sign in sign up
<<
Home , Argiletum, Basilica Ulpia, Casinum, Library of Celsus, Marcus Terentius Varro, Oppian Hill

Roman during the Late Republic and Early Empire: With Special Reference to the of Michael Kenneth Affleck

A thesis submitted for the degree of Doctor of Philosophy at The University of Queensland in 2012 School of History, Philosophy, Religion and

i

Abstract

This thesis considers the history and development of Roman libraries from the Late Republic through to the 2nd Century AD. It will begin by examining evidence suggesting that libraries existed in prior to the often-cited date of 168 BC, and then look at the steady development of private libraries through the last century of the Republic. It will discuss how these private libraries led to the creation and development of the first public libraries in the early Imperial era. As some scholars have suggested that these public libraries were of little significance, the thesis will analyse their arguments and assess the value of the Imperial libraries and how important they were, particularly to the writers of Rome. The development of private libraries in the first two centuries of the Empire, particularly those of the Imperial households, will be examined with a view to showing that they were innovative in design and reflected developments in the public libraries. There will subsequently be a particular focus on one writer of the 1st Century AD, Pliny the Elder, and by considering his sources for the and how he used them, conclusions will be drawn about the use he made both of his own library and of other resources available to him. The thesis will conclude by conducting a survey of Roman libraries according to methodologies used in modern libraries and discussing how the Romans dealt with various issues which modern libraries also face.

Michael Affleck PhD Candidate, School of History, Philosophy, Religion and Classics The University Of Queensland August 2012

ii

Declaration by author

This thesis is composed of my original work, and contains no material previously published or written by another person except where due reference has been made in the text. I have clearly stated the contribution by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical assistance, survey design, data analysis, significant technical procedures, professional editorial advice, and any other original work used or reported in my thesis. The content of my thesis is the result of work I have carried out since the commencement of my research higher degree candidature and does not include a substantial part of work that has been submitted to qualify for the award of any other degree or diploma in any university or other tertiary institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the General Award Rules of The University of Queensland, immediately made available for research and study in accordance with the Copyright Act 1968.

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright holder to reproduce material in this thesis.

iii

Publications during candidature No publications.

Publications included in this thesis No publications included.

Contributions by others to the thesis No contributions by others.

Statement of parts of the thesis submitted to qualify for the award of another degree None.

iv

Acknowledgements

My most heartfelt thanks must first go to my supervisor Dr Tom Stevenson, whose good advice, guidance and above all eternal patience with my endless prevarication are beyond my power to repay. Without him, this thesis would never have happened.

Thanks also to the participants in the Ancient Libraries conference at the University of St Andrews, Scotland, 2008, organised by Dr Greg Woolf, Dr Jason Konig and Dr Katerina Oikonomopoulou, for a tremendous experience and much encouragement that helped to shape this thesis.

I must acknowledge my parents, Alan and Margaret, who suffered more than most from the production of this work, having to endure their dining room table being stacked chest-high with papers and for years, in addition to other disruptions. For their love and support, I can only say all my love and thank you.

To my dear friends Robin and Sonja Craig, who embraced this project from the outset, encouraged me, made me resident historian to their monthly philosophy group and even allowed me to present portions of my work to the group, I say thank you from the bottom of my heart.

To my good friend Selena Trace, who came late to the project, but gave me good advice and encouragement over frequent coffees, I say thank you.

v

Keywords Private libraries, public libraries, Rome, bibliotheca, library surveys

Australian and New Zealand Standard Research Classifications (ANZSRC) 210306 Classical Greek and Roman History

Fields of Research (FoR) Classification FoR code: 2103, Historical Studies, 100%

vi

TABLE OF CONTENTS

Abstract ii

Declaration by author iii

Statement of publications during candidature iv

Statement of publications included in this thesis iv

Statement of contribution by others to this thesis iv

Statement of parts of this thesis submitted for the award of another degree iv

Acknowledgements v

Keywords vi

Australian and New Zealand Standard Research Classifications (ANZSRC) vi

Fields of Research (FoR) Classification vi

Abbreviations viii

Introduction 1

Chapter 1. Libraries in Rome from the Early Republic to the Flavian Era 5

Chapter 2. Rome’s Public Libraries: Boom or Bust? 34

Chapter 3. Private Libraries in the Early Empire 50

Chapter 4. The Library of Pliny the Elder 71

Chapter 5. A Survey of Roman Libraries 93

Conclusion 121

Ancient Sources 123

Bibliography 128

vii

ABBREVIATIONS

The entries below will not appear again in the bibliography. Ancient authors and texts are abbreviated according to the Oxford Classical Dictionary, 3rd ed., and will not be listed below.

Reference works

BMC Mattingly, H. ed. 1965-, Coins of the in the , : Trustees of the British Museum.

CIG Boeckhius, A. ed. 1977, Corpus Inscriptionum Graecarum, Hildesheim; New York: Georg Olms Verlag.

CIL Henzen, W., Huelsen, C., Mommsen, T., et al. eds. 1862-, Corpus Inscriptionum Latinarum, Berlin: Reimer.

Grilli Grilli, A, ed. 1960, Fragmenta, Milan: Istituto editoriale cisalpina.

IG Inscriptiones Graecae, 1981-, Berlin, New York: de Gruyter.

ILS Dessau, H. ed. 1892-1916, Inscriptiones Latinae Selectae, Berlin: Weidemann.

OED Oxford English Dictionary, 1989, Oxford: Clarendon Press (2nd ed.).

OLD Oxford Dictionary, 2011, Oxford: Oxford University Press (2nd ed.).

Pflaum Pflaum, H.G. 1960- , Les carrières procuratoriennes équestres sous le haut-empire romain, : P. Geuthner.

Platthy Platthy, J. ed. 1968, Sources on the Earliest Greek Libraries with the Testimonia, Amsterdam: Adolf M. Hakkert

P. Oxy The Oxyrhynchus Papyri, 1898-, London: Egypt Exploration Fund.

viii

Introduction

Between 36 and 27 AD, the Roman politician, soldier and polymath, C. Asinius Pollio, established a library in the Atrium Libertatis. Almost nothing is known today about his library, and even the location of the Atrium Libertatis itself is disputed. What is not in dispute, however, is that it was Rome’s first , ‘public’ in the sense that it was built in a public area, and ‘public’ in that it was theoretically at least open to all. It was, almost certainly, the first known library, as Pliny asserts, ‘to make the works of the property of the public.’1 In its construction and operating procedures, however, it was drawing on a long line of predecessors, from , Egypt and Greece. Even though the idea was not realised during his lifetime, the initial idea for a public library in Rome came from Julius ’s experience of the great Library of .2 Yet although the Romans drew on Greek models for their libraries, they improved on those models. With the Roman flair for innovation they developed new ideas in library construction, in layout, and in the way books were stored and organised. From their earliest days, Roman libraries showed a steady increase in sophistication. Documenting the development of Roman libraries will be one of the aims of this thesis. To develop his public library in Rome, Caesar chose , a fitting choice, as Varro was surpassed only by as Rome’s bibliophile par excellence in the last decades of the Republic. His qualifications were further bolstered by being the only known writer of a library manual in antiquity, De Bibliothecis, which unfortunately has not survived. Varro was one of three great bibliophiles of the late 1st Century, along with Cicero and Cicero’s great friend, correspondent and publisher Pomponius Atticus. Each of these men established and owned multiple private libraries in their various houses and villas. They succeeded earlier generations of library-owners who acquired their libraries by somewhat different means. It is Aemilius who is credited with bringing the first library to Rome, when he looted the royal library of Macedonia and gave it to his sons after the Battle of Pydna in 168 BC. This example was followed by and , who looted the libraries of and respectively and established them in substantial country villas for the edification of their fellow aristocrats. It was from these private libraries, rather than from Caesar’s experience in Alexandria, that the idea for Rome’s public libraries really developed. Rome’s public libraries were built on developments made in the private libraries of aristocrats of the last century of the Republic. As one scholar puts it, the public libraries of Rome were essentially the

1 NH 35.9. 2 Suet. Caes. 44. 1

Emperor’s own libraries thrown open to his friends and clients.3 Consequently, another of the aims of this thesis will be to examine the development of Rome’s private libraries, beginning well before the putative date of 168 BC, up to the end of the main period of Roman library-building in the era of and , with a special examination of the library of one particular writer, Pliny the Elder. However well-developed and sophisticated Rome’s public library system became, there are scholars who question how significant it was in the wider context of Roman society. One scholar has dismissed Rome’s public libraries almost entirely, arguing that they had no eminent scholars associated with them, nor completed any significant bibliographical work, nor even played any part in the transmission of .4 Other scholars have questioned the libraries’ significance due to the fact that they were apparently subordinate to other institutions and were never depicted on coins.5 As will be argued, however, there is compelling evidence to suggest that they were indeed significant institutions. Not least among the reasons for this view is the longevity of some of Rome’s public libraries. The Library of Palatine was established in the 30s BC. It was destroyed by fire in 363 AD, a lifespan of nearly 400 years. The Bibliotheca Ulpia in Trajan’s was established in 109 AD. It was still operating in 456 AD, a lifespan of nearly 350 years. This can be compared to the two most significant library collections in the world today, the British Library and the , both of which have had lifespans of little more than 200 years. The fact that the emperors continuously provided the funds and staffing which kept libraries going for this length of time is a key indicator of their significance. After the destruction of the first incarnation of the Library of Palatine Apollo in 80 AD, Domitian had the library rebuilt and went to the trouble of sending off to Alexandria for copies to replace the lost stock.6 These were plainly important institutions. The third and final aim of this thesis will be to demonstrate the significance and importance of these libraries, particularly to the writers of Rome. The first chapter will begin by examining the origins of libraries in Rome, going beyond the often-cited date of 168 BC to ascertain whether there is valid evidence for the existence of libraries earlier in the life of the Republic. From there the trail leads through the era of looted libraries, demonstrating the growing sophistication of private libraries from the age of Sulla and Lucullus through to the last years of the Republic and the great collections of Cicero, Varro, and Atticus, which preceded Caesar’s idea for a public library in Rome. The chapter will conclude with an examination of

3 Marshall 1976: 261. 4 Irwin 1964: 76. 5 Dix and Houston 1994: 710. 6 Suet. Dom. 20. 2 the first public libraries in Rome, beginning with Pollio’s initial effort in the Atrium Libertatis, then the first imperial libraries of and , before ending with the Flavian era, in such a way as to lead in later chapters to an examination of the library of Pliny. The second chapter will examine the value and importance of Rome’s public libraries. In response to questions posed by scholars about the importance of the libraries, it will be suggested that they were certainly significant. In particular, the chapter will look at the value of the libraries to Rome’s writers, especially using the testimony of writers such as , and about the importance of the libraries. The third chapter will evaluate private libraries of the early Empire. While in the late Republic there is copious written evidence for private libraries but almost no archaeological evidence, in the early Empire there is little written evidence but quite an amount of archaeological evidence. The chapter therefore begins with the only intact library ever discovered, the library in the Villa of the Papyri at Herculaneum. Next the analysis moves on to possible library remains in and on the in Rome, showing steady development and increasing sophistication. The chapter will then examine the private libraries of the Emperors, the palace libraries of Augustus and Domitian on the Palatine, possible libraries in Tiberius’ villa on Capri and ’s Golden House, and finally Hadrian’s Villa at Tivoli, where as many four putative libraries have been identified. It will be shown how developments in Rome’s public libraries influenced developments in the Emperors’ personal libraries. The fourth chapter continues the examination of private libraries in the early Empire by evaluating the library of one writer, Pliny the Elder. Due to the detailed information about his study methods recorded by Pliny and his nephew, the Younger Pliny, and because his voluminous Natural History includes numerous references to his sources as well as a comprehensive index, it is possible to gain impressions about Pliny’s library and how he used it. This chapter will examine his research methods, how he used his library to compile the Natural History, and ways in which he used the five authors to whom he most referred, Cato, Varro, , Homer and Juba. The final chapter will undertake a survey of Rome’s libraries utilizing survey methods employed in present-day libraries. The chapter will attempt to show the sophistication of Roman libraries and their practices relative to modern libraries. It will focus on areas such as income and expenditure, buildings and equipment, technical processes, staffing, stock and administration. It will provide a summary of Roman libraries and their practices, with a view to detailing how they coped with practical issues common to libraries of all eras, such as lighting, space, cataloguing and shelving.

3

It will also show that some practices which have only recently come back into use in libraries were in use in the libraries of Rome, for example the use of library space for purposes other than storage. The use of Roman libraries for purposes such as public recitation and meetings pioneered what is now becoming known as the use of libraries as the ‘third space’, wherein libraries are joining cafes, hair salons, bookshops and other institutions as public social spaces between home and work.

4

Chapter One: Libraries in Rome from the Early Republic to the Flavian Era

Communis opinio holds that the first library in Rome was the one procured from the royal palace of Macedonia in 168 BC by Aemilius Paullus and given to his sons.7 Certainly evidence of libraries in Rome prior to this is circumstantial at best, and few scholars are prepared to examine this possibility actively. Some, in fact, are bold enough to state categorically that for the first 500 years of its existence, Rome was without libraries of any kind.8 However, at least two scholars are prepared to posit the idea of pre-Aemilian libraries in Rome. Their theories are largely based on fragmentary and circumstantial evidence, but are of interest nevertheless. Casson argues for the existence of at least two types of library in Rome prior to the 2nd Century BC. First he cites the case of , who is alleged to have produced the first known works of Latin literature when he presented Latin renditions of Greek plays in a festival in 240 B.C.9 Casson declares that Livius must have had access to a substantial collection of Greek originals, which at this time would only have been available from Greek booksellers in the south of , and would have been beyond the resources of a freedman like Livius. Casson therefore reasons that the only possible source of Greek originals for Livius was the wealthy patron of whom Livius was formerly a slave and then a client, and who must have possessed a private library of Greek literature.10 This is logical and entirely possible, even likely. It is known that from the time of their first contacts with the Greek world in the south of Italy and beyond, the Roman upper classes had become deeply interested in Greek culture, and it seems entirely logical that they would procure, through purchase or conquest, collections of literature. This, after all, is what Aemilius Paullus did after his conquest of Macedonia, and there is no reason that it could not have happened at least a generation or two earlier, if not a century or more. There is evidence for two waves of violent contact with Greek civilization between the start of the 3rd Century B.C. and the middle of the 2nd Century. The second wave involves the Macedonian wars and the Syrian wars against Antiochus, and from this period ensued the Aemilian capture of the Macedonian royal library. The first wave involved the Pyrrhic and Punic Wars, and there seems no reason why similar capture and removal of libraries, albeit on a smaller scale, may not have happened then, without necessarily being recorded. It must be remembered that, if it were not for and Isidorus, both writing hundreds of years later, the Aemilian removal of the Macedonian

7 Plut. Aem. 28.7; Isid. Etym. 6.5.1. 8 Cf. Thompson 1940: 27; Dunlap 1972: 29. 9 Casson 2001: 61. 10 2001: 62. 5 library would also be unknown. As Frank points out, during the war in Sicily, the Roman nobility were exposed at length to the flourishing Greek culture of Syracuse and other cities, including libraries.11 It is certainly not beyond belief, and possibly quite likely, that at least a portion of these cultural riches were brought back to Rome, including the collections of Greek classics which Casson argues were required by Livius Andronicus and other early writers. Possible supporting evidence comes from , writing only a few years after the Aemilian conquest, who seems to infer that libraries were quite commonplace by his time, stating that ‘Inquiries from books may be made without any danger or hardship, provided only that one takes care to have access to a town rich in documents or to have a library close at hand.’12 Casson’s second type of library existing in the pre-Aemilian era also involves conversion of Greek plays into Latin by playwrights in the late 3rd and early 2nd Centuries, most notably by and . Casson argues that Plautus, in particular, a notably poor individual, had no access to a library of Greek originals through a wealthy patron as Andronicus did.13 Therefore, his only resource must have been collections held by theatre managers themselves.14 Once again this theory is based on rather than hard evidence. Certainly there is little doubt that managers bought plays directly from the playwright and therefore would have built up collections of plays which would have been presented periodically in the future.15 However, it seems likely, on the basis of two pieces of evidence, that the Greek originals themselves were held elsewhere. One place in which collections of plays might have been held is in the of the . Terence records that the aediles purchased Menander’s Eunuchus before it was performed.16 Clift takes this further and suggests that all plays, both originals and adaptations, were held in the state archives.17 This is a certainly a possibility, but evidence is lacking, and there is also the question, which will be examined later, of how well- developed the Roman archives themselves were at this time. A second piece of evidence suggests a more likely source of Greek plays available to playwrights at this time, with stating: ‘So when Livius Andronicus in the Second Punic War wrote a hymn which was sung by the virgins…the Temple of on the Aventine was officially granted, where the scribae and histriones might assemble

