanEgypt, to observe the distribution of the knownauthors and the dates frorn which the remains aremost numerous. Horner, as already indicated,predominates quite enormously. He was the in-dispensable subject-matter of education, and justas a knowledge of the Bible is regarded as an

essential part of the equipment of every one withany tincture of culture in this country, so it was withHomer in the Greek world. But it is noteworthythat the predominance of the _ liad over the Odysseyis just as great as the predominance of Homer overall other authors. Ofthe 8 manuscriptsofHomerrepresented in Oldfather s list, 221 are from theIliad, and only 6 from the Odyssf)'. O f the othergreat writers, Demost henes is the most fully repre-

sented, with 48 copies of one or other.ofhis orations,besides three commentaries and the more extensivework of Didymus, of which a substantial papyrusexists at Berlin. Next to hirn, as is only natural,comes Euripides, with 3 manuscripts; and afterhirn Menander, \vith 26, though the attribution of


same ofthese is daubtful. This, in view ofthe popularity of Menander and the extent to which hiscornedies lent themselves to quotation, is only whatone would have expected. Since the discovery of heCairo codex, which contains substantial portions of

four comedies, Menander may be reckoned withBacchylides, H yperides, H erodas , and Timotheusas an author who has, at least to some considerableextent, been rcstored to us from the sands ofEgypt.To thern one should perhaps add Ephorus, if, asseems probable, he is the author of the historicalwork discovered at Oxyrhynchus, and Aristotle asa historian, in virtue of the 'A87Jvalwv7To'AtTla.

After these follow Plato, with 3 manuscripts,

Thucydides with 21 , Hesiod with 2 0 (mostly fromthe Catalogues and the Theogonia, only four beingcopies of the Works and Days , Isocrates with 18,Aristophanes and Xenophon with 7 each, So-phocles with 12, and Pindar with 11. That Sapphoalso retained popularity is shown by the appearance of eight manuscripts, one of which is as lateas thc seventh century. The most noticeable gapsin the list are Aeschylus, Herodotus, and Aristotle.

Aeschylus is represented only by a single fragment,which has been doubtfully assigned to his Cariansor Europa; of his more famous works no trace hasbeen preserved. Of Aristotle there is only the'A87Jvalwv IIo'At'Tla, the Posterior Ana Jtics, and theIIpOTp'1T7'tK6s; nothing of the Ethics, the Politics,

5294


22/80

4Tlze Use oiBooks in Ancient Greece

the Rhetoric, or the Metaphysics, or of the collectionson natural history In view of the difficulty of boththese authors, it is perhaps not surprising that theydid not forln part of the curriculum of a smallprovincial community; but it is remarkable thatHerodotus, who is both easy and attractive, and hasa special interest for Egyptian readers, should berepresented only by 0 examples. Other authors ofwhom there is some substantial representation areAeschines 8), Apolionius Rhodius 8), Callimachus9 and commentaries), Hippocrates 6), and

Theocritus 6).Oldfather s list was supplemented up to the year

1945 by Laura Giabbani (Testi Letterari Greci di P r o ~

venienza egiziana, in the Pubblicazioni dell Istituto dzPapirologia Vitelii, 1947). This includes substantialadditions, made especially from British and I a l i ~ nexcavations at Oxyrhynchus. From one mound Inparticular, ofwhich Grenfell and Hunt were onlyable to re ach a small portion, but to which theItalians obtained fulier access, came a rich haul offragments, especially of Aeschylus and Callimachus.The Oxford excavators did not live to edit tp.eir

finds from this spot, some of which have beenpubli shed .by Lobel and his colleagues in 1941 and1948 (Oxy. Pap. xviii and xix). The Aeschylus textsinclude portions mostly very small) of his Niobe,Myrmidons,and Xantriae, and (more interesting fromtheir character and rather more extensive) the


satyric plays, r a f J K ~ SllOVTtEvS, L1tK71JOVAKOL p -fLTJBEVSllVpKatEvs, 8EWPOL IuBfLtauTaL. The Calli-machus fragments include several from the ALTtaand a few from other works. Other authors re-presented include Alcaeus, Hipponax, Hesiod,