11 Frank 1929: 105. 12 Polyb. Hist. 12.26.4 (Loeb translation). 13 Plautus’ poverty is confirmed by Gellius (NA 3.3.14), who says that at one time the playwright was reduced to arduous manual labour in order to support himself. 14 Casson 2001: 64. 15 Ter. Hec. 57, Eun. 20; Plaut. Cas. 7-12; cf. Beare 1955: 154 ff; Duckworth 1952: 73 ff. 16 Ter. Hec. 20. 17 Clift 1945: 6. 6 and make offerings, in honour of Livius because he used both to write plays and set them.’18 Jory argues that this represents the first evidence of among writers and actors in Rome.19 It has in fact been suggested that this may be the genesis of that much-debated body, the Collegium Poetarum.20 Regardless of the formal existence of such a body, the possibility of collegiality amongst writers in this period, and the ramifications of collectivity, including the possession of a designated meeting-place, point to the existence of a library collection which could have been accessed by members of the guild. It defies belief that if writers were operating collegially, they would not also have ensured that members had regular access to the necessary written resources for their work. The association of Livius Andronicus with this collective gives further weight to this idea. As already noted, Andronicus probably had access to the personal library of his wealthy patron. The possibility must be considered that he used this privilege in some way to provide resources for other playwrights, either through direct donations from his patron or through the copying of works held in the family library. The possible existence of these libraries takes the history of Roman libraries back a few generations, possibly a century. However, it is possible to find evidence, circumstantial but strong nonetheless, of libraries existing in Rome centuries before Aemilius Paullus. These libraries were to be found in the temples and meeting-places of Rome’s priests and augurs. Johnson is one of the few scholars to examine this possibility, albeit superficially. He argues that from the earliest times, Roman temples must have had collections of texts, for the education and training of priests if nothing else.21 This is fair enough, but the evidence indicates that the need for priests to have libraries goes much deeper than this, being fundamentally linked to the essential nature of Roman religion itself. Being essentially a contractual relationship between the gods and man, the gods insisted on strict fulfilment of all contractual obligations. This included, as Westrup notes, the stipulation that sacred rites must be performed in a manner exactly adhering to certain ritual formulae.22 Therefore, it seems that the rituals and formulas must have been recorded in detail to preserve their exactitude. Hence, there is a strong likelihood that, from the earliest days of Roman , there must have been the need for a considerable number of texts detailing these rituals and formulae. This is not undisputed. Scheid, for instance, strongly denies that there was ever a

18 Festus 492.29-36 (translation by Horsfall 1976: 79). The date of 207 B.C is supplied by 27.37.7, although this has been disputed and an alternative date of 249 B.C. suggested. Cf. Barwick 1933: 203 ff; Mattingly 1957: 159 ff. 19 Jory 1970: 226. 20 Cf. Sihler 1905: 1 ff; Horsfall 1976: 79 ff. 21 Johnson 1970: 69. 22 Westrup 1929: 6. 7 written tradition in Roman religion and argues instead that the forms of ritual were passed down orally.23 Yet he cites little other than the evidence of silence to support his case, arguing that no remains of such texts have survived and that the fragments of alleged sacred texts that have come down to us are actually from treatises on religion.24 The fact that nothing of substance from ritual texts has come down to the present day is not surprising, however, given the cult of secrecy which was still being practised by both priests and augurs right down to the time of Cicero.25 Evidence for the existence of such ritual texts is limited, but worth examining. Livy is the main source, saying for example, ‘He chose next as Pontifex Numa Marcius…and to him he entrusted written directions, full and accurate, for performing the rites of worship; with what victims, on what days, in what temple, sacrifices should be offered.’26 It could be argued that this reference, dating to a period for which Livy himself admits records were scarce, could represent little more than myth.27 However, a later reference by Livy seems to indicate that the existence of such texts was to some extent commonplace by the 3rd Century B.C: ‘(Marcus Aemilius) issued an edict that whoever had books of prophecies or prayers or a ritual of sacrifice set down in writing should bring all such books and writings to him before the first of April.’28 This suggests that while the specific direction from Numa recorded by Livy may not have been accurate, there was a tradition of rituals being recorded in written texts. Varro also refers to the pontificii libri, which seem to have recorded the exact details of sacrificial ritual.29 Similarly, makes a reference to the libri caerimoniarum.30 Dionysius of Halicarnassus refers to Numa’s commentaries on religious ritual, which Tullius Hostilius specifically directed the priests to make available for public consumption, indicating that the priests held some type of collection of texts relating to ritual.31 The possibility therefore exists that Roman priests, from very early times onward, maintained specialized libraries of texts relating to their rituals and practices, fulfilling both an educational function and the need for exactitude in the performance of rituals. There is also considerable evidence for the priests having held collections of other kinds of texts as well. A particularly interesting reference regarding the aedileship of Flavius around 300 B.C. comes from Livy: ‘He published the formulae of the civil law, which had been filed away in

23 Scheid 2006: 18. 24 2006: 17. 25 Cic. Dom. 33; 138. On priestly secrecy, cf. Savage 1945: 157 ff. 26 Livy 1.20.5-6 (Loeb translation). 27 Livy 6.1.2 (Loeb translation). 28 Livy 25.1.12 (Loeb translation). 29 Varro, Ling. 5.98. 30 Tac. Ann. 3.58 31 Dion. Hal. Ant. Rom. 3.36.4. Cf. Livy 1.31.8; Plin. HN 13.87. 8 the secret archives of the pontiffs.’32 As Oakley notes, there is considerable speculation among scholars as to what these legal formulae actually entailed.33 Cicero elaborates by describing them as actiones.34 It seems likely that they were forms of words used for bringing a claim before a court.35 The significance of Flavius’ action has generally been seen as wresting the practice of jurisprudence away from the secretive clutches of the pontiffs, although scholars are skeptical about how secret these formulae actually were.36 However, it also seems to suggest that the priests, from an early period, held collections of texts associated with legal matters. While it is drawing a long bow to suggest that the priests held fully-fledged legal libraries, there is evidence to indicate that since the priests were the earliest legal authorities in Rome, and thus held an important advisory capacity in regards to legal matters, they would have needed access to a considerable amount of written material relating to the law. There is no explicit mention of the existence of libraries in Rome between Aemilius Paullus’ conquest of Macedonia and the Sullan era, except for the reference by Polybius mentioned above.37 However, there seems little reason to doubt that private libraries became common among wealthier Romans. The evidence indicates that in this period there developed a great interest in Greek philosophy and literature among upper-class Romans, and this continued through until the end of the Republic and beyond. This is really the beginning of what could be described as the era of otium. As defined by Cicero, the phrase cum otium dignitate, ‘leisure with dignity’, can have several interpretations, but it has been particularly identified with the idea that a Roman noble should use his leisure time for intellectual pursuits.38 In this period, for the first time, the Roman nobility can be seen to have taken an active interest in reading, writing and intellectual discussion. To all intents and purposes, this interest began with Aemilius Paullus’ looting of the Macedonian royal library and his handing over of it to his sons for transport back to Rome. It was one of Paullus’ sons, Scipio Aemilianus, who is most identified with the growth of interest in Greek philosophy and literature among the Roman upper classes, and so he makes an excellent starting point for a discussion of this period. The existence or otherwise of the ‘Scipionic circle’, an alleged clique of noble Romans surrounding Scipio Aemilianus who had a particular interest in Greek philosophy and culture, has been

32 Livy 9.46.5 (Loeb translation); Cf Posner 1972: 172 33 Oakley 1997: 609 ff. 34 Cic. De or. 1.186; Att. 6.1.8. 35 Jolowicz and Nicholas 1972: 90. Cf. Schulz 1946: 9; Stein 1966: 11 ff. 36 1946: 10. 37 See n. 6 above. 38 Cic. De or. 1.1; Sest. 98; Fam. 1.9.21. For discussion of the possible interpretations of otium, see Balsdon 1960: 46 ff. 9 a matter of some debate, although it seems clear that on at least two occasions Cicero refers to a set or group surrounding Scipio, Laelius and the philosopher Panaetius.39 In fact, if such a group did exist, it might be more appropriate to call it the ‘Panaetian circle’, since association with Panaetius seems to be the common factor linking the various members. However, even if no such clique existed, it is clear that a number of Roman nobles maintained an active interest in intellectual pursuits in general, and in the study of Greek philosophy and literature in particular. The evidence concerning Scipio Aemilianus’ interest in Greek culture is substantial: ‘And surely this community has produced no men of more splendid fame, more weighty influence or more polished manners than Africanus, Gaius Laelius and Furius, who at all times and in public had about them most accomplished personages from Greece.’40 Scipio is particularly associated with Panaetius, with whom he apparently had an intimate relationship, living with him and travelling to Greece with him.41 It seems inconceivable therefore that a man with such a deep interest in Greek culture and philosophy did not possess a library of some size. It is entirely possible that the basis of this collection was the royal library seized from Macedonia. If so, Scipio’s library could have been quite vast, certainly the largest private library in Roman hands before the 1st Century B.C. A number of other prominent Romans of the period, some connected with the ‘Scipionic circle’, others not, are attested as having literary interests. Publius Rutilius Rufus is noted by Cicero as being a man of wide reading in Greek.42 Postumius Albinus is recorded by both Polybius and Cicero as being a man of letters.43 Of particular interest is Marcus Fulvius Nobilior, who erected the Temple of Hercules Musarum in around 187 B.C, dedicated to the Nine , because of an interest in literature and his friendship for the poet .44 The literary associations of this temple are well- attested. It became a gathering-place for poets to read their works.45 The next tangible development in the history of Roman libraries occurs early in the 1st Century B.C. When the soon-to-be dictator Lucius Cornelius Sulla captured Athens in the course of his campaign against Mithridates, it is recorded that he seized and returned to Rome the library of Apellicon of Teos, a philosopher and bibliophile who had taken up arms against the Romans.46 This

39 Cic. Fin. 4.23; Amic. 19.69. For debate on the Scipionic circle, see Astin 1967: 294-306; Zetzel 1972: 173 ff; Forsythe 1991: 363 ff. 40 Cic. De Or. 2.155 (Loeb translation). 41 Cic. Tusc. 1.33.81; Acad. 2.5. 42 Cic. Brut. 114. 43 Polyb. Hist. 39.1; Cic. Brut. 81. 44 Cic. Arch. 27; Plin. HN 35.66; Eum. Rest. 7; cf. Platner 1965: 255; Richardson 1977: 355 ff. 45 Hor. Sat. 1.10.39; cf. Stambaugh 1978. 46 Strab. Geog. 13.609; Plut. Sull. 26.1; Athen. Deip. 5.214. 10 library apparently contained the accumulated library, or part thereof, of and Theophrastus.47 Sulla’s motivation for taking the library remains obscure and, according to Dix, unexplored.48 It cannot be doubted that the library was significant to him, because it is recorded as being one of only two treasures from his Greek campaign that he kept for himself.49 He had no previous record of a particular interest in Greek literature or philosophy, although he may have acquired one during the course of his campaign, as he is recorded as having spent considerable time with the book-lover and Hellenophile Titus Pomponius Atticus.50 It is possible that Sulla’s motivation may have been, as Dix argues, a desire to give his children a Greek education, as Aemilius Paullus had done.51 It might also have been a strike against the philosophers who had been prominent in the campaign against the Romans, for, as has already been noted, Apellicon himself had taken up arms against the Romans on Delos. After the library was returned to Rome, it seems to have been installed in Sulla’s villa on the Bay of Naples, although the exact location of the villa is a matter of dispute.52 It is not recorded whether Sulla himself made use of the library, but after his death it seems to have passed to his son and again come to prominence.53 Most notably, the grammarian Tyrannio obtained access to it. According to , Tyrannio was able to do so by courting the .54 This is the first record of a Roman library having dedicated staff and is evidence of the growing sophistication of the libraries being held in the private residences of Roman aristocrats. The next library to be given prominence in the literature is that of Lucius Licinius Lucullus. The most detailed description yet of a Roman library is given by Cicero, Plutarch and Isidore.55 Cicero’s description of the library of Lucullus, as located in his son’s villa at (almost certainly inherited from his father) is the first clear description of the layout and contents of an aristocratic Roman’s library: ‘I was down at my place at Tusculum, and wanted to consult some books from the library of young Lucullus, so I went to his country-house, as I was in the habit of doing, to

47 Exactly how much of Aristotle’s library, and in what quality, not to mention how Apellicon acquired it in the first place, has been a huge and long-running debate. Cf. During 1956; Keaney 1963; Chroust 1973; Lindsay 1997; Dix 2004. 48 Dix 1986: 16. 49 The other item was a gold statuette of Apollo from Delphi: Plutarch, Sulla 29.6. 50 Nep. Att. 4.1. 51 Dix 1986: 20; Plut. Aem. 6.8. 52 App. B. Civ. 1.104 records that it was at Cumae, but Val. Max. 9.3.8 and the 4th Century De viris illustribus 75.12 seem to suggest Puteoli as the location. Cumae is accepted by most because specifically mentions a villa, whereas Valerius and De. vir. ill. do not. Cf. D’Arms 1970: 30-33. 53 Cic. Att. 4.10.1. cf, Harris 1995: 56 54 Strab. Geog. 13.609. The librarian has been tentatively identified as L. Cornelius Epicadus, who is mentioned by Suet. Gram. 12. Cf. Hillscher 1892: 363 n. 8; Marshall 1983: 62 55 Cic. Fin. 3.2.7-3.3.10; Plut. Luc. 42.1-2; Isid. Etym. 6.5.1. 11 help myself to the volumes I needed. On my arrival, seated in the library I found Marcus Cato…he was surrounded by piles of books on Stoicism.’56 It is quite possible that the incident is fictional (Cicero admits elsewhere to having invented similar conversations), however, as Dix notes, the setting is almost certainly genuine.57 From this description it can be gleaned that the library was large enough for at least two people to sit in, with chairs and presumably tables. This contrasts with the only known Roman private library preserved from antiquity, the room identified as the library in the Villa of the Papyri at Herculaneum, which was small, filled with bookcases, and usually seen only as a storeroom, with books being removed to read elsewhere in the residence.58 However, there is no way of knowing how typical of Roman private libraries the Herculaneum library was. There are grounds to believe that it was not typical, as it seems to have been a relic collection, assembled during the century preceding the Vesuvius eruption and preserved without being added to by the collector’s descendants.59 Cicero’s description also gives some hint of what Lucullus’ library contained. The presence of Peripatetic and Stoic literature indicates that at least part of the collection consisted of philosophical material. This would be in keeping with Lucullus’ character as described by Plutarch and Cicero, who both mention his interest in Greek philosophy.60 This at least is entirely in keeping with the contents of the Herculaneum library, which consisted almost entirely of the writings of the Epicurean Philodemus. It seems likely therefore that the typical library of a Roman aristocrat at this time was at least partially dedicated to philosophical works. However, there are clues to other works which Lucullus’ library may have contained. Plutarch records that in his younger days Lucullus may have authored a historical work in Greek on the Marsic War.61 If this is the case, it would seem likely that his library contained historical material. Another clue to the contents of Lucullus’ library comes from the fragments of Cicero’s lost dialogue Hortensius, in which Cicero used the library as the setting. In the course of the dialogue, one of the characters asks for a list of tragedies available in the library.62 It would seem likely therefore that the library contained plays and possibly other literary works. The conclusion seems to be that Lucullus’ library was a varied and well-balanced collection, catering for its owner and his friends’ varied tastes and requirements. Another important revelation from Cicero’s description of Lucullus’ library is the apparently free-and-easy use of the collection by his fellow aristocrats. Cicero, Cato and presumably others of

56 Cic. Fin. 3.2.7 (Loeb translation). 57 Cic. Fam.. 9.8.1; Dix 1986: 74. 58 Cf. Sider 1990: 537. 59 1990: 539. 60 Cic. Acad. 2.1.4; Plut. Luc. 42.2. 61 Plut. Luc. 1.5. 62 Grilli 1960, fr. 8. 12

Lucullus’ acquaintance are able to enter the villa at will and make use of the collection as needed.63 This impression is supported by Plutarch, who specifically notes that Lucullus made his collection available to all.64 This is some of the first evidence for the network of mutual library use and borrowing of books among the nobility that would become prevalent in the Late Republic, and would eventually become a significant factor in the establishment of Rome’s public libraries. It is from Cicero that the only reference to the location of Lucullus’ villa comes. Lucullus is known to have had at least one urban residence and four country villas.65 Only the one at Tusculum is mentioned as having a library, though Cicero is known to have had libraries in at least three of his residences, so it is possible that Lucullus had multiple libraries as well. The exact location of Lucullus’ Tusculan villa has been a source of debate, but recently the remains of a particularly large residence, known as Villa 76 or Villa Centroni, have been tentatively identified as Lucullus’ villa.66 It seems likely that the majority of Roman aristocrats had their libraries in their country residences, as almost no record exists of a library in an urban residence. This is in keeping with the concept of otium cum dignitate, as expressed by Cicero, that a gentleman should spend his leisure hours in intellectual pursuits. That Lucullus ardently followed this practice is confirmed by Plutarch: ‘His libraries were thrown open to all, and the cloisters surrounding them, and the study-rooms, were accessible to the Greeks, who constantly repaired thither as to an hostelry of the Muses….Lucullus himself also often spent his leisure hours there with them, walking about in the cloisters with the scholars.’67 Plutarch and Isidore supply the only evidence as to how Lucullus acquired his library. Isidore records that Lucullus followed the well-worn path beaten by Aemilius Paullus and Sulla, in that he looted his library from a defeated enemy, in this case Mithridates of Pontus.68 This is certainly plausible, since Mithridates is known to have had wide-ranging cultural and literary interests.69 Lucullus’ known interest in literature and philosophy would surely have driven him to seek out and take any such collection from Mithridates’ capital in Sinope. However, Plutarch, in an odd comment, throws another possibility into the mix: ‘He got together many books, and they were well-written, and his use of them was more honourable to him than his acquisition of them.’70 This is Plutarch’s only reference to Lucullus’ acquisition of his library, and it seems to suggest that the library was not