Cratinus, Sappho, Euphorion, Pindar, Corinna,Lysias, and of extant works SOlne lines of Sophocles,a few of Aeschylus (Agamemnon and Septem) andmore considerable portions ofPlato s Phaedo. Apartfrom these the Homeric record is increased by143 items from the Iliad and S8 from the Odyssey;Euripides has additional items, Demosthe-nes 28, Hesiod 16, Isocrates 14, Plato 13, Sophocles13, and Herodotus 9; but except for the Aeschylus

fragments which come from a single hoard of thesecond century and are mostly by a single scribe),they do not materially modify what is said aboveabout the extent of our losses of Greek literature.The authors considerably represented are the same.The unidentified fragments in Miss Giabbani slist are much more numerous than the identifi-able, being 745 as against 2 2 2 Homeric (includingcommentaries) and 157 other known authors.

A few of_them 83 out of a total ofI

124) hadalready been included in Oldfather s list.On the whole, when it is remembered that these

papyri come mainly from the rubbish heaps of smallprovincia l towns, the range ofliterature representedmust be regarded as fairly substantial. I t shows that


23/80


Greek literature was widely current among theordinary Graeco-Roman population; that it held aprominent place in education, and that there was areading public of considerabl e size. I t can have norelation to the extent ofliterature that was available

in a great literary centre such as Alexandria, An-tioch, Athens, or the other important towns of theGreek world. Ofthis a better idea may be obtainedfrom the collections of Athenaeus or Stobaeus, andthe numerous quotations scattered about in otherGreek authors. The papyrus discoveries dispose,however of the suggestion that such compilations.were derived mainly from anthologies; for If somuch literature existed in the small towns and

villages ofEgypt, thereis

no ground for questioningthc much wider comprchensiveness of thc greatlibraries, to which scholars had access.

The distribution of the papyri in time is also instructive. I t will be understood that conclusions onthis head are necessarily precarious, partly becauseof he element of chance that attends the discoveriesof papyri, and partly because. datings of manuscripts can seldom be exact. Palaeographers differ intheir

opinions asto date, and

oftencan venture

onlyon approximate dates, such as Ist-2nd century .Still, the range of variation between experiencedpalaeographers is not very great; and if the manuscripts to which double dates are assigned aredivided equally between the centuries given as

The Use o ooks in ncient Greece 7alternatives, the results as taken from Oldfather slist, which again depends on the original publicarions) are as folIows:

3rd cent. B.C. 68 including one of the late 4 th )2nd cent. 42

Ist cent. 49I st cent. A.D. I 72nd cent. 34 13rd cent. 3044 th cent. 835 th cent. 7 86th cent. 297th cent. 3

Miss Giabbani scatalogue gives substantial ly the

same picture. The figures are approximately asfolIows:

3 rd cent. B. c. 572nd cent. 36Ist cent. 37Ist cent A.D. 692nd cent. 3 113rd cent. 2 44th cent. 825th cent. 6

6th cent. 477th cent. I I

Only approximate as these figures may be, themain results stand out unmistakably. They showthat the period of greatest dissemination of readingwas in the second and third centuries of our


24/80


era. This s the period when the Graeco-Romanoccupation of Egypt was at its height. During thePtolemaic period the infiltration of a Greek population and the assimilation of Greek culture by thenatives were steadily growing. (The higher figures

for the third century B.C. , as compared with thosefor the second and first may be accounted for bythe larger discoveries of papyri of that century,especially in the form of mummy cartonnage.) Afterthe Roman conquest the Graeco-Roman population which was mainly Greek-reading, greatly increased; and the first three cen u ries of the Empiremark the climax ofGraeco- Roman culture in Egypt.The drop that takes place in the fourth century is

very marked and s to be accounted for part ly bythe general decline of Roman civilization andpartly by the spread of Christianity which divertedattention from pagan literature. From this declinethere was no recovery until the Arab conquest inthe seventh century extinguished Christi an andpagan literature at once.

These figures of course relate only to Egypt, butthere s no reason to doubt their general applica

bility to the Hellenistic wrld. The causes whichoperated in Egypt operated also in Syria and AsiaMinor, and _may be assumed to have producedsimilar results. We are entitled therefore to drawgeneral conclusions as to the dissemination ofbooksand the practice of eading in the Hellenistic world.

The Use o ooks in ncient Greece 39During the last three centuries before Christ reekliterature was spreading over the wide regions adIninistered by the successors of Alexander. Themain centres notably Alexandria but also Antio chPergamurn and the other great cities of the Near

East were the seats of libraries and the hornes ofscholars; and Greek literature was the natural heritage ofthe Greek-speaking population throughoutthe Hellenistic kingdoms. There was a large outputof literature much of i t in the shape of commentaries and collections a good deal of t scientific andmedical. There was also a general habit ofreadingthe great works ofprevi ous ages especially Homer,and after hirn Demosthenes Plato Euripides and

Menander. During the first three centuries of theRoman Empire the same habits continued thenwith the spread and official recogniti on of Christi-anity came an abrupt decline ofhumanistic culture.Christian literature increases but pagan literaturedeclines until both alike are submerged in therising flood of Mohammedanism.


25/80

T H E P P Y R U S R O L L

N the previous chapter some account has been

given of the use of books and the practice ofreading in the Greek world from the origins of

Greek literature down to the time when the spreadof Christianity begins seriously to affect the predominance of pagan literature and Hellenismpasses into Byzantinism. So far litde has been saidof the material character and appearance of thebooks in which Greek literature was preserved. Thisis not a matter ofmerely antiquarian interest. The

extern al form of books has at all times affected andbeen affected by their contents. The materials available for writing have facilitated or impeded theoutput ofliterature. Fashion and convenience havedictated the size and shape of books and therebyhave affected the scale and character of their con-tents. Authors have planned their works to suit theprevalent scale of books or on the other hand, thescale ofbooks has been altered to meet the demand

for a particular content.As

will be shown later, thedemand for volumes containing the whole of theaccepted Christian Scriptures had much to do withthe adoption of the vellum codex as the predominant form ofbook from the fourth century onwards.Similarly in the thirteenth century the growth of

r-

J..::J

:t- Kat SupOpat. from which he professes toderive his knowledge of early e r s i ~ nhistory is tobe trusted. Coming nearer to the Greek world wehave the statement of Herodotus (v. 58) that theIonians had from antiquity ealled books SupOlpabecause onee, when papyrus was scaree, they hadmade use of goatskins and sheepskins. He adds that

even in his own time many barbarous peoples usedskins as writing materials. No doubt he would haveincluded under this head the peoples of Syria andPalestine, where we know leather to have beenregularly used. The Talmud requires all copies ofthe Law to be written on skins, and in roll form;

The Papyrus Roll 45and many examples of such rolls are in existence.In this the Talmudists were no doubt only confirming the existing and tradition al practice; andsuch evidence as exists tends to support this view.The copies of the Hebrew Scriptures which were

taken to Egypt in the third century B.C., for thepurposes of the Septuagint translation, are expressly said to have been written on S cpOpa . Thestatement in Jer. xxxvi. 3 thatJehoiakim used thcscribe s scraping-knife (Tcp [vpcp T V ypap..p,aTWS) todestroy the roll of Jeremiah s prophecics impliesthat they were written on a material stronger thanpapyrus . A knife was (as in the Middle Ages) partof a scribe s equipment for making corrections on

leather or vellum,just as a sponge was for the writeron papyrus.I t may be added that the Hebrew manuscripts

discovered in 1947 in a cave ne ar the Dead Sea arealmost all on leather. They include a nearly complete copy of the book of Isaiah, assigned by thosewho have studied i t to thc late second or early firstcentury B.C. ; and some other Biblical fragmentsare reported to be as early as the fourth century B.C.

We cannot thcrefore exclude the possibility thatworks ofGreek literature maysometimes have beenwritten on leather; but we have no direct evidenceof it, and in any case the practice can only be

Letter of Aristeas , ed. Thackeray, in Swete s Introduction tothe Old Testament in Creek pp. 5 19, 549.