63 It must be noted that Cato would have had every right to be there, as he was the guardian of the young Lucullus, cf. Varro, Rust. 3.2.17. 64 Plut. Luc. 42.1. 65 Cf. Dix 2000: 446. 66 McCracken 1942: 337 ff; cf. Ashby 1907: 121 ff. 67 Plut. Luc. 42.1-2. 68 Isid. Etym. 6.5.1. 69 App. Mith. 16.112; Strab. Geog. 13.1.155; Plut. Pomp. 37.2; Plin. HN 25.3.5. 70 Plut. Luc. 42.1. 13 acquired in one piece, but accumulated over time. Furthermore, he makes a clearly adverse moral judgement on how Lucullus acquired it. Swain argues that Plutarch is condemning Lucullus’ forced acquisition of the book from Mithridates.71 However, Plutarch had recorded similar looted acquisitions by Aemilius Paullus and Sulla without a hint of condemnation. It seems likely, therefore, that Lucullus acquired at least part of his library by some other means which Plutarch found distasteful. In contemplating a means of acquisition which an ancient source writing about Late Republican Rome might condemn, one’s mind immediately jumps to . Plutarch makes a point of noting Lucullus’ closeness to Sulla, including becoming guardian of the dictator’s son Faustus.72 Is it possible that Lucullus benefitted from Sulla’s by acquiring the libraries of proscribed men? Given Lucullus’ interest from an early age in culture and philosophy, it would certainly seem plausible that the libraries of the proscribed would be of great interest to him. Another possibility is raised by Lucullus’ guardianship of Faustus Sulla. Is it possible that Lucullus abused his position of trust in this respect to take books from Sulla’s library? There is no evidence of this, but it is possible that such an act would have inspired Plutarch’s condemnation. Questions regarding the legality of Lucullus’ acquisition of his library apply to looted libraries in general. As already indicated, Plutarch seems not to condemn Paullus and Sulla for the manner in which they acquired their libraries, but the evidence regarding the legality of generals appropriating war booty suggests that, on the face of it, they should have received the same castigation that Lucullus does. There has been a long-running debate on the subject of ‘manubiae’ and ‘praeda’, and what exactly these two terms, used by Latin writers with a degree of interchangeability, referred to. The long-accepted view is that praeda referred to the actual booty as it was seized from defeated enemies, while manubiae referred to the money gained from the sale of the booty.73 This definition, however, is now largely discredited, and there is a general acceptance that praeda refers to booty in general, particularly that either captured by the soldiers or presented to them as gifts or donatives after the battle or campaign, while manubiae refers in some way to the general’s share of the booty.74 The debate has now shifted to asking what legal rights the general had with regards to his manubiae. Shatzman argues that the general had the right to appropriate a certain amount of the booty and that

71 Swain 1990: 144 n. 106. 72 Plut. Luc. 2.1, 4.4. 73 Cf. Shatzman 1972: 177-205; Orlin 1997: 117-139; Churchill 1999: 85-116. 74 Praeda: Livy 1.55.7-9, 4.53.10, 9.37.10, 10.30.10, 10.45.14, 23.11.3, 30.45.4, 39.5.17, 45.34.1-56; Cic. Leg. Agr. 1.4.12, 2.22.59-60; Nep. Cim. 2.4. Manubiae: Livy 1.55.7, 23.11.3, 36.35.2; Vitr. 5.5.8; Frontin. Strat. 4.3.15; Plin. NH 34.93; Cic. Pis. 90; Leg. Agr 1.4.12, 2.22.59-62; Rep. 2.27.31; Dom. 38.102; Prov. cons. 11.28; Tac. Ann. 2.53; Florus 1.34.6. 14 there were no legal restrictions on his use of it. Furthermore, Shatzman stresses that no general was ever put on trial for illegally appropriating booty.75 Churchill goes to the opposite extreme and argues that manubiae were considered to be public property and the general had a responsibility to use it for the public good, and that generals were indeed tried for misappropriating booty.76 Orlin takes the middle ground, arguing that while the general had jurisdiction over the distribution of booty, there was a general expectation based on custom rather than law that manubiae would be used in the public interest. It seems likely that Orlin is correct about restrictions being customary rather than legal, given that Rullus felt constrained to propose a law, which provoked Cicero’s ire, aimed at legally restricting a general’s rights over manubiae.77 Nevertheless, there were a number of prosecutions or other actions against magistrates throughout the history of the Republic that patently involve misappropriation or wrongful distribution of booty.78 None of these cases infer that it was wrong for generals to keep a share of the booty, only that it was culpable not to give a fair share of the booty to the soldiers and the treasury. So where does this leave the looted libraries? There are four recorded cases of libraries being taken as booty by Roman conquerors. Those of Aemilius Paullus, Sulla and Lucullus have already been detailed. The fourth case relates to the removal of booty from the city of Asculum by Pompeius Strabo in 89 B.C. The Senators were apparently displeased by Strabo’s failure to share the booty.79 However they waited until after his death to prosecute his son Magnus, specifically in regard to hunting nets and books which Strabo had given to him from the booty of Asculum.80 This makes an interesting comparison with the case of Aemilius Paullus, who also gave looted books from the Macedonian royal library to his sons, but seems not only to have escaped without condemnation, but was even praised for his restraint in refusing to appropriate booty from the Macedonian conquest.81 The difference would appear to be that Strabo had failed to give a due share of booty to the treasury, and so his son was prosecuted, while Paullus had refused to appropriate a share of the booty, which was presumably then divided up between the treasury and the army, and so his acquisition of the library was allowed to pass unchallenged. In both cases, the books were incidental to the

75 1972: 177. 76 1999: 87. 77 Cic. Leg. Agr. 1.4.12, 2.22.59. 78 A non-exhaustive list would include P. Servilius , cos. 495, who was denied a triumph for not turning in booty to the treasury (Dion. Hal. 6.30.20); T. Romilius Rocus and C. Veturius Circurinus, cos. 455, prosecuted and fined for not giving a share of booty to the soldiers (Livy 3.31.4-6; Dion. Hal. 10.48-9); M. Valerius Potitus, cos. 410, prosecuted for not giving soldiers a share of booty (Livy 4.53.10); M. Livius Salinator, cos. 219, fined for unequal division of booty (Frontin. Strat. 4.1.45; Vir. Ill. 50.1); and M’. Acilius Glabrio, prosecuted for not turning in booty to the treasury (Livy 37.57.12-14). 79 Oros. 5.18.6. 80 Plut. Pomp. 4.1. 81 Polyb. 18.35.5. 15 misappropriation or otherwise of more valuable loot. This would seem to argue that libraries were not considered monetarily valuable enough to be an object of contention in the distribution of booty, except where, as in the case of Pompey, they were perhaps a convenient excuse to prosecute for the misappropriation of more valuable but perhaps not as identifiable loot. This would seem to indicate that the mild condemnation of Lucullus’ acquisition by Plutarch relates to something other than his seizure of the library from Mithridates. However, if in fact his acquisition of Mithridates’ library did in some way transgress whatever customs existed regarding the appropriation of manubiae, then it is possible that Lucullus expiated his sin by, in effect, turning his library into a public monument. As already stated, generals were inclined to use their manubiae to produce public monuments, although, as Orlin notes, it is almost impossible to determine whether this was a legal or customary obligation.82 It has long been thought that the greatest beneficiary of manubial building was the State religion, with a number of temples being vowed from the proceeds of booty. However, Orlin argues that, in fact, only a small number of temples were actually built from manubiae, and that rather more secular structures were built with these funds.83 A number of secular projects built from manubiae can indeed be found in the sources, including porticoes, aqueducts and walls.84 However, it is also possible that generals were able to use manubiae in respect of their own houses and still have this classed as public usage of the funds. This is possible because, as Wiseman notes, a Roman aristocrat’s home was at least partly public:85 ‘For men of rank, who from holding offices and magistracies, have social obligations to their fellow-citizens, lofty entrance courts in regal style, and most spacious atriums with plantations and walks of some extent in them, are appropriate to their dignity. They also need libraries, picture galleries, and , finished in a style similar to that of great public buildings, since public councils….are very often held in the houses of such men.’86 It is therefore possible that Lucullus was able to expiate any transgression in his acquisition of Mithridates’ library by, in effect, turning it into a public institution, by, as Plutarch says, throwing the library open to all.87 This is in fact confirmed by Plutarch, who emphasizes that Lucullus’ use of the library was more honourable than his acquisition of it.88 Even if Lucullus did not feel the need to expiate some transgression by turning Mithridates’ library into a public institution, he might still have done so out of sheer vanity. As Wiseman notes, Roman generals were wont to display the spoils of

82 Orlin 1997: 122. 83 Orlin 1997: 126. 84 Livy 33.27.3, 43.4.6-7; Plin. NH 34.13; Vell. Pat. 1.11.3. 85 Wiseman 1987: 395. 86 Vitr. 6.5.2 (translation Morgan). 87 See n. 58 above. 88 See n. 64 above. 16 their conquests in their homes.89 It may therefore have been Lucullus’ desire, given his interest in literature, to display Mithridates’ library as spolia, just as other Roman aristocrats displayed arms captured from enemies. Whether from personal vanity or a perceived need to expiate a transgression, if indeed he did turn his library into a public monument, this would make his library, and not that of Asinius Pollio, the first public library in Roman history, and thus a very significant step forward in the development of Roman libraries. Following on from what could be described as the century of looted libraries, the last half- century of the could be described equally well as the true period of otium cum dignitate. It was in this period that the private libraries of Rome’s aristocrats reached their apogee, with individuals owning numerous country villas, each with its own library. Three leading Romans are associated above all with this period, namely Varro, Atticus and Cicero, and it is worth examining in detail the contribution each made to the development of Roman libraries. was unquestionably Rome’s foremost expert on libraries. He produced the only known work on libraries from the Classical world, De bibliothecis, which unfortunately now exists only in a few fragments. It is no coincidence, surely, that when decided to set up the first public library in Rome, he chose Varro to undertake the task.90 There are a number of references to Varro’s great renown as a scholar. He was, for example, ‘the most learned of Romans. He composed a vast number of erudite works, and possessed an extraordinary knowledge of the Latin language, of all antiquity and of the history of Greece and Rome.’91 Ironically, given his noted scholarship and knowledge of libraries, nothing is known about Varro’s own libraries. The only mention of his library is a very brief reference by Cicero, wherein he enquires whether Varro’s library has a garden in it.92 This may give an indication that Varro’s library was of considerable size. Presumably Varro’s collection would have been quite large, probably rivalling, if not surpassing, Cicero’s own. He is known to have had villas at Tusculum and Cumae, as well as a farm at , which was later seized by Antony, but restored to Varro by Caesar’s order.93 Something of the nature of Varro’s library can be gleaned from his reference in De re rustica to fifty Greek authors on agriculture whom he has presumably consulted.94 It is a fair assumption that if his library contained so

89 Wiseman 1987: 395. 90 Suet. Caes. 44. 91 Quint. Inst. 10.1.95; cf. Dion. Hal. 2.21.2; Arn. Adv. Nat. 5.8. 92 Cic. Fam. 9.4.1. 93 Tusculum and Cumae: Cic. Fam. 9.1.2; Casinum: Cic. Phil. 103-105. Cf. D’Arms 1970: 46-47. 94 Varro, Rust. 1.8-10. 17 many works in one subject area, it is likely that his overall collection, taking into account the breadth of learning for which he was praised by and others, must have been quite sizeable. The second of the great library-owners of the Republic’s last half-century is the Hellenophile businessman Titus Pomponius Atticus. Despite being one of the best-known figures of the late Republic, Atticus is a curiously voiceless figure. None of his writings have survived, and his voluminous correspondence with Cicero is a conversation only half-heard, with Atticus a mute listener as Cicero harangues, beseeches, jokes with and occasionally berates him, his own contribution missing, perhaps deliberately so, argues Carcopino.95 Yet Atticus is known to posterity as a Roman bibliophile par excellence, second only to Cicero. While Cicero is known to have had at least three libraries and possibly as many as eight, Atticus is almost certain to have had just three libraries, though it is quite possible that they were the largest private collections of any Roman aristocrat during the Republic. That Atticus’ library resources must have been considerable is indicated by Cicero’s repeated requests to borrow books, or for Atticus to do research for him.96 Carcopino argues that Atticus did not actually lend books, citing an instance of Cicero purchasing a copy of Serapion’s Geography in 59 BC.97 Yet in 49, Cicero mentions returning Demetrius’ book On Concord to Atticus, indicating that if Atticus had ever had a policy against lending books to his friends, he had abandoned it, or was prone to making exceptions to the rule in certain cases.98 Atticus is known to have had just three definite dwelling places among his numerous properties. His main residence in Italy was apparently his house on the Quirinal, left to him in his uncle’s will, a modest dwelling according to Nepos.99 There seems little doubt that it contained a substantial library, as Cicero records browsing among books on consolation there after the death of Tullia.100 Nepos errs in asserting that Atticus owned no suburban villas. His reference to a property owned by Atticus at Nomentanum almost certainly alludes to a villa at Ficulea mentioned by Cicero.101 As Ashby notes, although Ficulea’s exact location is unknown, it was certainly on the Via Nomentana.102 No specific reference exists to a library in this villa, but it seems likely that if Atticus, and for that matter Cicero, spent any time there, it would have contained a collection. Almost certainly, however, Atticus’ most important library was not even in Italy. This ardent Hellenophile’s

95 1951: 432. 96 Book requests: Att. 8.11.7, 12.6; 9.9.2; 12.6.2; 13.8.31, 31.2, 32.2, 39.2, 44.2, 48.2. Research requests: 13.30.2, 33.2. 97 1951: 472 ff; Cic. Att. 2.4.1. 98 Att. 9.9.2. 99 Nep. Att. 13.2. 100 Att. 12.14.2, cf. 14.4. 101 Nep. Att. 14.3; Cic. Att. 12.34.1. 102 1906: 61 ff. 18 preferred home away from home was his property at Epirus in Greece. As Marshall notes, it is the only property that we know of in which Atticus took an active and continuing interest.103 It was here that Atticus created a library and literary centre he called the Amaltheum, which was lavishly praised and then copied by Cicero at his property in Arpinum.104 It is from Atticus’ Amaltheum that useful information can be gained about how private libraries during the Late Republic were furnished and decorated. Cicero records that the Amaltheum was decorated with epigrammata.105 He also praises its decoration with the statues of writers and philosophers, speaking fondly of ‘being in that niche of yours under Aristotle’s statue.’106 Pliny records that Atticus even published a book on the subject of portraits.107 This of great importance in the development of Roman libraries, as it is later recorded by Pliny that Rome’s first public library – that of Asinius Pollio in the Atrium Libertatis – was decorated with the portraits of famous writers, of whom Varro was the only living one. The voluminous nature of Cicero’s correspondence makes it possible to draw some tentative conclusions about the likely contents of Atticus’ libraries. Due to Cicero’s numerous, often repeated, requests for loans from Atticus’ collection, it is certain that his libraries contained poetry, geography, oratory, history and philosophy.108 It can also be surmised with reasonable certainty that his collection contained Procilius’ On Geography, Varro’s Peplographia, and works by Libo, Casca, Timaeus, Theopompus, Heracleides, Apollodorus, Sophocles and Polybius.109 In his dialogues, Cicero has Atticus cite Philistius, and Lysias, in addition to criticisms of Stratocles and Clitarchus.110 Considerable interest in the Latin is indicated by comments on Fabius Pictor, Cato, Piso, Fannius, Vennonius, Coelius Antipater, Quadrigarius, Sempronius Asellio, Licinius Macer and Sisenna.111 Cicero also puts effusive praise of into Atticus’ mouth.112 Furthermore, indications of the likely contents of Atticus’ libraries can be gauged from the type of works that he himself wrote. Although virtually nothing has survived, he is known to have published a work in

103 1986: 94. 104 Att. 1.13, 16, 2.1. 105 Att 1.16. 106 Att. 4.10.2. 107 NH 35.11. 108 Poetry: the poems and tales of Amalthea (Att. 1.16) and Alexander of ’ poem On Geography (2.20, 2.22). Geography: the works of Serapion (2.4, 2.6) and ’ The Descent (13.31.1). Oratory: Tyrannio’s book On Accents (12.2, 12.6) and Varro’s Orations (13.48). History: Brutus’ Epitome of the annals of Marcus Coelius Antipater (13.8) and a history by Cotta (13.44.3). Philosophy: Demetrius Magnes’ On Concord (8.11.7, 8.12.6, 9.9.2), Dicaearchus’ On the Soul (13.32.2) and ’ On the Gods and On Pallas (13.39). 109 Procilius: Att. 2.2.2; Varro: 16.11.3; Libo and Casca: 13.44.3; Timaeus: 6.1.18; Theopompus: 2.6.2; Heracleides: 15.4.3; Apollodorus: 12.23.2; Sophocles: 2.7.4; Polybius: 13.30.2, Rep. 2.2.7. 110 Brut. 42, 293; cf. Byrne 1920: 29. 111 Leg. 1.5-7. 112 Cic. Brut. 3.1. 19