29/80

The Papyrus Roll

supposed to have existed in very early times. 1 Allthe evidence goes to shovv that the one material ingeneral use in the Greek lands at least from thesixth century B a onwards was papyrus. The passage just quoted from Herodotus shows that he,writing in the middle of he fifth century, could notconceive of a civilized people using any materialother than papyrus, except under the pressure ofnecessity. All the copies of earlier Greek authorsknown to him must therefore have been written onpapyrus, and these may be presumed to have ex:-tended back for at least two or three generationsbefore his time. We are thereforejustified in takingit as certain that the use of papyrus covers at leastthe period ofthe lyric poets, and there is no reasonwhy it should not be carried back even to the beginnings ofGreek literature. 'Te know that papyruswas used in Egypt as far back as the third millennium, i f not earlier; and there is no other material,with the passible exception ofleather, which wouldhave been easily available far the Greeks. If, therefore, there was writing (as I ha ve endea voured toprove) in the days ofHomer, it is a probable corol-

I The reference to skins in a fragment of Euripides (fr. 627,Nauck 2, quoted by Gardthausen)-

, .1..8' \ .I..El.CTt.V yap, t a t OL r pat p,l\ayypa-rEts7ToAAwv ylp,ovaaL Aoglov Y1Jpvp.aTWV

cannot be taken to rest on real archaeological knowledge, hut isintended to suggest great antiquity, and implies a tradition tothis effect.

The Papyrus Roll 4 7

lary that the nlaterial used was papyrus; and quitecertainly it was the material in principal use duringthe great days of Attic literature and throughoutthe Hellenistic period.

In describing, therefore, the papyrus book weare describing the main vehicle ofliterature in theclassical world; and for this we have ample evidence, both from literary allusions and from theexistence of actual specimens.

Papyrus,2 the '\vriting material, was manufactu red out of the pith of a water-plant, Cyperuspapyrus, which in antiquity grew plentifully in thewaters of the Nile. I t was not unknown in otherparts ofthe ancient world, hut Egypt, and particularly the Delta, was the main place ofits cultivation.To-day it survives onl y in the upper reaches oftheNile, far beyond the fron iers ofEgypt, and sporadically also in Sicily and Syria. Theophrastus 3 and.Pliny4 describe it as a plant growing in 6 feet of

I The assertion of Varro, quoted by Pliny Nat. Hist. xiii. 1 I),that the use of papyrus as a material for books was only discoveredafter Alexander s conquest of Egypt and the foundation of Alexandria is negligible, for we have large numbers of Egyptian hookswritten on papyrus from about 2 0 0 0 B C downwards, and thestatement is inconsistent with the references in Herodotus andelsewhere. Pliny himself did not believe it, remarking shortlyafterwards c. 13), ingentia quidem exempla contra Varronissententiam de chartis reperiuntur .

In the following pages I have freely used two earlier articlesof my own, The Papyrus Book' The Library, 1926), and AncientBooks and Modern i s c o v ~ r i l J s(Caxton Club, Chicago, 19 2 7).

3 Hist. Plant. iv. 8. 3. 4 Nat Hist. xiii. 1 I


30/80

Tlze apyrus Roll

water or less, with a total height of as much as 5feet, and a root as thick as a man s wrist. Differentparts of it were used for different purposes-forfuel, for boats, for ropes, for sails; but the use thathas given it a world-wide reputation s that of itspith for the manufacture ofwriting materiaL

The supply of papyrus appears to have been agovernment monopoly in Egypt, and to have beenfarmed out to individual contractors. Among theTebtunis papyri s a document ofthe second centuryNo. 3 0 8 ) containing a r eceipt for 2 0 , 0 0 0 papyrus

sterns, bought from two JLtU }WTatpvpiOv Kat P ~ f O Valyta Ov [Jo>..l/LWvos fLEptOS.

The locus classicus on its manufacture s Pliny, NatBist xiii. I I 12 . The unit of manufacture was thesingle sheet K6MT)JLa). The pith having been cutwith a sharp knife into thin strips, these strips werelaid down in two layers, in one of which the fibreswere placed horizontally, in the other vertically.The two layers were then fastened together bymoisture, glue, and pressure until they formed onefabr ic-a fabric which, though now so brittle thatit can easily be crumpled into dust, probably had astrength nearly equal to that of good paper. This sshown by the fact that pumice-stone, in addition toa mallet and ivory or shell polishers, was used togive it a smooth surface. Thc central portion of thepi h was best, and was therefor e used for the highest

:I See Appendix.