Greek on Cicero’s consulship, a Annalis, a volume of portraits, a series of epigrammatic verses and a number of genealogies of aristocratic Roman families.113 The writing of the Liber Annalis accords well with Cicero’s citation of Atticus’ comments on the earlier Latin annalists. Overall, his writings indicate that he was particularly interested in the and its greatest families, and in conjunction with references to such writers as Thucydides, Polybius and Theopompus would seem to indicate that his libraries were probably well-stocked with history texts. Atticus’ other great contribution to Roman literature was as Rome’s first significant publisher, though the extent and significance of his publishing operations are now disputed. The conventional view is that Atticus’ publication business was Rome’s first large-scale publishing house, employing a huge establishment of copyists and run by Atticus with the intention of making a profit.114 The opposing opinion holds that Atticus’ publishing was a boutique operation, motivated by literary zeal rather than profit, on a small scale, largely concerned with copying and distributing the publications of Atticus’ friends.115 The available evidence can be manipulated to support either view. There is no doubt that Atticus employed a large number of copyists and other slaves engaged in literature-related duties.116 However, it is apparent that some of these slaves performed more than one duty, a fact which suggests a lack of the type of specialization which would be expected in a large-scale operation.117 In addition, as Marshall notes, Atticus no doubt employed large numbers of slaves in other parts of his business, such as agriculture, but as Cicero, the main source, is not interested in cattle-herding, for example, nothing is mentioned about them, thus giving a false impression of the actual size and importance of Atticus’ publication operation.118 Further evidence is provided by the fact that Atticus did not publish all of Cicero’s works. In fact only seven of Cicero’s books are definitely known to have been published by Atticus.119 In addition, only one work by another author is definitely known to have been published by Atticus.120 Furthermore, Cicero himself had works published by other, unnamed publishers, and this could well indicate that Atticus was not necessarily the pioneer of Roman publishing, and that his establishment was not large enough to the market.121 Nevertheless, as the first Roman publisher whose activities in that field are described in detail, and

113 Cicero’s consulship: Att. 2.1.1; Nep. Att. 18.6. Liber Annalis: Nep. Att. 18.1-2; Cic. Brut. 14-15. Portraits: Nep. Att. 18.5. Verses: Att. 1.17.5. Genealogies: Nep. Att. 18.3. Cf. Byrne 1920: 36 ff; Marshall 1986: 58 ff. 114 Carcopino 1951: 467; cf. Ogilvie 1980: 14. 115 Sommer 1926: 422; cf. Kenney 1982: 20; Marshall 1986: 125; Phillips 1986: 237. 116 Copyists: Att. 12.40, 13.44. Library-slaves: 4.5, 4.8, 13.33. Readers: 16.2. Unspecified: 5.3, 5.20. 117 For instance, Salvus, a copyist, also does readings at dinner: Cic. Att. 13.44, 16.2. 118 1986: 124. 119 Orat. (Att. 14.13); Acad. 2 books (13.32); Fin. 1 book (13.1); Lig. (13.44); De Gloria (16.2, 3, 6); 2nd Phil. (15.13, 16.11). Cf. Byrne 1920: 16 ff. 120 Hirtius’ Anti-Cato: Att. 12.40, 48. Cf. Byrne 1920: 17. 121 Att. 13.12, 21a, 22. 20 possibly as a template for the future operations of Roman publishing, Atticus is an important figure, and his operation is further evidence of the growing sophistication of literary activity in the last half- century of the Republic. However, by a large margin, the most important library-owner of the Late Republic was Cicero, the Roman bibliophile par excellence. As Reid says, ‘[h]is appetite for every kind of literature was insatiable.’122 Virtually everything of significance known regarding the private libraries of Rome’s aristocrats comes from Cicero’s letters and books. Atticus and Varro may well have possessed more books in total than Cicero, but it is doubtful whether they lived their lives as surrounded by books as Cicero. He is definitely known to have had libraries in his city residence and in at least three, and most likely six, of his eight villas, although there seems little doubt that in fact he had libraries in every single one of his known residences. There is little doubt that Cicero’s love of books began in his childhood, as he reminisces about how his invalid father devoted his life to study, and in all probability established a substantial library at Cicero’s childhood home in Arpinum.123 When Cicero inherited this house, he re-modelled it himself into a retreat suitable for the pursuit of intellectual labours. Inspired by Atticus’ Amaltheum, a library and literary centre at Atticus’ home in Epirus, Cicero set out to create his own version at Arpinum, although, as Moore notes, it is unlikely that it was ever finished.124 It was located on an island in the stream Fibrenus, near Cicero’s villa.125 Although the Amaltheum was never finished, the existence of a library somewhere in Cicero’s villa at Arpinum seems almost certain because Cicero was in residence here while he wrote some of his most important works, most notably De legibus and part of De finibus, as well as revising .126 Undoubtedly Cicero’s best-known libraries were in his homes at Tusculum and Antium. Cicero’s Tusculan villa was one of his favourites, purchased from Vettius in 68 B.C., having previously been owned by Sulla and Catulus, and it remained in Cicero’s possession for the rest of his life.127 Its exact location remains undetermined, despite a host of possibilities being proffered over the centuries.128 It certainly seems to have been a residence of considerable size, since it boasted not one, but two gymnasia. These were not gymnasia in the sense that had developed originally in Greece,

122 Reid 1885: 6. 123 Leg. 2.1-3. 124 Att. 1.16; Moore 1906: 126. 125 Leg. 2.1. 126 Att. 13.12. 127 Att. 2.1.11; Vettius and Catulus: Att 4.5.2; Sulla: Plin. NH 22.12. 128 Ashby (1910, 232 ff) argues that none of the sites match all the necessary criteria obtained from Cicero’s own description of the location, and he is supported in this by McCracken (1935, 261 ff). 21 facilities for athletics and physical training, but gymnasia of the type that developed subsequently, devoted to intellectual and philosophical study. The palaestra, originally an open area for athletic pursuits surrounded by colonnades, had evolved into a suitable location for intellectual activities. As described by , the colonnades were equipped with ‘exedrae, with seats in them, where philosophers, rhetoricians and others who delight in learning may sit and converse.’129 This concept was enthusiastically embraced by the Roman aristocracy, and the palaestra structure would become significant in the development of Rome’s public libraries. Although no substantial description of Cicero’s gymnasia survives, it can be assumed that they followed the basic Greek model of a quadriporticus surrounded by colonnades, especially since Cicero was careful to give his gymnasia Greek names. At Tusculum the upper gymnasium was known as the , the lower as the .130 Both gymnasia were apparently decorated heavily. Cicero records the purchase of statues and ornaments for the Academy, and of ‘statues of Megaric marble and Hermes figures in Pentelic marble with bronze heads’ for both gymnasia.131 In addition there were ‘bas-reliefs for insertion in the stucco walls of the hall, and for two well-covers in carved relief.’132 Not surprisingly, in view of the intellectual pursuits that took place in Greek gymnasia, it is comprehensively recorded that these gymnasia contained libraries.133 Accordingly at Tusculum Cicero included libraries in at least one of his gymnasia. The library in the Lyceum appears to have been at least large enough for Cicero and his brother to be seated in it.134 Comparison is made with the library in Lucullus’ villa, wherein Cicero and Cato were able to be seated, and with the library of the Villa of the Papyri, where the library itself was only a small room suitable for the storage of books and reading was apparently done outside in the colonnade area.135 However, given Vitruvius’ description of the Greek gymnasium as featuring exedrae in the colonnade where intellectual discussion took place, it is likely that this is what also happened in Cicero’s gymnasium, as Cicero records building ‘some new reading-rooms (exhedria) in a little colonnade at my Tusculan villa.’136 Rather than reading and discussion taking place in the library itself, it was in fact undertaken in the exedrae in the colonnade. This is almost certainly what happened in the Herculaneum villa, and it is likely to have been the case in Cicero’s villa as well.

129 De arch. 5.11.2; cf. Makowiecka 1978: 9 ff. 130 Lyceum: Div. 1.8; Academy: Tusc. 2.3. 131 Att. 1.4, 6, 8, 9. 132 Att. 1.10. 133 Platthy 1968: 28-35; cf. Makowiecka 1978, 10 ff. 134 Div. 2.8. 135 Lucullus’ library: Fin. 3.2.7; Villa of the Papyri: Casson 2001: 70 ff; Sider 1990, 537 ff. 136 Fam. 7.23.3. 22

Cicero’s other significant library was at his villa at Antium. This villa had once belonged to Q. Lutatius Catulus and was apparently purchased by Cicero before 60 B.C.137 That he had a library of some size here from early on is apparent from comments to Atticus in 59 B.C, where he attests that ‘I either enjoy myself with my books, of which I have rather a lot at Antium, or else count the waves.’138 That the library was very important to Cicero is indicated by the trouble and expense he devoted to properly sorting and cataloguing it. After having Tyrannio arrange the library, he requests Atticus to supply him with trained library-slaves (librarioli) to assist Tyrannio ‘to glue pages together….[and] get some bits of parchment to make title-pieces, which I think you Greeks call sillybi.’139 He later names these two slaves as Dionysius and Menophilus, and expresses great satisfaction with the results: ‘Since Tyrannio has arranged my books, the house seems to have acquired a soul.’140 It is evident from these passages that the maintenance of libraries was now carried out by professional slaves. This had been first hinted at with Sulla’s library, and its putative librarian Epicadius.141 However, it is obvious by Cicero’s time that multiple staff members were needed to organize and run the large sophisticated libraries owned by Rome’s aristocrats. That these practices were relatively new to the Romans is indicated by the use of Greek words for library terminology, for instance Cicero’s use of the word sillybi to describe the tags which indentified books.142 This is evidence of the increasing sophistication of Roman libraries at the end of the Republic leading in to the fully professional and diversified staffing employed in Rome’s Imperial public libraries. Cicero also provides important information about how libraries were acquired. As previously indicated, the main method of acquisition for significant libraries prior to the late Republic seems to have been from looting of conquered territories. Cicero was not a noted military man, so this obviously was not an option for him. Yet he does seem to have continued the tradition of block acquisition, albeit by purchase, gift or bequest. His earliest block acquisition appears to have been his repeated requests in 67 for Atticus to secure for him a library from Greece.143 As Dix notes, Cicero’s intention was to acquire a complete collection of Greek works already assembled.144 That acquisitions of this type were expensive is clear from Cicero’s indication that lack of funds prevented him from

137 Att. 2.1. 138 Att. 2.6. 139 Att. 4.4a. 140 Att. 2.8. Cf. Lerner 2009: 20 141 See n. 48 above. 142 This is, however, almost the only instance of the word being used in this context. Later Latin writers used either index or titulus to describe these tags.: Cic. De Or. 2.61; Martial 3.2.11, 13.3.7; Ovid, Trist. 1.7, 1.109; Pont. 1.17; Livy 38.56.6; Plin. NH, pref. 26; Quin. 2.14.4; Plin., Ep. 5.6.42; Suet. Cal. 49.3; Claud. 38.3; Gell. NA 11.16.2; cf. Clark 1909: 30 ff.; Too 2010: 54 143 Att. 1.7, 10, 11. 144 Dix 1986: 99. 23 securing the library immediately. However, it is likely that economies of scale meant that block acquisition was ultimately cheaper than individual purchases, for Cicero certainly contemplated other bulk acquisitions, most notably the purchase of the library of Faustus Sulla by auction after that individual went heavily into debt in 49.145 Cicero seems to have avoided the expense of other bulk acquisitions through receiving books by inheritance and by gifts from grateful clients.146 In addition, he cut down on costs in the acquisition of one collection by sharing the expense with his brother, with the added efficiency of avoiding duplicates in their respective collections.147 Although Cicero certainly did acquire books individually, he seems not to have favoured Rome’s fledgling book-trade as a source. References to bookshops in Rome during the Ciceronian era are scarce, and it can be presumed that bookshops were few and probably quite small.148 Cicero is certainly disparaging about the quality of the product provided by Rome’s bookstores.149 This is not surprising, as good copyists were prohibitively expensive, and small establishments probably had to make do with less qualified staff. Hence the poor quality of the copies produced.150 For this reason, Cicero and other Roman aristocrats appear to have relied on their own staff and those of their friends for quality copies. Cicero’s main sources for individual texts seem to have been either gifts from friends or borrowing texts from friends and having copies made by his own staff. A particular source among Rome’s aristocrats of the day seems to have through complimentary copies from authors who had dedicated their books to their receiver. Cicero appears to have received books in this manner from Varro, Caesar, Brutus, Claudius and Gaius Trebonius.151 Another source was receiving draft copies of new works, sent to close friends for comment and suggestions. Cicero certainly sent his work to Atticus for comment prior to publication, and he is also known to have received drafts from Caesar, Brutus, Hirtius and Aulus Caecina.152 Certainly, in Late Republican Rome, the importance of private networks in substituting for a rudimentary book-trade and the lack of public libraries cannot be overestimated. From at least the middle of the third century to the last days of the Republic, it is obvious that there was steady development in the prevalence and sophistication of Rome’s private libraries. From

145 Att. 9.11; Plut. Cic. 27.3; cf. Dix 1986, 102; Kleberg 1973, 1 ff. 146 Cicero almost certainly inherited his father’s library (Phil 2.40). He also received as a gift from L. Papirius Paetus a bequest of books from Claudius (Att. 1.20). The legal aspects of library bequests certainly exercised the minds of Roman legal experts, cf. Suet. Pers. Flac. 1; Ulp. Dig. 32.52; Paulus, Sent. 3.6.51. 147 Q. Fr. 3.4, 5. 148 Cic. Phil. 2.9.21; 55.4. 149 Q. Fr. 3.5.6; cf. Strabo, Geog. 13.1.54. 150 Seneca (Ep. 27.6) quotes the price of an experienced copyist at 100,000 sesterces; cf. Marshall 1976: 254; Dix 1986: 103. 151 Varro, Ling. 5.11; Cic. Brut. 72.253; Fin. 1.3.8; Tusc. 5.1.1; Fam. 3.4, 11, 15.21; cf. Starr 1987, 214 ff. 152 Att. 13.21.2; Fam. 16.22.1; Att. 12.40, 13.50, 15.1a; Fam. 6.7; cf. Starr 1987: 213 ff. 24 small collections of Greek plays, the libraries owned by Roman aristocrats had grown in size and sophistication into networks of substantial libraries, with professional staffing, and rooms and even buildings designed to house the collections, with the nobility supplementing any lack in their libraries with informal networks of borrowing and copying from their friends. The next logical step in the development of Roman libraries was to take the libraries from inside the private homes of the aristocrats and place them in public buildings. That this step happened almost exactly at the same time as the final collapse of the Republic and the development of the Imperial system is no coincidence. The necessary impetus that was required for the establishment of public libraries was supplied by the concentration of power in the hands of one individual. In this, the Romans followed a lead established by the Greeks four centuries earlier. It is a peculiar fact of the ancient world that public libraries, so associated in the modern world with democracy and egalitarianism, were almost always the product of dictatorships. Neither Classical Athens nor the Roman Republic, the latent democracies of the ancient world, provided the impetus for the establishment of public libraries. However, the Greek tyrants, the Hellenistic monarchs and the Roman emperors did. In this lies one of the key differences between the modern conception of the public library, a taxpayer-funded public institution open to all, essentially a product of 19th Century liberal democracy, and the Classical concept, a collection established and funded by an individual, either a ruler or a private citizen, housed in a public building, but almost certainly open only to an intellectual and literary elite. Curiously, the impetus for both seems to have been the same, namely developments in education and a subsequent rise in literacy. In the 19th Century, the development of public education and a subsequent demand for books led to the development of the public library system in Britain and the United States.153 Similarly in 6th Century Greece, rising literacy and interest in intellectual and literary matters almost certainly provided the catalyst for the first public libraries. As Makowiecka argues, public libraries could only develop within societies able and willing to use them.154 As previously stated, however, the essential difference between the rise of public libraries in the Classical World and in the 19th Century was the role of dictators in their establishment in Greek and Roman societies. Impetus for the establishment of libraries was created by the evolution of the gymnasium from a centre for athletic training into a place for education and intellectual discussion.155 Certainly there is copious evidence for the presence of a library in the Ptolemaion in Athens, and it is likely that other

153 Cf. Sesso 1968: 267 ff. 154 1978: 8. 155 Cf. Makowiecka 1978: 9 ff. 25 gymnasia elsewhere also had substantial libraries.156 Yet it seems likely that the first public library in Greece was that established by the tyrant Peisistratos in Athens in the late 6th Century B.C.157 The tyrant Polycrates of Samos established a library at about the same time.158 Certainly it was a dictator, in title as well as in fact, who was the first to broach the establishment of a public library in Rome. In 46 B.C., Julius Caesar, recently returned from battling the Pompeians in Africa, decided to ‘open to the public the greatest possible libraries of Greek and Latin books, assigning to Marcus Varro the charge of procuring and classifying them.’159 The choice of Varro to head the project was an eminently sensible one. Varro’s literary credentials were second to none and he possessed substantial library holdings of his own. There were also political considerations. Varro had recently come over to the Caesarean camp after having been a staunch Pompeian, and there is little doubt that Caesar was keen to keep him loyal by offering him a substantial project that also appealed to Varro’s love of literature. An element of gratitude was also probably present in Caesar’s choice, for Varro had recently dedicated his work Antiquitates rerum humanarum et divinarum to the dictator.160 Although Caesar’s motivation for planning the establishment of a public library is not stated, apart from being an element of his general plan for the ‘adornment and convenience of the city’, there is little doubt that he was heavily influenced by having experienced the greatness of the Ptolemaic Library in Alexandria, itself the culmination of several hundred years of Greek library development.161 It has even been argued that Caesar intended to stock his new library by removing part of the Alexandria collection to Rome.162 The evidence for this idea is tenuous, but it is entirely possible that Varro sought to obtain books from Alexandria for his collection by means of copying the desired texts. This happened about a century later, when, after the Palatine Library was damaged by fire, Domitian requested copies of books be made in Alexandria to replace those lost in the fire.163 Given that, in the Late Republic, those interested in literature and with substantial collections of their own made extensive use of copying of books owned by friends, it seems logical that Varro would have sought to obtain books in this manner from the largest library in the world, and a library which certainly held

156 Platthy 1968: 28-35. 157 Gell. NA 7.1-2; Isid. Etym. 6.3.3-5. 158 Athen. Deip. 1.3, A. 159 Suet. Caes. 44. 160 Lactant. Div. Inst. 1.6.7; cf. Horsfall 1972: 121 ff. 161 Suet. Caes. 44.Cf Canfora 1989: 67 ff. 162 Parsons 1952: 284 ff. Caesar has been blamed indirectly for the destruction of part of the Alexandria library during his stay. Either part of the Library itself or a storehouse of books was apparently burnt down during his efforts to repel the Egyptian army. It has been argued that the destroyed warehouse was where Caesar was storing the looted books prior to taking them to Rome. Cf. Plut. Caes. 49.7; Cass. Dio 42.38; Amm. Marc. 22.16.14; Makowiecka 1978: 27 ff. 163 Suet. Dom. 20. 26 copies of every Greek text that the new library would require. In any event, such speculations are hypothetical. Due to Caesar’s assassination, his library was never built. It was left to Asinius Pollio to build Rome’s first public library. Unlike Varro, Pollio was an adherent of Caesar from the start, and distinguished himself in fighting against the Pompeians and then in a campaign against the Parthini in Illyria, for which he was awarded a triumph in 39 B.C. It was from the proceeds of this campaign that he began building a public library in the Atrium Libertatis.164 As with Caesar’s project, his exact motivation is obscure, though he had shown a great interest in literature early in life, when he was an associate of Catullus, and certainly after his retirement from politics, when he arranged a programme of public recitations of literary works and became a noted literary critic. Almost certainly, as an adherent of Caesar, he desired to fulfill the dictator’s unfinished project.165 It is quite possible too that he made use of whatever resources Varro had gathered for the library, a clue to this being provided, as Merivale argues, by his placing of Varro’s bust in the new library, the only living author so honoured, according to Pliny.166 The date of construction of Pollio’s library, as well as the exact location of the Atrium Libertatis, remains unknown. A definite post quem for the library’s establishment is provided by the death of Varro in 27 B.C. Since Pliny has recorded that Varro’s bust was placed in the library during his lifetime, the library was definitely in existence before this date.167 It is probable that it was built in the mid-30s. The library’s location, however, has been a subject of much debate. The location of the Atrium Libertatis, despite a number of references in the sources, is simply not known with any degree of certainty. Early suggestions identified the Atrium Libertatis with the Aedes Libertatis, a temple erected on the Aventine in 238 B.C., but this idea has now been dismissed.168 The only reference to the location of the Atrium Libertatis comes from Cicero. Referring to the construction of Caesar’s Forum, Cicero wrote: ‘So Caesar’s friends (I mean Oppius and myself, choke if you must) have thought nothing of spending 60 million sesterces on the work which you used to be so enthusiastic about, to widen the Forum and extend it as far as the Hall of Liberty.’169 Prior to the reign of Trajan, north-west of Caesar’s Forum there was apparently a saddle of higher ground between the Capitol and the Quirinal. It has always been argued that the Atrium Libertatis was in an elevated

164 Plin. NH 35.9; Isid. Etym. 6.52. 165 Zecchini 1982: 1280. 166 Plin. NH. 7.115; Merivale 1890: II.403 n. 3. 167 Cf. Purcell 1993: 145 n. 84. 168 Lanciani 1889: 184; Pierce 1922: 71; Platner 1965: 297; Schanz 1935: 25. 169 Cic. Att. 4.16 (Loeb translation). 27 location because Livy records that in 169 B.C. the censors, having been charged with treason, ascended (escenderunt) to the Atrium Libertatis, where their offices were located, from the Forum.170 For this reason, allied to Cicero’s reference, the saddle has been the favoured location for the Atrium Libertatis. This ridge was destroyed when a massive excavation was undertaken to accommodate Trajan’s Forum in the early 2nd Century.171 Conveniently, the censors’ office was defunct by then, the last census having taken place in the Flavian period. Thus the building might have been considered expendable and demolished during the building of Trajan’s Forum. However, references to the Atrium Libertatis as late as the 6th Century have caused problems for the idea that the building was located here, since obviously in some respect the building must have survived the demolition of the likely location.172 For this reason, scholars who believe that the Atrium Libertatis was located here have had to come up with alternative explanations for the building’s survival. Some, for example, have argued that parts of the Atrium were incorporated into Trajan’s Forum.173 At least one scholar, however, has rejected the Trajanic location for the Atrium Libertatis altogether. Purcell has argued that the likely location of the Atrium was in fact the structure on the Capitol now known as the . Long identified as a vast repository for state records built by Sulla, Purcell claims that the centrality of Roman record-keeping has long been overestimated, and that instead of one large repository for the archives, records were kept in rooms attached to the individual offices of the state, and that the Tabularium represents just such a suite of public offices.174 He goes on to argue that the Tabularium was in fact the Atrium Libertatis. He argues that the ridge between the Capitol and the Quirinal was not high enough to justify Livy’s use of ‘escenderunt’, but that the Capitol itself is more likely to have merited that description.175 He also cites links between and , and associations between the Capitol and the customs involving the registration of new citizens, with which the offices of the Atrium Libertatis were involved.176 He also believes that the archival functions of the building would have fitted it out suitably for Pollio’s library, in that the Romans made few distinctions between

170 Livy 43.16.13. Among other reasons, this is why the Atrium Libertatis could not have been located on the Aventine. To ‘ascend’ the Aventine from the Forum would have involved a walk of some considerable distance and exiting the city through the Servian Walls first. In retrospect, it seems highly improbable that the important censorial offices would have been located so inconveniently distant from the Forum. Cf Posner: 1972: 179 171 Trajan’s Column marks the height of the land excavated and thus conveniently indicates the elevation of the vanished ridge; cf. Coarelli 2007: 118 ff. 172 Cassiod. Var. 5.16; CIL 6.407, 1094. 173 Anderson (1984: 26) argues, based on a fragment of the Severan Marble Plan which has the word ‘LIBERTATIS’ placed over the hemicycles of the Ulpia, that these hemicycles were recycled portions of the Atrium. Richardson (1977: 104 ff) believes that the libraries of Trajan’s Forum were actually the relocated libraries of the Atrium Libertatis, as does Callmer 1944: 156 ff. 174 Purcell 1993: 142. 175 Purcell 1993: 146. 176 Purcell 1993: 145 ff. Cf. Ov. Fast. 4.624; Livy, 45.15.1-5. 28 archives and libraries, and that therefore the building made an excellent choice for the location of Pollio’s library.177 There is some evidence that might support Purcell’s theory. If the library of the Atrium Libertatis was in fact located on the Capitol, this might explain the existence of the mysterious Capitoline library, whose destruction in 188 A.D is recorded, but which is otherwise completely unknown.178 The balance of evidence, however, seems to favour the view that the Atrium Libertatis was located in the vicinity of Trajan’s Forum. For example, the Forum seems to have taken over some of the functions of the Atrium, including the manumission of slaves and the registering of citizens.179 There is also the interesting fact that part of the Forum, as Coarelli notes, appears to be a reproduction of the Atrium Libertatis.180 The complex comprising the Basilica Ulpia and the two libraries adjoining it matches the known descriptions of the Atrium following Pollio’s rebuilding of the structure, which also appears to have consisted of a basilica as well as Pollio’s two libraries.181 This would help to explain the later evidence which hints that the Atrium Libertatis survived in some form until at least the 6th Century.182 Disputes about its location aside, there is no doubt that the library of the Atrium Libertatis was extremely important to the development of public libraries in Rome. A number of features incorporated in Pollio’s library would become standard features of all the known libraries in Rome. Makowiecka argues that as Pollio had no precedents in Rome, he had to use Greek libraries as a template for his structure.183 That certainly seems to be the case, yet in one very important respect Pollio had to be innovative. As already noted, when Caesar entrusted Varro with the task of constructing Rome’s first public library, he had specifically requested two libraries, one Greek and one Latin. As Varro’s library was never built, Pollio had to become the first ever to construct a bilingual public library. The essential nature of Roman culture, heavily dependent on Greek literature, but with its own burgeoning Latin counterpart, demanded that libraries cater equally to both languages. This was unprecedented in the Greek world. It is known that the incorporated Greek

177 Purcell 1993: 144 ff. 178 Oros. 7.16; Jer. Chron. 2.174. It has, however, also been suggested that the Capitoline library was in fact Hadrianic, possibly associated with the Athenaeum. Cf. Boyd 1915: 19 ff; Platner 1965: 56. 179 Sid. Apoll. Carm. 2.544 ff; Hist. Aug. Hadr. 7.6; Vopisc. Aur. 39.3. 180 Coarelli (2000: 7 ff) argues that the religious, political and historical significance of the Atrium would have required Trajan to make some restitution for its destruction, hence the possibility that he ordered the Atrium to be reproduced within the Forum complex. 181 The existence of a basilica in the Atrium Libertatis after Pollio’s refurbishment and enlargement of the complex is surmised from the discovery of an inscription which mentions a ‘Basilica Asinia’: CIL 6.4330; cf. Coarelli 2000: 8; 1993: 170. 182 Of some interest in this context is a later inscription which refers to the Forum of Trajan as the ‘Atria Traiani’: CIL 6.9446. 183 Makowiecka 1978: 28 ff. 29 translations from other languages, but they would have been, at best, a tiny fraction of the total.184 No evidence exists for any other library in the Greek world including any more than a token component of foreign language texts. Neither is there much evidence to suggest that Rome’s private libraries during the Republic were linguistically demarcated. Certainly the looted libraries of Paullus, Sulla and Lucullus would have been almost wholly Greek, and it is only the self-acquired libraries of literary- minded aristocrats such as Cicero, Atticus and Varro in the last half-century of the Republic that are likely to have contained sizeable amounts of Latin literature. The only evidence suggesting that private libraries had segregated Greek and Latin sections comes from the Villa of the Papyri at Herculaneum.185 The other major innovation of the Atrium Libertatis library, as recorded by Pliny, is that it was the ‘first library to make men’s talents public property.’186 For the first time, an author’s works would be available, without purchase, to others who were not necessarily his friends or members of his intimate circle of book-sharing. Roman authors were relatively quick to perceive the value of this. Ovid, languishing in exile on the Black Sea coast, laments that his books have been banished from Rome’s public libraries as well, thereby depriving him of valuable readership. Referring specifically to the Atrium Libertatis, he says, ‘Nor did Liberty allow me to touch her halls, the first that were opened to learned books.’187 In this context, Ovid mentions the next two public libraries to open in Rome after Pollio’s library. However, whereas Pollio had established his library in the very last days of the Republic, where, theoretically at least, the private citizen maintained some power, the next two libraries were built at the very beginning of the Imperial era, and symbolically represent the way things had changed in Rome. They were established and run under the aegis of the Emperor, and these were the terms under which all future public libraries in Rome would operate. Augustus was responsible for the establishment of the first public library built under the Imperial system. It was established in association with the Temple of Apollo on the Palatine, and came to be known variously as the Bibliotheca Apollinis, Bibliotheca Palatina or Bibliotheca Templi Apollinis in Palatio.188 The Temple itself was vowed by Augustus in 36, following his victory over

184 The Septuagint is the best-known of the foreign holdings of the Library, but the writings of the native Egyptian Manetho and the Babylonian Berossus, as well as Zoroastrian texts from Persia, were also held by the library. Cf. Parsons 1952: 175 ff; Fraser 1972: 330 ff; Barnes 2000: 67. 185 The library found in the Villa of the Papyri was wholly Greek, though the discovery of one case of Latin texts in the Villa has led to suggestions that another separate Latin library, as yet undiscovered, is located somewhere within. As the interests of the putative owner of the library seem to have been almost entirely confined to Greek philosophy, however, it is possible that the Latin texts found were the only ones held, and that no such Latin library exists. 186 Plin. NH 7.115. 187 Ov. Tr. 3.1.71-2. 188 Fronto, Ep. 4.5; Suet. Aug. 29; CIL 6.5188. 30

Sextus Pompey in Sicily and was dedicated in 28.189 However, it was not until at least 24 B.C. that Augustus ‘added porticoes with Greek and Latin libraries.’190 Like the Atrium Libertatis, the location of the Palatine Apollo complex has been disputed. It was always assumed, based on the literary evidence, that it was located on the north-east side of the Palatine.191 However, archaeological investigations have revealed that the largely Domitianic structures on the south-west match the descriptions of the Palatine Apollo complex, and are likely to be reconstructions or embellishments of the Augustan works.192 These structures include fragments of a building consisting of two identical chambers set side by side with a prominent recess in the centre rear wall of each chamber and smaller niches, 18 in number, set into the side walls, with a raised podium beneath broken by flights of steps.193 These remains have been identified as part of a library of the Domitianic era and the evidence indicates that it was built on the site of the Augustan Bibliotheca Apollinis.194 It is known that the library was damaged or destroyed in the fire of 64 A.D., and that Domitian restored the library and ordered replacement books from Alexandria, so there is every chance that the Domitianic library replicates the structure originally built by Augustus.195 What remains therefore of Rome’s first Imperial library demonstrates the further development of library construction and practice. The bicameral structure pioneered by Pollio is carried on, as indeed it would be in all subsequent Roman libraries. The presence of niches and a podium is a significant development in library construction. Niches for the storage of books are unknown from Greek libraries, and it has been postulated that they were an Egyptian idea, borrowed from New Kingdom architecture for the Library of Alexandria, and that subsequently the idea was imported to Rome.196 However, Makowiecka argues that such an exotic origin is unnecessary, and that the niches were a purely Roman idea, made possible by the use of masonry and brick, and that niches of this sort were common in Roman architecture in general.197 The podium running beneath the niches is another feature which would be replicated in later libraries, and there is little doubt that it was intended to

189 Vell. Pat. 2.81.3. 190 Suet. Aug. 29; Prop. 2.31. However, Richmond (1914a: 201) argues, based on a reading of Eleg. 3.1, that the libraries were not ready until 23 or 22 B.C. 191 Verg. Aen. 8.98; Prop. 4.1; Ov. Tr. 3.1; cf. Huelsen 1928: 66 ff. Bishop (1956: 190 ff) argues that Tac. Hist. 1.27 and Mart. 1.70 also point to a north-eastern location. 192 Cf. Richmond 1914a: 193 ff; Platner 1965: 18; Bishop 1956: 187 ff. 193 Cf. Makowiecka 1978: 30 ff; Strocka 1981: 309; Casson 2001: 82. 194 Richmond 1914a: 204 ff; Coarelli 2007: 142 ff. 195 Suet. Dom. 20. 196 Wendel 1949: 412 ff. 197 Makowiecka 1978: 33. 31 make access to the shelves easier. However, unlike the niches, it is likely to have been borrowed from Greek architecture.198 Much more is known about the Bibliotheca Apollinis with regard to its collection than the Atrium Libertatis. Only poetry is known for certain to have been held in Pollio’s library, although its contents would certainly have been more varied than that. Poetry was also a strong feature of the Bibliotheca Apollinis, as Horace, Ovid and Martial attest. It is known from a scholiast of that the library also contained ‘books of civil law and liberal studies.’199 From the letters of Fronto it can be seen that the library contained orations from the Republican era.200 Another possible inclusion in the library was the collection of revered , though it is not certain whether these were held in the library itself, only that they were in the Palatine Apollo complex.201 The second Imperial library established in Rome cannot be wholly credited to Augustus, although he was certainly involved in its construction. It was established by his sister Octavia within the in memory of her son Marcellus, and was built on the northern side of the , near the . Although some remains of the Porticus are extant, no trace of the library remains.202 The exact date is unknown, but since Marcellus died in 23 B.C, it cannot have been earlier than that. Credit for building the library has been disputed between Augustus and Octavia, and some of the literary evidence appears contradictory, but as Richardson argues, it is clear from Plutarch and Pliny that Augustus was responsible for refurbishing the old Porticus Octavia (erected c. 168 B.C) and naming it after his sister, but that Octavia established the library within it.203 Like the Atrium Libertatis and Biblliotheca Apollinis, it consisted of Greek and Latin libraries.204 Nothing is known of its collection apart from Ovid’s reference to it as the third library from which his works are barred.205 The next two public libraries built in Rome were both constructed in the reign of Tiberius. During his reign, Tiberius had a temple constructed in honour of the Divine Augustus on the Palatine.206 The first definite references to a library in this temple are provided by Pliny and

198 The remains of the library at Pergamum have a podium running beneath the shelf area, although its use is disputed. Cf. Makowiecka 1978: 15 ff. 199 Schol. ad Juv. 1.128; Wessner 1941; Lerner 2009: 22 200 Fronto. Ep. 4.5. 201 Suet. Aug. 31. 202 Cf. Olinder 1974: 103 ff; Harris 1995: 57; Coarelli 2007: 271 ff. 203 Plut. Marc. 30.6; Plin. NH 34.31; Richardson 1976: 63. 204 CIL 6.4433, 4435. 205 Ov. Tr. 3.1.69-70. 206 Some of the sources credit the construction to Livia, while others say that Tiberius began the temple, but that Caligula completed it. See Plin. NH 19.94; Vell. Pat. 2.130.1; Tac. Ann. 6.45; Suet. Tib. 47, Cal. 31; Cass. Dio 52.10.2, 56.46. 32

Suetonius.207 The Bibliotheca Domus Tiberianae is credited to Tiberius, although no references to it exist earlier then the 2nd Century.208 As the name suggests, it was attached to the Palace of Tiberius and its remains have previously been identified with one of the apartments adjacent to the central court.209 Recently, further evidence has come to light in the form of a rediscovered manuscript by . Writing of the destruction by fire in 192 of the libraries on the Palatine, Galen refers to the Domus Tiberiana library and mentions that prior to its destruction its contents were already deteriorating through theft and damage due to the marshy location of the library.210 This has led to some debate about its location.211 It is presumed that, like the Bibliotheca Apollinis and Bibliotheca Porticus Octaviae, the libraries of the Templum Divi Augusti and Domus Tiberiana were divided into Greek and Latin sections, although no firm evidence exists. The end of the reign of Tiberius brought a hiatus in the construction of public libraries by the Emperors. No libraries are recorded from the reigns of Caligula, Claudius or Nero. It would not be until the reign of Vespasian that the next public library would be erected in Rome. By the 2nd decade of the 2nd Century, however, there were at least 7 imperial libraries operating in Rome. The next chapter will examine what, if any, impact these libraries had on the writers of Rome, and how the libraries affected the manner in which writers wrote their books, published them and made them available to their readers.