PAPYRUS ROLL BEFORE OPENING

Th oll


31/80

The apyrus oll 49class of writing material; the portions nearer therind were employed only for inferior qualities. Thesize ofthe sheets in which the material was manu-factured differed according to the length in whichthe strips could be cut without weakness or fracture.The best quality was that in which the horizontalstrips were longest ; and our ancient authorities andmeasurements from existing specimens concur toprove that although specimens exist of sheets aswide as 5 inches, yet normally about 9 inches wasthe width of a sheet of he best papyrus, while thoseof more ordinary quality might measure 6 or 5inches or even less. In the Roman market differentqualities of papyrus, with their different sizes, wereknown by different names Claudia, Augusta, Livia,hieratica, amphitheatrica, Fanniana Saitica, Tae-niotica, emporetica, in descending order of merit),.but this statement of Pliny represents only Romanpractice. There is nothing to show that the sameclassification existed in Egypt, and it is impossibleto identify the several categories in the papyri thathave actually bee n found.

All that can be said is that in the best papyriin which the quality of the material is obviouslysuperior, the width ofthe sheets ofwhich the roll iscomposed is usually greater. A few examples maybe given from papyri in the British Museum. Severalofthe best Egyptian papyri have sheets ofas muchas Io i inches in width, and in some they exceed 12

5294

Th P R ll Th Papyrus R ll


32/80

The Papyrus Roll

inches. In the Papyrus of Nu the sheets actuallyreach a width of 15 inches. The Ani Papyrus, prob.ably the finest extant Egyptian book, has sheets of12 - 13 inches. The Hunefer Papyrus has sheets varying between 10 and I I t inches. O n the other hand,in the Greenfield Papyrus, which is a finely writtenhieratic roll, they are not more than 8i inches. Theheight of Egyptian papyri is often very great. TheGreenfield Papyrus measures 19 inches in height;the Harris Papyrus I is 17 inches, the Ani Papyrus15 inches, and the Papyri of Nu and Nekht I3 tinches.

For Greek papyrus ro11s the.measurements areconspicuously smaller. Probably the finest Greekliterary papyrus is a copy ofBook 111 ofthe Odyssey, Brit. Mus. pap. 27 I). This is composed of o ~ -JuiTa measuring 13 X 9 i n c h e s ~A fine manuscript atBerlin, containing a commentary on Plato s Theae-tetus, has sheets measuring I2 t X 10 inches. In

8

theBacchylides Papyrus, also a fine manuscript, theymeasure 91 X 8 or 9 inches; in the principal Hyperides MS. B.M. papp. 108 115), 12 X 10

,inches. Other exai,nples are as follows:- B.M. pap. 13 2 , Isocrates, De Pace, 11 X 7 to 81 in.Bodl. Gr. class. A. 1 P), Iliad, ii, 1 0 t X 1 0 i in.B.M. pap. 742, Iliad, ii, 1 0 X 8 i in. about).B.M. pap. 128, Iliad, xxi, xxiv, 91 X 5 to 6 in.B.M: pap. 134 . Hyperides, In Philippidem, 91 X 7 in.

Other papyri- of exceptional height are B.M. pap.

The Papyrus Roll

736 ll. viii), which is 12t inches high; P. Oxy. 843Plato, Symposium and 844 Isocrates, Panegyricus ,

both I2t inches; P. Oxy. 448 Od. xxii, xxiii), 1 1 1inches; and P. Tebt. 265 ll. ii), I I t inches; but thewid h of he o > J . . ~ l 1 a T aof hese is ei her unascertainable or unrecorded. P. Tebt. 268 Dictys Cretensis)surpasses a these in height, measuring 13 inches;but it iswritten on the back of a non-literary document, and these not infrequently exceeded thcheight measurements of literary manuscripts. Thetallest Greek papyrus known to me B.M. pap. 268)is a tax-register, measuring I5t inches in height;but the c o > J . . ~ l 1 a T aare only 5 inches wide. I t may beadded that two very carefully written petitionsB.M. papp. 354 and 177 , for which no doubt

papyrus of good quality was selected, are written onsingle sheets measuring 8i and 6 inches in wid hrespectively.