207 Plin. NH 34.43; Suet. Tib. 74. Earlier references to the library of the ‘new temple’, believed by some to refer to this library, are provided by Martial (4.53, 12.3), but this is not universally accepted. Cf. Boyd 1915: 12-13. 208 Fronto, Ep. 4.5; Gell. N.A. 13.20.1; Vopisc. Prob. 2.1. 209 Lanciani 1889: 144-47; Boyd 1915: 15. 210 Gal. Peri Alupias 18-19. 211 On the basis of the Galen manuscript, Tucci (2008: 144) has identified the location of the Domus Tiberiana library as lying between the northern corner of the Palatine and the Forum. Jones (2009: 394 ff) argues that the marshy and low-lying land Galen is referring to is actually in Antium, and therefore gives no clue to the location of the Domus Tiberiana library. 33

Chapter Two: Rome’s Public Libraries: Boom or Bust?

It was shown in the last chapter how Rome’s libraries developed from primitive beginnings in temple archives, playwrights’ collections and modest private libraries, through steadily more sophisticated and extensive private collections, with professional staff, to the first public libraries, initially the work of private citizens, then of the Emperors themselves. By the time the first century of the current era dawned, Rome had at least two Imperial libraries, housed in public spaces, run by professional staff, and, at least theoretically, open to all. Within the next 70 years, at least three more Imperial libraries were constructed. By the beginning of the second century, Rome had a public library system unequalled anywhere else in the ancient world, and the like of which would not be known again in the West before the 19th Century. Yet, for all the resources which the Emperors undoubtedly poured into these edifices, a nagging doubt remains about whether they were of any real significance in Rome’s public life, and especially in respect of Rome’s literary culture. It could be expected that these libraries would have been seen as a boon by Roman writers, with resources available for use by all, not bound by ties of patronage or friendship with wealthier individuals who possessed their own large collections. Yet, in reality, Rome’s writers damn these libraries with faint praise. Not only are there comparatively few mentions of the Imperial libraries in Roman literature, but we possess almost no details of how they were run, how they were used, or even who used them. The possibility must be considered that these libraries were just not as significant to Rome’s literary culture as has been assumed. This chapter will examine Rome’s public libraries, in an attempt to ascertain whether they really made a difference and whether they constituted a real advantage for Roman writers in the early Empire as compared to their counterparts in the Late Republic, who could only rely upon their own private collections, which were of necessity restricted by their financial circumstances, or those of friends or patrons. Dix and Houston raise concerns about the significance of Rome’s public libraries, noting that no library in Rome was an independent structure, but that all were simply subordinate parts of larger complexes.212 They argue that this may indicate that the public libraries in Rome were not considered significant enough to be established as institutions in their own right, but were afterthoughts attached to other existing or newly built structures. It is interesting to note that the known libraries elsewhere in the Roman world were free-standing structures. The Libraries of Pantainos and Hadrian in Athens, the Library of Celsus in Ephesus, and the library in Timgad in North Africa all appear to have been

212 1994: 710. 34 independent institutions.213 It can also be assumed, though not with absolute certainty, that the library pledged by to the town of Comum was an independent structure, judging from the pattern established by the private benefactors of the Pantainos, Celsus and Timgad libraries.214 That this is not the case in Rome is self-evident. For whatever reason, it seems that nearly all the Imperial libraries were established as subordinate parts of other institutions. The Bibliotheca Apollinis, Bibliotheca Templi Divi Augusti and Bibliotheca Pacis were associated with temples, the Bibliotheca Domus Tiberianae was, as its name suggests, part of the Imperial palace complex. The Bibliotheca Porticus Octaviae was part of a complex that included a and two temples. Even the largest and most influential of the Imperial libraries, Trajan’s Bibliotheca Ulpia, was simply part of his Forum complex, sandwiched between the eponymous column and the Basilica Ulpia. However, it must be noted that simply because the libraries were structurally associated with larger institutions does not mean that they were administratively subordinate. On the contrary, the evidence indicates that the Emperors took care to establish and maintain a distinct administrative structure to oversee and run the libraries. There is ample inscriptional evidence to show that specialized functionaries were employed to work in the libraries and that a distinct hierarchy of administrators supervised them.215 The reason why the Imperial libraries were structurally subordinated to other institutions probably comes down to a mixture of Roman practicality and existing precedents. As was shown in the previous chapter, temples in Rome had a long history of archival storage and this would seem to indicate a precedent for attaching libraries to the temples. In some cases, there may also have been religious motives, with libraries being attached to the temples of gods who had particular affiliations with cultural pursuits – Palatine Apollo being a very good example. On the other hand, the libraries outside Rome seem almost wholly to have been private ventures, Hadrian’s library in Athens being a rare exception. For private benefactors, there would be no benefit in attaching their libraries to any existing structure, if for no other reason than that it would diminish the personal glory of the philanthropist, which was undoubtedly a major motivation for donating the library in the first place. On the whole it seems very difficult to say that being structurally subordinate indicates that Rome’s Imperial libraries were of less significance than they have been given credit for. Other possible indicators need to be examined before this conclusion can be drawn.

213 Cf. Sisson 1929: 50 ff; Pfeiffer 1931: 157 ff; Callmer 1944: 172 ff; Johnson 1980: 651 ff; Strocka 1981: 320 ff; Casson 2001: 112 ff; Burrell 2009: 78 ff. 214 Plin. Ep. 1.8; CIL 5.5262; cf. Dix 1996: 85 ff. 215 Cf. Boyd 1915: 41 ff; Houston 2002: 139 ff. 35

Dix and Houston point out as a further indicator of the relative insignificance of the public libraries that none of them appear on coins.216 To some extent this is a corollary of their previous indicator: as subordinate parts of other institutions, the libraries would hardly be depicted on coins in their own right. What should appear on coins, therefore, is the parent institution. In fact, to some extent, the parent institutions do not appear very often on coins either. Of all the institutions which housed Imperial libraries, the one which most frequently appears on coins is the Templum Divi Augusti, which is attested on coins of Caligula after its dedication in 37 and on coins of Antoninus after its restoration in 157.217 Although the Bibliotheca Ulpia does not appear on coins, various other constituent parts of Trajan’s Forum are depicted, most notably the eponymous Column and the Basilica Ulpia.218 The Forum itself as an institution is depicted on coins which show its monumental entrance.219 The Domus Tiberiana may be depicted on a sestertius of Domitian, although there is some doubt about the coin’s authenticity.220 No coin depictions of the Porticus Octaviae or the Templum Pacis are known, though it has been suggested that a coin of Caligula which was thought to depict the Templum Divi Augusti may in fact be a representation of the Templum Apollinis Palatini. Richmond argues that coins of Tiberius from 34-36 show the completed but undedicated Templum Divi Augusti as a Corinthian temple, while the temple on the coin of Caligula is Ionic.221 He also points out that Vitruvius describes the Templum Apollinis Palatini as diastyle, with widely separated columns, a trait which matches the depiction on the Caligula coin.222 He cites further literary evidence of the Templum Apollinis Palatini having sun chariots and archaic Victories on its pediment, which are visible in the coin depiction.223 Yet although Richmond makes an interesting case, it cannot be proven conclusively that this is the Templum Apollinis Palatini, if for no other reason than the fact that neither temple is extant to provide solid proof one way or the other. Furthermore, Bennett makes the point that depictions of buildings on Roman coins cannot be assumed to be accurate representations, for the idea being represented by the building’s depiction was more important than how it looked on the

216 1994: 710. 217 BMC 41-43, 69; 549-51, 916, 1652, 1718. 1729, 2051, 2063-66, 2070, 2072, 2077, 2079, 2098. 218 It is possible that the Bibliotheca Ulpia, arguably the most likely of the Imperial libraries to be considered worthy of depiction on a coin because of its size and durability, was simply not suitable because its layout, with twin libraries facing each other adjacent to the Column, did not lend itself to effective depiction on coins. Column: BMC 449-50, 451-5, 522-4, 565-8, 971-2, 993-5, 1003-5, 1016, 1024; Basilica: BMC 492, 982-3. 219 BMC 509-10. 220 Nash (1961: 1.371) regards it as genuine, but Hill (1989: 102) argues that it is probably a forgery, though he also indicates that a medallion showing the same depiction but cast from a different die may well be genuine. 221 1914: 203 ff 222 Vitr. De arch. 3.3.4. 223 Prop. 31.11; Plin. NH 36.13. 36 coin.224 It could well be that details such as Ionic or Corinthian styles, or diastyle representations, were simply not very significant in coin depictions. While it is obvious that no library is depicted on a Roman coin, and that even representations of the parent institution are not particularly common, it is interesting to note that libraries are not the only Roman institution to be neglected in coin depictions. Two excellent examples of this are supplied by baths and theatres. No-one can doubt that baths were extremely important to the Roman way of life, and were the subject of huge expense by emperors, much more than was ever spent on libraries. The and Diocletian, for example, are among the largest and most expensive structures ever built under the Imperial aegis, perfect subjects, it might be thought, for numerous coin depictions. In fact, neither of those two structures is known to have been depicted on coins. The only Roman bath-house definitely attested on a coin is the Alexandrinae, which appear on silver and bronze coins of Alexander Severus, who restored and enlarged the Thermae Neroniae and renamed them after himself.225 Similarly, theatres, which were a cultural institution like libraries and of which a number were built by emperors, are another institution whose depictions are few and far between on coins. A coin of Gordianus struck in Bithynia, and depictions of successive wooden theatres erected in the Campus Martius upon coins of Domitian and Severus, are rare examples.226 Therefore, if libraries can be considered to have been insignificant because they were not depicted on coins, then these other institutions can be considered similarly insignificant. In fact, it is simply not possible, based on these superficial external indicators, to say how significant Rome’s Imperial libraries were. The libraries’ importance can only be determined by examining them from the viewpoint of the people who used them, the readers, and, perhaps more importantly, the writers of Rome. Much more damning is Irwin’s complete dismissal of the public libraries as institutions having any meaningful importance in the Roman literary field whatsoever. He points out that no eminent scholars were associated with them, that they produced no significant bibliographical or editorial work, nor did they have any noticeable effect on the development of Latin literature or any part in the transmission of Latin classics to posterity.227 However, Irwin’s criticism of the Roman public libraries is almost entirely based on a comparison with the Library of Alexandria, which is to some extent invalid, because the purpose and set-up of the different library systems were entirely different. The Library of Alexandra was an adjunct to the Museion, a cloistered community of the best scholars in the

224 1999: 151 ff. 225 Stevenson 1964: 790. 226 1964: 786; Hill 1989: 46. 227 1964: 76. 37

Hellenistic world, aptly described by Timon of Phlius as a ‘bird-cage of the Muses.’228 It was in no sense a public library, even in the limited sense in which Rome’s Imperial libraries merited the title. It is unlikely that anyone other than scholars, professional staff, members of the royal household and important visitors such as Caesar ever gained access to the Library. Even though Caesar himself took his inspiration from Alexandria in developing his own idea for a library in Rome, there is no evidence that he intended it to be a similarly elitist institution. Indeed states that he intended to ‘throw open to public use as many libraries as possible.’229 In any case, Caesar’s library was never built, and the template upon which Rome’s libraries were built was established by Asinius Pollio in the Atrium Libertatis, which, as Pliny indicates, was explicitly intended for public use.230 None of the subsequent Imperial libraries shows any indication of having been intended to form part of a cloistered community of scholars, although of course scholars would have been the heaviest users of the libraries. In reality, a comparison between the Library of Alexandria and the Roman libraries is a clear case of comparing apples with oranges, and thus insupportable. Nor is it entirely valid to criticise the Roman libraries for having no eminent scholars associated with them. While they can certainly not boast of having the equivalent of Callimachus, Eratosthenes or Zenodotus on staff, there were several figures with eminently respectable academic and literary credentials who worked in the Imperial libraries. The most notable figure associated with the libraries was Gaius Iulius Hyginus, whom Augustus placed in charge of the Library of Palatine Apollo, a freedman, possibly of Spanish extraction, although Suetonius indicates that he might have come from Alexandria.231 He was a pupil of the eminent historian Alexander Polyhistor, and became a close friend of Ovid, who addressed Tristia 3.14 to him.232 His literary output was quite impressive, though almost nothing has survived.233 His works included a treatise on agriculture, a book on Italian cities, works on the Trojans, on bees, a book of sayings and a number of works on religion, one of which was quoted by .234 However, his most significant work appears to have been a

228 Athen. Deip. 1.22d. 229 Suet. Caes. 44. 230 Plin. NH 35.9. 231 Suet. Gram. 20. 232 Cf. Van de Woestyne 1929: 33 ff. 233 The work Fabulae, which has survived and is credited to an author named Hyginus, is now universally believed to have been written by a later author. Cf. Rose 1930: 36; Grant 1960: 1; von Albrecht 1997: 877. 234 Macrob. Sat. 3.8.4. 38 commentary on , which is twice quoted by Gellius.235 Some scholars have gone so far as to credit him with making a lasting contribution to the Virgilian canon, although this is now disputed.236 Nearly as prolific a writer as Hyginus was Gaius Melissus, another freedman, native to Spoletium, well-educated and formerly a servant of Gaius Maecenas, arguably the most notable patron of Augustan poetry. With these solid literary credentials behind him, it is perhaps not surprising that Augustus entrusted him with the task of establishing and directing the Bibliotheca Porticus Octaviae.237 His only known work was the Ineptarium, a collection of jests and witty sayings attributed to well-known people, a substantial work running to 150 volumes. However, he also invented a new type of play called trabatae, apparently based on an earlier form of theatre called togatae, comedies concerning the activities of the equestrian classes. As with Hyginus, he was acquainted with Ovid, who records that his plays were light-hearted in tone.238 Yet another friend of Ovid, Pompeius Macer, was placed in charge of the Imperial libraries by Augustus, according to Suetonius.239 Not only a long-time friend and travelling companion of Ovid, but related to Ovid’s wife, Macer was a writer of epic poetry.240 This poetry is praised by Ovid, but Macer is not otherwise mentioned as one of the significant writers of the Augustan era, though two of his poems have at least survived for posterity.241 His political connections must have been excellent, since not only was he placed in charge of the Imperial libraries by Augustus, but he was also procurator of , and then became a close friend of Tiberius.242 Having such an influential and highly-placed individual in charge of them must have given the public libraries some clout. It is worth mentioning here two further figures associated with Roman libraries whose literary credentials are much more impressive than those of Hyginus and Melissus, though the evidence is far more tenuous – Suetonius and Tacitus. The career inscription found at Hippo in Algeria records that Suetonius held the post a bibliothecis, a post which is known from inscriptions of the era to have been associated with the Imperial libraries in Rome.243 It is now generally believed that Suetonius held this position under Trajan, though no further information exists to shed light on this