I t may be taken, therefore, as established byexperience that a papyrus sheet intended for a rollon which a work of Greek literature might be inscribed rarely, if ever, exceeded 3 X 9 inches, whilesomething like 10 X 7 would be more common fora book of moderate pretensions. O n the other hand,pocket volumes of poetry might be of much lessheight. The papyrus containing the Mimes ofHerodas is about 5 inches high, and a Hibeh papyrus ofthe third century B.C. , containing a comedy, isalmost the same. The smallest papyrus roll known


33/80

Th P R ll Th P R ll


34/80

54 The Papyrus RollP. Grenf. 4 lI. xxi-xxiii), 35 ft.P. Oxy. 224 Euripides, Phoenissae), 34 ft.B.M. 108+1 5 Hyperides, three orations, In-

complete), 28 ft.P. Oxy. 26 Demosthenes, IIpoolfLLa), c. 28 ft.P. Oxy. 27 Isocrates, I I p ~ ~ v T L 8 6 a w S ) ,25 ft.

P. Oxy. 843 Plato, Symposium), 231 ft.P. Oxy. 844 Isocrates, Panegyricus), 231- ft.P. Oxy. 16 Thucydides, Bk. iv), 23 ft.B.M. 128 li. xxiii, xxiv), 20 ft.P. Tebt. 265 li. ii), c. 19 ft.B.M. 132 Isocrates, IIpt E l p ~ v Y J S ) ,14 ft.

The British Museum Odyssey papyrus, alreadyreferred to as the handsomest specimen of Greekbook-production, would have required 7 feet if i tcontained only

Book 111,or

2 I feet if it originallyincluded ;Rooks 1-111. The Bacchylides Papyrusnow measures about 15 feet, hut we do not knowhow much is missing; and the same is the case withthe papyrus of Herodas, of which the survivingportion measures 14t feet, with a height of only 5

. inchesThe net result thenwould appear to be that 35

. feet may be taken to be the extreme limit of anormalGreek literary roll. The only two instances whichseem to require a greater length are P. Petrie 5 (3rdcent. B.C. of Plato's Phaedo and P. Oxy. 225 (Istcent.) of Thucydides, Book 11, which would each

. ~ a v e-occupied about 50 feet. Either, therefore, thebook in each of these cases occupied two rolls, or

The Papyrus Roll 55they must be regarded as exceptional. The generalrule appears to be weIl established on a wide basisofproof.

On the roll thus forrned the writing was arrangedin aseries of colurnns aA{oEs). I t is clear that

thc roll was made up before i t was written on; thescribe did not write his text on separate sheets andthen unite thern to form a roll, for the writing frequently runs over -the junct ion of the sheets. TheK O A A ~ f L a T asheets of papyrus) are therefore quitedistinct from the aAlos columns of writing). Inthe case ofpoetical texts the width ofthc column isfixed by the length of he lines. Thus in a papyrus ofthe second book of the Iliad at Oxford Bodl. Gr.

class. A. I P))written in an

exceptionallylarge

hand, the width of the column of writing is about7 inches, or 91 inches including the margin. Asimilar manuscript of the same book in the BritishMuseum B.M. pap. 742 = P. Oxy. 20) has aboutthe same measurements. In the British MuseumOdyssey the column is about 5 inches wide, or 6inches with the margin; in the Bacchylides MS.they vary between 4 and 5-i inches, includingmargin. In the earliest known literary papyrus, th cMS. ofTimotheus at Berlin, they vary between 6

and 10 inches; but this is an exceptional case, the

] The Berlin papyrus ofGenesis no. 9II ) ends at eh. xxxv. 8with the tide i'EVE ULS' KOUJLOV whieh seems to show that the wholebook was divided between two rolls.

The Papyrus Roll


35/80

The Papyrus Roll

writing being a large, heavy uncial, which takes upa great deal of room.

In prose vV'orks, where the scribe was at libertyto choose his own length ofline, the measurementsare much smaller. The following are some examples,

including margins, for which ab outt

inch must beallowed in each case:

Louvre Hyperides, 4 in.P. Oxy. 842 (Ephorus, Hellenica) , 4 in.P. Oxy. 843 Symposium), 31 in.Berlin Theaetetus commentary, 31 in. (with excep-

tionally wide margin).B.M. papa 132 (Isocrates), 31 in.P. Petrie 5 B.M. papa 688 (Plato s Phaedo) , 3 in.B.M. papp. 108+115 (Hyperides), 2 in. (mar-

gins rather wide).B.M. papa 133 (Demosthenes, Epistles , 3 in.B.M. papa 134 (Hyperides, In Philippidem), 2 in.P. Oxy. 666 (Aristotle, IIpoTp1TTtKos),I n.

In general it may be said that 3 inches or moreis exceptionally wide, 2 inches or less exceptionallynarrow, o ~ the actual column ofwriting. Between2 and 3 inche s is the normal width in a well-writtenpapyrus. I The large British Museum Hyperides,which is a good specimen, has columns of about

I Of 38 manuscripts of which the dimensions are given byMilne Gat. o Literary Papyri in the British Museum, 1927), 27 havecolumns of writing from 2 to 3 inches in width; in 5 they are lessthan 2 inches, in 6 they are more than 3 inches. Exceptionalmanuscripts, such as the }1 J1Jva{wv llOAtrda are not incIuded.

The Papyrus Roll 7


36/80

The Papyrus Roll 7

inches of writing, with about inch of margin.Literary texts written in non-literary hands, suchas the A87]valwv nOAI/Tla where one column extends to as much as I I inches of writing, cannot betaken as evidence of the normal methods of bookprod uction.

The number oflines in a column, and the num-her of letters in a line of prose, naturally dependto some extent on the size of the writing. I t will,however, be useful to givesomefigures , since textualcritics not infrequently base calculations as to conjectural emendations on estimates of the probablecontent of a line or a column. The number oflinesin a column of a given papyru s roll does not usuallyvary much, though it is seldom absolutely uniform.On the other hand there s often considerable varietyin the numher of etters in a line. Scribes made noeffort (as was the practice in medieval vellum

J It s the use of narrow columns that seems to be the point ina passage in Suetonius's Caesar (c. 56), to which my attention wasdrawn by Professor H. E. Butler:

'Epistolae quoque eius ad senatum ex.tant, quas primusvidetur ad paginas et formam memorialis libelli convertisse,cum antea consules et duces nonnisi transversa charta scriptasmitterent.

Caesar's predecessors had contented themselves with writing theirdispatches across the width of a sheet of papyrus in one broadcolumn. Caesar (writing probably at greater length) sent his in theform of a small roll, with columns of the narrow width usual inworks of prose literature. The expression is obscure to us, thoughdoubtless clear to Suetonius's contemporaries; but this sc::ems themost probable explanation.

The Papyrus Roll The Papyrus Roll 59


37/80

The Papyrus Roll

manuscripts) to enclose their column of text in aprecise rectangle. The outer right-hand) edge ofthe column was allowed to beragged and irregular.There were strict rules as to the permissible divisions of words between two lines, l a n d the scribe

would extend or reduce the number of letters in aline in order to reach the end of a word or a permissible division. If the ine was short, a small fillingmark was often used. Afewfigures will serve toshow characteristic lengths of columns and oflines,and the limits ofvariation:

Lines in Letters incolumn. line.

Louvre Hyperides (2nd

cent. B.a.)26-28 30 (27-33)

B.M. Hyperides, nPhilippidem (I st cent.) 26-28 17 16 - 19)

B.M. Hyperides, nDemosthenem &c. Istcent. A.n.) 28 (27-3 0 ) 16 (13-18)

P. Oxy. 16 (Thucydides,Ist cent.) 49-5 2 19 - 23

B.M. 13 2 (Isocrates, lateIst cent.) c 5 20-24

P. Oxy. 844 (Isocrates,2nd cent.) 39-4 1 18 (14-19)

I These are stated n my Palaeography o Greek Papyri (p. 3 1 .Z Figures outside brackets give the more usual nw:n b.er of lines

or letters, those inside brackets the observed VarIatIOns. Thelatter are not based on exhaustive examinations, so that evengreater variations may exist.

The Papyrus Roll 59Lines in Letters incolumn. fine.