235 Gell. NA 1.21, 16.6. 236 Reynolds and Wilson (1991: 27) credit Hyginus with inaugurating Virgilian scholarship, while Zetzel (1981: 33) points out that Hyginus’ emendation of the phrase uelati lino at 12.120 is used in some modern versions. However, Nettleship (1898: lvii), Goold (1970: 162) and Horsfall (1987: 178) are dismissive of Hyginus’ contribution. 237 Suet. Gram. 21. 238 Ov. Pont. 16.4. 239 Suet. Iul 56.7. 240 Ov. Am. 2.18; Pont. 2.10. Cf. Schwartz 1951: 184 ff; Syme 1978: 73 ff; Evans 1983: 141 ff. 241 Anth. Pal. 7.218, 9.28. 242 Strabo, Geog. 13.618. 243 Townend 1961: 99 ff. 39 tantalizing fragment.244 The evidence for Tacitus’ involvement with an Imperial library is even more tenuous, namely a reference by the 14th Century Italian writer Guglielmo Da Pastrango to Tacitus having been the librarian of Titus’ private library.245 The accuracy of this reference is a matter of debate, but Reed argues that Da Pastrango may have had access to a preface to Tacitus’ works, now lost, which gave details of the writer’s early career.246 Although figures such as Hyginus and Melissus do not rank with the luminaries who toiled in the Library of Alexandria, their academic credentials are significant, and there is no doubt that their talents were perfectly suited to establishing and directing the Imperial libraries as the Emperors envisaged them.247 Essentially their duties would have been practical, administrative functions. The Imperial libraries were civil service establishments and no doubt the Emperors expected them to be run in an efficient no-nonsense manner. In essence, the Emperors took the idea of the Library of Alexandria and applied basic Roman common-sense practicality to it. For this reason, it is pointless to criticise the Imperial libraries, as does Irwin, for not producing the type of bibliographical and editing work that the Alexandrian Library did, since this was not their function. Another of Irwin’s criticisms of the Imperial libraries concerns their alleged failure to play a part in the survival and transmission of classical literature down to posterity. He goes so far as to say that the chain of transmission which kept classical works intact through the Dark Ages bypasses the public libraries altogether and depends solely on the survival of private libraries. The best that can be said for Irwin’s case here is that it is not proven, and that he relies heavily on the argument from silence. There is simply no evidence which proves one way or another that particular libraries preserved or lost particular texts from the Classical period. It could be argued that public libraries, which by their very nature were highly visible, would have been more likely to have come under attack from either militant Christians, invading barbarians or simple act of God than private libraries, so that there would have been more chance of the latter surviving. This was certainly the case, for example, with the Library of Palatine Apollo, which was burnt down on March 18, 363 AD, possibly a case of arson by rioting Christians.248 On the other hand, it is known that the Bibliotheca Ulpia survived until

244 Wallace-Hadrill 1983: 5. 245 Mendell 1957: 237. 246 1976: 312. Oliver (1979: 224), replying to Reed’s article, disputes this and maintains that Da Pastrango either concocted the reference or garbled an existing line. 247 Also worth mentioning as a library director with sound academic credentials, although he falls well outside this time- frame, is Julius Africanus, who was given the task by Alexander Severus of establishing and directing the Bibliotheca Panthei. Africanus is well-known as a chronographer and friend of the Christian scholar Origen, and his best-known work is the Kestoi, an encyclopaedic miscellany, of which fragments dealing with military tactics and agriculture survive. See P. Oxy iii.412; cf. Granger, 1933: 157-61. 248 Amm. Marc. 23.3.3. 40 at least 456 AD, and possibly until the end of the 6th Century.249 It is also known that the Forum of Trajan, wherein the Bibliotheca Ulpia was located, was a centre for classical education from the 4th to 6th Centuries, and the existence of the largest library in Rome would almost certainly have been a significant factor in this.250 Having thus survived at least two centuries of Christianity and certainly the initial barbarian invasions, there is no reason to suppose that at least some of the texts in these libraries, or copies of them, would not have been available when the first monastic libraries began to appear. Reynolds and Wilson optimistically point out the survival of classical texts in the 6th Century and make the point that the cultural and intellectual surge of the 4th and 5th Centuries provided the necessary infrastructure that enabled books and learning to survive until the monastic libraries began to take over the functions of the classical libraries.251 There is no way to know what part the Imperial libraries played in this survival, but given that at least one of them was still in existence at this period, and, furthermore, was almost certainly involved in the continuing process of education, there is reason to believe that their part in the survival of classical literature into the Middle Ages was hardly non-existent. These criticisms of Rome’s public libraries are largely invalid because they do not relate to the real function of the libraries. The libraries must be judged on how successfully or otherwise they fulfilled their purpose for the people for whom they were intended, the writers and their reading audience. There is no doubt about the function for which these institutions were constructed, above and beyond the need to show imperial benevolence and to advertise imperial power. The sources make this function quite clear. Suetonius, describing Caesar’s initial plan for a public library, says that his purpose was to ‘make available as large a collection of Greek and Latin books as possible for public use.’252 The Elder Pliny described the founder of Rome’s first public library, Asinius Pollio, as ‘the first to make the genius of men public property by dedicating a library.’253 Ovid describes his own works in the libraries built by Augustus as ‘mixed with the memorials of learned men, and made public through the gifts of the leaders.’254 The purpose of these libraries is therefore clear. They were designed to make the works of writers available to as wide an audience as possible, and so their success or failure must be judged on how well they served these two markets, readers and writers.

249 Sid. Apoll. Epist. 9.16; Gregory, Ep. 8.28; cf. Riche 1976: 30, 176 n. 254. 250 Cf. Marrou 1932: 93 ff; Reynolds and Wilson 1991: 40. 251 1991: 81. 252 Suet. Iul. 44. 253 Plin. NH 35.10. 254 Ov. Tr.. 2.419-20. 41

Before attempting to judge how successful or otherwise the public libraries were in serving their clientele, the question of how wide that clientele was must be examined. In short, the vexed question must be posed: how ‘public’ were the public libraries of Rome? Scholars have almost universally adhered to the tried and true belief that the libraries were solely the province of the literate and literary classes. As one scholar puts it, they were no more than the ‘Emperor’s libraries generously thrown open to his friends and clients.’255 Yet, if this was the case, the fortunate classes appear to have been reluctant to use them, or at least to write about using them. There are virtually no references from writers of the first century or the early second century about actually using the libraries for readership or research purposes. Writers such as Pliny and Suetonius are happy to extol the of the Imperial libraries as a noble idea, to make the works of great writers available to all, but show no great enthusiasm for stepping through the portals themselves. Ovid and Horace, as will be shown later, were taken with the idea of the libraries as places where their own works would gain a wider readership, but say little about visiting the libraries themselves. It is not until the middle of the second century that references to actual use of the libraries begin to appear. It is ironic that the class of citizen regarded by scholars as most likely to use the public libraries was exactly the class that least needed to use them. Modern public libraries are largely utilised by the less-advantaged classes of society. Yet scholars are of the view that the exact opposite prevailed in Rome, so that the classes who probably had most need of free access to books were those whose financial resources meant they needed them the least. The reason for this, of course, is the perceived view that literacy rates were extremely low in the Roman world, a condition that does not prevail today in countries with extensive public library systems. Therefore it is the question of literacy that weighs most upon the breadth of the readership of the public libraries and must therefore be examined next. The most recent authoritative statement on literacy in the ancient world has come from William Harris, and his findings on the level of literacy in the Roman world are blunt to the point of brutality. After examining the main surviving evidence, most notably inscriptional evidence, the papyri from Egypt, the graffiti from Pompeii and other chance finds such as the Vindolanda tablets, he states that it is unlikely that the literacy rate, in the Western provinces at least, ‘even rose into the range of 5- 10%.’256 In other words, fewer than 1 in 10 people in the heart of the Empire was capable of reading or writing. Perhaps not surprisingly, this extremely low figure has provoked a backlash from scholars. Early reviews of Harris’ book cautiously accepted the fact that literacy in the Roman world was likely to be low, but demurred at the fact that Harris had put it so low, much lower in fact than any Third

255 Marshall 1976: 261; cf. Dix 1994: 282 ff. 256 1989: 272. 42

World country today.257 This caution eventually developed into a full-scale reaction in the form of a collection of solicited papers from eminent scholars, most of whom did not deny Harris’ contention that literacy in the Roman world was low, merely that the figure he gave was too low, based on the evidence. The main weakness in Harris’ theory, as identified by James Franklin, comes from the graffiti in Pompeii.258 Harris is dismissive of the Pompeian graffiti, brushing over it in about five pages, as compared to the dozens of pages he devotes to papyri from Egypt. The obvious reason is that the Pompeian graffiti stand virtually alone in comprising a written record which can demonstrably be ascribed to the whole population of a given area, not just the intellectual and literary elite. The sheer number alone suggests, as Franklin notes, a widely literate population.259 Tanzer, who examined 15,000 graffiti from Pompeii, states that the graffiti, ‘despite numerous errors in spelling and grammar, give us the impression of a high degree of literacy among the lower classes.’260 In fact, it could be argued that the poor grammar and spelling is perhaps as good an indicator as any that these graffiti were not written by well-educated people of the higher classes, but by poorly-educated people, who nevertheless had sufficient literacy skills to daub snatches of poetry on the walls of their city. This is confirmed by Franklin, who notes evidence of graffiti clearly written by the labourers on a building site and by prostitutes in a brothel, definitely not occupations likely to have been undertaken by people with literary pretensions.261 It is undoubtedly the literary graffiti that are the most intriguing and of the most interest in terms of this study. Literally thousands of scraps of verse as well as quotations from known Latin poets have been found on the wall. This is of interest because it indicates not only literacy, but an interest in literature. Among the poets identified are Virgil, Ovid, Propertius, and Ennius.262 Harris’ answer to these literary graffiti is to cite evidence that the same hands produced a quantity of this graffiti, so that the phenomenon is put down to a ‘few hundred literate slaves and….schoolboys from highly respectable families.’263 This may account for some of the graffiti, but the sheer number of literary graffitos that were daubed on the walls of Pompeii over a long span of time would tend to argue against this idea, though there is certainly an argument to be made that literate slaves would have

257 Bowersock 1990: 37; Kenney 1991: 168. 258 1991: 77. 259 1991: 81. 260 1939: 83. 261 1991: 97. 262 Wallace 2005: 80 ff. Virgil is the most quoted poet. Milnor (2009: 309 ff) records 64 Virgilian graffiti. The most common quotation by far is the first line of the Aeneid, arma virumque cano, of which Milnor records 16 variations. Cf. Della Corte 1940: 171 ff; Franklin 1996: 175 ff. 263 1989: 260. 43 become literate freedman, and if they had developed a taste for literature, they would then have become a ready market for it when they had the freedom to purchase books, or perhaps, to utilise the public libraries in order to access it. It certainly appears that, beginning in the late Republic, there was a desire among the lower classes for the sort of knowledge that is derived from reading. Cicero casually notes the ‘delight that is taken in history by men of the humblest station, who have no expectation of participating in public life, even mere artisans.’264 Certainly it seems likely that someone as astute as Caesar recognised this quite early, since there remains a strong possibility that his Commentaries on the Gallic War were written with an eye to consumption by the general public, though, as Wiseman notes, more likely to be heard than read.265 However, as Rome passed into the Imperial era, it seems increasingly likely that people of the lower classes were reading the works produced by the supposedly literate classes. Increasing numbers of writers make clear in their prefaces and dedications that they expect their works to be consumed by the widest possible readership. Dionysus of Halicarnassus states in the preface of his work that he aims to offer satisfaction ‘to any who desire mere undisturbed entertainment in their reading of history.’266 Not only history but other types of works were apparently aimed at a wider audience than the conventional upper-class readership. The Elder Pliny notices works, apparently including his own, which are ‘written for the common people, for farmers and workers, and for those who have nothing else to do.’267 Martial states that his works are written for those who enjoy watching bawdy mimes.268 On the whole, there seems to be a definite realisation among Rome’s writers that there was a wider audience for their works than their own class. If the evidence of the graffiti from Pompeii can be relied upon, there was a base of functional literacy amongst the lower classes, at least in the urban centres, and it appears that these people had an appetite for literature. As Woolf notes, comparatively few would have had full literacy as we understand it, but nevertheless they could make use of texts.269 Best argues that the interest in literature indicates that the masses of Rome had a literacy level much higher, for example, than that of the Middle Ages.270 Even more, perhaps, than the desire for reading for its own sake, was a desire to be educated, or at least seem educated, for the sake of status. It is clear that during the early Imperial era, there was a definite upswing in the trappings of education among the lower classes, especially noveau-riche

264Cic. Fin. 5.52. 265 1998: 4. 266 Dion. Hal. Ant. Rom.1.8.3; cf. Schultze 1986: 133 ff. 267 Plin. NH 1. 268 Mart. Pref. 1.12-16. 269 2009: 53. 270 1965: 72. 44 freedmen. As Best notes, these people seized upon education and reading as a symbol of increased status.271 It is these people whom pillories mercilessly in the . His protagonist, the wealthy freedman Trimalchio, boasts to his friends about his Greek and Latin library, which, he claims, is the equal of the Imperial libraries themselves. That people like Trimalchio actually existed is confirmed by Seneca, who speaks scornfully of those who collect books merely as decorations for their homes.272 It is possible to imagine that people such as this would have been, as well as consumers of books, a possible clientele for the public libraries. It might well have seemed valuable, in terms of social status, to be seen to be using the libraries, and to mix with others who used them. The question must be asked, however, whether the lower classes, apart from wealthy parvenus like Trimalchio, could have afforded the books they obviously desired, for whatever reason? As Phillips states, the evidence which indicates that the common people were able to read books has little credence unless books were actually affordable.273 The question of book prices in the ancient world is another vexed issue. There is little direct evidence of how much books cost at any stage during antiquity. In fact, the best evidence probably comes from Rome during the Imperial age, because, as Reichmann notes, it was during this era that the best-organised book-trade prior to the invention of printing existed.274 Certainly evidence provided by the writers indicates that the book-trade under the Empire was far more reputable than that of the Republic. For the first time, booksellers are actually referred to by name, the places where they did business are mentioned, and Martial even ventures a description of what a typical bookshop would have looked like.275 Martial indicates at one stage that his little book Xenia cost only four sesterces.276 As Phillips indicates, the daily wage of a common labourer at the end of the 1st Century was probably about three sesterces, so a book like Xenia was evidently affordable by all but the lowest paid workers.277 However, as Xenia was a very small book, and probably produced cheaply, it can no doubt be safely inferred that this was at the lower end of book prices. Other books, in particular deluxe editions and rarer items, were certainly far more expensive.278 Phillips states optimistically that most books would have been within the price-range of artisans and men of modest means, though it seems likely that works other than small volumes of

271 1965: 73. 272 Sen. Tranq. 9.5. 273 1985: 36. 274 1938: 43. 275 Locations: the Vicus Tuscus (Hor. Epist. 1.20.1); behind the Templum Pacis (Mart. Ep. 1.2.7); along the (Mart. 1.3.1, 1.117.9); along the Vicus Sandelarius (Gell. NA. 18.4.1); the Sigillaria (Gell. 2.3.5, 5.4.1); cf. White 2009: 271 ff. Martial’s description: 1.118. 276 Mart. 13.3.1. 277 1985: 37. 278 Martial (1.117) records that a deluxe edition of his works cost 5 denarii. Gellius (2.3.5) notes that a collector purchased an early copy of the second book of Virgil’s Aeneid for 20 aurei. 45 poetry, for example, would have been beyond their regular reach. A work such as Pliny’s Natural History, even in its cheapest edition, would almost have been beyond the means of even the most determined artisan to purchase, not to mention, as Wiseman notes, having been an incredibly large and cumbrous object, in scroll form, for an average man without the assistance of servants to purchase and carry home.279 Yet such difficulties make it more likely that some of the less-well-off readers, having a desire for literature but being unable to purchase it, would have been willing users of the public libraries. There is another class of reader in Rome whose presence is often neglected, but who, having come from the same class as those commonly believed to have frequented the libraries, should not be excluded, namely educated women. There are enough references to well-educated women in the literature of the Empire to indicate that they were not uncommon and may well have formed a largely unsuspected readership. The Younger Pliny extols the literary tastes of his wife Calpurnia, saying that she ‘has developed a taste for literature, possesses my books....and learns them by heart.’280 Propertius speaks of educated women with well-developed literary tastes.281 Agrippina the mother of Nero wrote an autobiography which was seen by both the Elder Pliny and Tacitus.282 A poetess named Balbilla, who came under Hadrian’s patronage, was noted for her mastery of Greek verse.283 Undoubtedly there were many more educated and literate women whose names have not survived. It is worth wondering whether this largely unremarked readership is to be considered as a possible clientele for the public libraries. There is no explicit evidence to say so, but as matters stand there is only a handful of references to educated Roman men using the libraries, so it is not surprising that no references to women using them exist. It might be that women, largely excluded from male literary circles and such events as recitationes, found a dignified and safe place to exercise their love of reading in the Imperial libraries. Based on the evidence, it seems likely that the potential clientele for the public libraries of Rome was much bigger than most scholars over the years have been prepared to accept. The tried and true dogma of libraries being only the haunts of a narrow group of writers and intellectuals simply fails to convince. Certainly writers and intellectuals dominate the small amount of evidence for library usage, but then these were the people responsible for the evidence. There is no direct testimony, but given the relatively high level of literacy discovered in the Pompeii graffiti, and the obvious interest in