P. Oxy. 843 (Plato, 2nd-3rd cent.) c 7 28 (23-32)

P. Oxy. 193+1182 (De-mosthenes, 2nd cent.) 31-34 10 - 13

Chester Beatty Deu ero-nomy (2nd cent. codex) 3 1-3 8 16 (13-19)

Chester Beatty Daniel(3rd cent. codex) 45-4 6 18 (16-19)

These figures, which might be considerably extended, show that columns ofless than 25lines are,to say the least, rare, and that ines with less thananormal 16 letters are equally so. The normalfigures may be anything betwee n 25 and 45lines to

a column,1 and ab out 18 to 25 letters to a ine.There are columns of exceptional height or smallness ofwriting which exceed these dimensions, butthey are rare, and there may be a few manuscripts

. which fall bel ow them; but even the Herodas papyrus, which is only 5 inches in height, and may betaken to represent a pocket volume of poetry, has15-19 (usual ly 18) ines to' a column. Hibeh Pap.6 a comedy), which is ofthe same height, has 3

lines to a column. I t is also clear that conjectureswhich assume a fixed number oflines to a column,

I O f 0 pagan manuscripts listed by Milne, of which theseparticulars are given, 47 have from 25 to 45 lines to a column;12 have less, II have more. Only 3 have less than 20 (all poetry),and only 6 more than 50 3 prose, 3 poetry).

6 The Papyrus Roll The Papyrus Roll 61


38/80

The Papyrus Roll

or of letters to a line, are not to be depended onwhen dealing with papyri.

The size of the margins, as in a modern book,varies with the pretensions of the book to beauty ofappearance. In the large Hyperides MS. the upper

marginis 2t

inches, the lower 3 inches with inchbetween the columns; in the Berlin Theaetetus commentary the measurements are respectively Iinches, 2t inches, and I nch. I t will be seen that theancient baak-designer realized the true proportionsof margins much as a modern book-designer does,though the upper margin is perhaps rather larger inproportion to the lower than modern taste prefers.In the humbler class ofbooks, as now, margins aremuch curtailed; and most of thepapyr i that havebeen discovered, which come from provinc ial Egypt,are of this kind. I t is only the more handsomel ywrittenmanuscripts (which would betheleastlikelyto be disfigured by additions ofthis kind) that givemuch scope for those marginal notes and additionswith which conjectural criticism sometimes makessuch free play.1

At the beginning of a roll aspace equivalent toab out the width of a column seems often to havebeen left blank,2 no doubt with the object ofgivingthe reader something to hold the roll by when read-

I Amherst Pap. ii. 13 has exceptionally wide margins betweenthe columns, apparently for the reception of scholia.

2 Examples are to be seen n the )1() 1]va.lwv IIoAL7 la. andHerodas papyri.

py

ing it, and also of protecting the text.from injurythrough accidental tearing. This space was notutilized, as might have been expected, to receivethe tide of he work. Titles, when they appe3:r at all,are appended at the end, as in early printed books.This seems to imply that when a roll had been readit was left with its end outside; and, given ordinaryhuman nature, it is more probable that this was sothan that the reader took the trouble to re-roll i t inorder to bring the beginning to the outside. Hence anewcomer would only have t o look at the exposedend of the roll to ascer tain its contents.

Certai n external additions have to be mentionedto complete the description of the papyrus book.Vve know from references in Latin literature thatin books with any pretensions to style rollers wereattached to the ends of the papyrus, and theserollers were ornamented with projecting knobs(cornua, umbilici , which might be of various shapesor colours. 2 So far as I am aware, no examples ofthese have yet been found; so they must not betaken as characteristic of he cheaper class ofbooks.In some cases, however, the ends of the roll arestrengthened by an extra thickness of papyrus; andI have seen some burnt papyrus rolls which hadquills attached to one end, to serve as rollers. In

I It should be observed that this s a personal view, not onegenerally accepted.

e ~ c ethe phrases ad umbilicum pervenire (Martial, iv. 9or exphcatum usqueadsua cornualibrum (Horace, Epod. xiv. 8).

The Papyrus Roll The Pap rus Roll


39/80

SOHle cases the roll was provided with a wrapper ofparchment membrana), to protect i t when not inuse . and these could be made ornamental bycolouring. The purpose of lettering on the back ofa modern book was served by projecting labelsaOJwoL), of papyrus or vellun1, on which the title

A book-box capsa) containing rolls with sillybi. FromAntichitd di Ercolano.

of the book was inscribed. This hung outwards asthe rolls lay on the shelves of boo

kenyon frederic g. , books and readers in ancient greece and rome 1951

Documents