279 2007: 384. 280 Plin. Ep. 4.19.1. 281 Prop. 1.2.26. 282 Plin. NH 7.46; Tac. Ann. 4.53. 283 CIG 3.4725-4730. 46 literature of the Roman proletariat, it seems highly likely that some proportion of the libraries’ clientele must have been from the more intellectually disadvantaged classes. The libraries must have seemed like an enormous boon for groups such as educated slaves, freedmen, artisans, men of good education but modest means, and perhaps even women. Even for people who could not read, there is ample evidence that libraries were used for public recitations, the subject of which is dealt with at length in Ch. 5. The evidence is not there because these were not the people who were writing about the libraries. Is it really possible that they were kept out? It seems that writers of the early Empire realised quite quickly that having their works admitted to the Imperial libraries would give a boost to their potential readership. While there are only a handful of references to writers actually using the libraries, there is copious evidence to suggest that they were very keen to have their books placed into the public libraries. Horace makes no secret of his desire to see his books accepted into the public libraries. He sees several advantages to himself and other writers in, as he puts it, ‘filling with volumes that gift so worthy of Apollo.’284 He is aware that by having his books enshrined, as it were, in the Emperor’s own libraries, he will himself be enshrined in the canon of great Latin authors, an idea he anticipates will form in imitation of the canon of great Greek writers already established and time-honoured. As Horsfall notes, he shrewdly inveigles Augustus to offer patronage to himself and other writers by offering to fill the empty shelves of the Palatine library.285 Since admission to the Imperial libraries entailed, as it were, official recognition by the Emperor of the worth of his work, it seems likely such admission helped both to increase his readership and establish him in the canon of Latin poets for posterity. He regards potential admission to the public libraries as a ‘spur to our bards to seek with greater zeal Helicon’s verdant lawns.’286 It has already been noted in Ch. 1 that Pliny records that one of the functions off the public libraries was to make writings as widely available as possible.287 The book which is admitted to the library will gain another advantage, in that it will be preserved for posterity, whereas a book sold from a shop will be in constant use and be passed from hand to hand, and eventually wear out. As Horace says: ‘When you have been well-thumbed from handling by the vulgar crowd and begin to grow dirty, you will either silently feed the moths or escape to Utica.’288 Eventually the book’s fate is to end its days ‘teaching boys their ABC in the city

284 Hor. Epist. 2.1.217-18. 285 1993: 62; cf. Dupont 2009: 152. 286 Hor. Epist. 2.1.219. 287 41; H.N. 35.10 288 Hor. Epist. 1.20.11-13. 47 outskirts.’289 By contrast the book that is preserved in the library will not only enshrine its author in the canon of classical poets, but will be preserved for future generations to read. As Horace puts it: ‘My monument shall it be, and raise its head over royal pyramids....and the long lapse of immemorial time. I shall not wholly die.’290 This belief that books would be preserved in the public libraries beyond their normal lifespan is supported by later references from Gellius. He records that while browsing in the Bibliotheca Ulpia, he came across early edicts of the .291 He also records that later he was sitting with friends in the Domus Tiberiana library, when he was presented with a book by the author M. Cato Nepos.292 This author was the grandson of Cato the Censor and father of Cato the Younger, which would place the book as having been written either in the late 2nd Century or the early 1st Century BC, close to two hundred years before Gellius discovered it. These two anecdotes are impressive testimonials to the preservation powers of the public libraries, which obviously formed a great lure to Horace and other writers who desired literary immortality. It is undoubtedly Ovid who most effectively demonstrates, by of having been excluded from them, the desire of early Imperial authors to have their works placed in the public libraries. Having been exiled to Tomis on the Black Sea for some unknown offence, Ovid wrote in Tristia of his feeling that his books, in effect, had been exiled too. His poems literally send one of his books back to Rome to seek entrance to the three public libraries then in existence. His book first seeks a home in the Library of Palatine Apollo, ‘seeking my brothers....and as I sought to no purpose, the guard who presides over the holy place commanded me to depart.’293 The book seeks also to enter the library ‘close by the theatre’, the Bibliotheca Porticus Octaviae, and the ‘halls of Liberty’, the Library of Asinius Pollio, but is likewise rejected.294 This reflects one of the negative aspects of having libraries under the control of the Emperors, as an Emperor who could admit books as a sign of Imperial recognition could equally exclude books as a sign of Imperial displeasure. Some scholars have doubted whether Ovid’s books were actually excluded from the Imperial libraries.295 Yet there is little doubt that Augustus did in fact exercise his powers of censorship over books admitted to the libraries. Suetonius records that he directed Pompeius Macer to exclude some juvenilia written by Caesar,

289 Hor. Epist. 1.20.17-18. 290 Hor. Carm. 3.30.1-6. 291 Gell. NA 11.17. 292 Gell. NA 13.20; cf. Houston 2004: 2. 293 Ov. Tr. 3.65-67. 294 Ov. Tr. 3.1.69-71. 295 Evans 1983: 53. 48 apparently regarding them as inappropriate to the memory of his divine father.296 In addition, it seems probable that he had the books of Titus Labienus and Cassius Severus burnt when they criticised the Imperial system. Perhaps Ovid got off lightly.297 Further evidence for the importance of the libraries to the scholars and writers of Rome comes from the recently discovered Galen manuscript. Galen writes of the loss of rare books and copies kept nowhere else but the libraries that were destroyed in the fire of 192, as well as copies of more common texts which were known to be accurate versions. Galen further records that the libraries contained collections, which were known by the names of those who originally collected them, and which were either gifted or bequeathed to the library, and whose books were stored in the library under the name of their collectors.298 The names of these collectors indicate that the holdings were up to 450 years old.299 It is clear that for scholars the libraries constituted an importance resource of books and copies of books that could not be easily obtained, or perhaps not obtained at all from other sources. Nicholls goes so far as to say that Galen’s manuscript indicates that the libraries were an indispensable resource for serious scholars.300 This might be going too far. After all, writers during the Republic were able to work without public libraries. However, this does appear to be the first recorded instance of a Roman scholar making extensive use of the public libraries for research purposes. While writers such as Pliny the Elder and Gellius made frequent allusions to the public libraries, there is no direct evidence that they used them as major research resources. It seems clear that the public libraries of Rome offered significant advantages to readers and writers in Rome in the Early Empire as opposed to the Late Republic. They fulfilled their purpose of making the best works of ancient writers available to the people, and the evidence indicates that the people were willing and able to use those resources, though unfortunately direct evidence is lacking to show that they actually did. It is clear that they had enormous potential to open up access to classes previously disadvantaged by the closed-shop collection and exchange of reading resources during the Republic. For writers, they offered significant advantages in terms of making their works available to a wider audience and enshrining their works in the canon of great authors for posterity. Judged on their own terms, therefore, the public libraries of Rome can be considered a qualified success. In subsequent chapters, the usage of libraries, both public and private, by one writer, Pliny the Elder, will be examined.

296 Suet. Iul. 56.7. 297 Sen. Controv. 2.4.11, 10 pref. 4-10; Tac. Ann. 1.72.3, 4.21.3. 298 Gal. Per. Alup. 13. 299 Nicholls 2011: 132. 300 2011: 124. 49

50

Chapter 3: Private Libraries in the Early Empire

The first two chapters have examined, first, the origins of libraries in Rome and the development of private libraries, and, second, the beginnings of public libraries during the Early Empire. This chapter will complete the logical progression by examining the continuing development of private libraries during the early Imperial era, especially those the Emperors themselves maintained in their own villas and palaces. This chapter will culminate in an examination of what is known about the private library of Pliny the Elder. It is a peculiarity of the study of libraries in Rome that while the Republican era yields numerous and fairly detailed written descriptions of private libraries, but almost no surviving remains of libraries, the situation is reversed in the Imperial era. Written descriptions of private libraries are few and sketchy, and some are of doubtful veracity. Archaeological remains, however, are quite plentiful, particularly for the libraries the Emperors established in their palaces and villas. Of course the best-known private library which has been discovered is that of the Villa of the Papyri in Herculaneum, although, as was discussed in the first chapter, there is a possibility that this is an intact relic of the Late Republic. Two more, possibly private libraries are known from Pompeii and Rome itself, and these will be examined. With regard to Imperial private libraries, the existence of libraries has been surmised in the Domus Augusta, the Domus Tiberiana, the , the Villa Iovis on the Isle of Capri, and the Villa Adriana, and these buildings furnish much interesting information about the continuing development of libraries during the Imperial era. The library of the Villa of the Papyri makes a good starting-point, because to some extent it bridges the chronological divide between the Late Republic and the Early Empire. Although the bulk of the collection was certainly assembled before the death of its putative owner, palaeographical evidence indicates that some of the copies were produced as late as the early decades of the 1st Century AD.301 As far as can be determined, the library itself dates from the Late Republic and thus is of limited value in assessing later developments in Roman private libraries, but it makes a good base point for examining how private libraries developed after the end of the Republic. The basic description of the Herculaneum library is well-known, but worth re-capping in brief. The library consists of a small rectilinear room, just 3.2 metres square, with shelves both along the walls and in the centre. It is generally agreed that the library was far too small to be used for reading purposes, and that reading was done elsewhere, with the library a storage room only. This is supported

301 De Lacy 1993: 179. 51 by the finding of carbonised papyri at other locations in the villa, most notably in what has been described as a ‘reading room, half again as large as the library and adjacent to the peristyle.302 This is problematic when compared to the little that is known about Republican private libraries. The only reasonably detailed description of such a library comes from Cicero, describing his visit to the library of Lucullus, where he finds Cato sitting in the library surrounded by scrolls.303 On the other hand, Plutarch’s description of Lucullus’ villa, describes the house as possessing cloisters and study-rooms, and mentions Lucullus walking about with his guests discussing philosophy.304 This description is reminiscent of the layout of the Villa of the Papyri at Herculaneum, with its peristyles, walkways and what may have been study or reading rooms. Therefore it seems likely that literary descriptions of Republican libraries cannot be relied on to provide an accurate picture of what a typical library looked like, if in fact such a thing as a ‘typical library’ existed. It is much more likely that private libraries differed according to their owner’s needs and desires. There is no doubt that the highly specialised library in the Villa of the Papyri represented a personalised collection appropriate to its owner’s requirements. The collection itself represents another problem in taking the Herculaneum library to be in any way typical of Republican libraries. As previously noted, it is composed almost entirely of Epicurean texts, and indeed of texts of just one writer, Philodemus. This bias is at odds with what is known of other private libraries. Although evidence indicates that the libraries of Cicero and Lucullus, for example, contained a high proportion of works on philosophy, it is known that they also contained literature, history and works on oratory, among other genres. The specialisation of the Herculaneum library has led scholars to offer various explanations for the nature of the collection. The conventional view is that the villa belonged to Cicero’s enemy and Caesar’s father-in-law, Lucius Calpurnius Piso.305 Evidence from Cicero himself has been taken to indicate that Piso might have been Philodemus’ patron.306 However, as other scholars have pointed out, the connection between the Villa and Piso rests on no solid evidence. These scholars contend that the library was either assembled by Philodemus himself, or by an admirer, or even a patron, who need not have been Piso.307 Perhaps the

302 Sider 1990: 537. 303 Cic. Fin. 3.2.7. 304 Plut. Luc. 42.1-2. 305 Frank 1928: 83; Bloch 1940: 491; Maiuri 1974: 38. 306 Cic. Pis. 68. 307 Sider 1990: 538; De Lacy 1993: 180. 52 most pertinent comment comes from Johnson, who states that the only certainty is that the villa contained a specialist collection for the use of the villa’s owner and his circle.308 Taken on its own merits, the Herculaneum library cannot be considered a typical example of a Late Republican private library. Its small size and extremely specialised collection clash with literary depictions of other libraries of the period. Yet when compared with another putative library, this one from Pompeii, it can be seen that the Herculaneum library still might mark a starting point from which libraries in the Early Empire were to develop. A room, whose status as a library has not been confirmed and is still a subject of debate, exhibits some features similar to both the Herculaneum library and to later, more advanced libraries. The room in question is in the so-called House of Menander in Pompeii. Excavations of the peristyle area of the house uncovered a small room on the south side of the peristyle. Only slightly larger than the library in the Villa of the Papyri, being some 10 m2, it is adorned with a floor mosaic of the Second Pompeian style of the mid 1st Century BC, depicting a satyr and a nymph. Rows of holes in the southern and eastern walls indicate that shelves were present here at some stage of the building’s existence.309 The floor mosaic has panels delineating spaces adjacent to the eastern and southern walls, wherein, it is believed, large items of furniture once stood. Maiuri was the first to suggest that these may have been large cabinets for holding books, a view supported later by Richardson.310 This view is not wholly accepted, with Ling, for example, suggesting that the room may have been a cubiculum, and that the furniture items were beds.311 However, there is a general consensus that the room might have been a library at some stage in the building’s existence, although not at the very end, when it was almost certainly turned into a storeroom.312 The strongest support for the room having been a library comes from its proximity to the peristyle and, especially, the exedrae that adorn it. In particular, one of the exedrae, larger than the others, is adorned with portraits of two individuals, one of whom has been identified as the poet Menander, from which the building derives its modern name. The portraits also depict tables containing tragedian masks, further emphasising its literary connotations. It is almost certain that this alcove and possibly the others in the peristyle were used for reading, as was probably the case with the Villa of the Papyri. Although Richardson notes that the library room was big enough to contain a

308 2006: 496. 309 Bruce 1986: 514. 310 1933: 88; 1977: 398; 1988: 160. 311 1997: 79. 312 Allison 2006: 308. The storeroom hypothesis is supported by the finds in the room, which included lamps, fragments of glass bottles, nails and bronze fittings. Cf Allison 1997: 82 ff. 53 small desk or a chair, this seems unlikely when these alcoves were so close by and clearly intended for reading.313 It is almost certain that if these exedrae were used for reading, a library was close by, and this room seems the best candidate. The similarities with the Herculaneum library are apparent. Both feature a small room, almost certainly used for storage of the books alone, with shelves and cabinets. Both are readily adjacent to a peristyle, with exedrae where reading was carried out. Like the Herculaneum villa, there is even the possibility that reading was carried out in the peristyle. In the Villa of the Papyri, boxes of scrolls were found in or near the peristyle. In the House of Menander, the finds included strap hinges, possibly from a chest, found in the west ambulatory of the peristyle.314 It is possible that, like the Herculaneum villa, these might have been used to transport books from the library to the peristyle. It is also possible, assuming that the library was removed from the identified room late in the villa’s history and stored elsewhere in the house, further from the peristyle, that chests or cupboards were used to store the books near the reading areas. That a library continued to exist somewhere in the villa after being removed from this room seems to be indicated by the fact that, while the library room mosaic is of the Second Pompeian Style, the decoration in the reading alcove is of the Fourth Pompeian Style. The library almost certainly existed in the adjacent room at the time the mosaic was installed because the floor niches for the cabinets were established at that point. It seems likely that, since the owner of the villa had the Menander and the other literary figure put in the reading alcove either just before or after the library had been removed from the room, his literary interest continued, and therefore some form of library continued to exist elsewhere in the villa. The literary character of the decoration in the reading alcove seems to indicate that, unlike the Herculaneum library, this library was of a less specialized nature. Certainly it would seem to indicate that the owner had an interest in Greek theatre. Structurally and functionally, however, the two libraries are very similar. Both were small rooms where books were stored on shelves and in cabinets, and reading was undertaken in the nearby peristyle. It is possible that this may have been the general pattern for libraries in the Late Republic and very early Empire. As Bruce indicates, it was a close follow-on from the Greek model of a library.315 This was almost certainly the type of library utilised in the gymnasia, which made much use of the peristyle with exedrae for study and teaching.316 However, later developments in Roman libraries saw divergence in a completely different direction.

313 1977: 514. 314 Allison 1997: 70. 315 1986: 514. 316 Makowiecka 1978: 8 ff. 54

Aside from private libraries of the Emperors in their villas, which will be dealt with soon, another possible private library has been located. Lanciani details the excavation of a house on the Esquiline in 1883. One of the rooms he describes as a ‘spacious hall, the walls of which were plain and unornamented up to a certain height, but beautifully decorated above in stucco work.’317 He further describes decoration of fluted pilasters, enclosing plain spaces decorated with medallions about 60 cm in diameter. Fragments of plaster found at the bottom of the wall clearly showed that the medallions had contained painted portraits. One of the medallions bore the inscription ‘APPOLLONIVS THYAN…’318 Clark provides the dimensions of the room, approximately 7 m by 4.5 m, and the height of the stucco work, which was about 1.2 m above the floor.319 It seems very possible, as Lanciani states, that this room was a library, with bookcases around the walls below the stucco work. No date for the building is provided, but given the inscription honouring Apollonius of Thyana and the presence of a Mithraeum elsewhere in the building, it is likely to date from the early 2nd Century AD at the earliest. What this room clearly demonstrates is how the private library has developed since the time of the Herculaneum and Pompeii libraries. Although its dimensions are reasonably modest, only about twice the area of the Herculaneum library, the decoration indicates that it was not intended to be simply a storage room. Although not large enough to have held a sizeable quantity of furniture, this was a clearly a room that was intended to be seen and used. The presence of the decorative portrait medallions is a new feature, but, as will be shown, one that will become common in Roman libraries. It is in essence a development of the portraits of writers from the House of Menander, but this time inside the library. The continuation and development of these features will be shown in an examination of the most significant private libraries of the Imperial era, those of the Emperors themselves. The on the Palatine is well-attested by ancient sources, but its location remained a mystery until well into the 20th Century. From about 36 BC, after acquiring the house of Hortensius on the Palatine for himself, Augustus began buying up properties around it with a view to extending the house. Before he could develop it, however, lightning struck the site and it had to be consecrated. Undaunted, Augustus gave the land over to become the site of the Temple of Apollo, and contented himself with a more modest dwelling adjacent.320 Early scholarship debated the location of

317 1889: 193. 318 1889: 194. 319 1909: 23 ff. 320 Vell. Pat. 2.81.3; Suet. Aug. 29.3; Dio 49.15.5, 53.1.3. 55 this site vigorously.321 It was not until the late 1950s that a complex of rooms was discovered in a quite obvious location, west of the Temple of Apollo and just below the House of Livia, appearing as a rather inviting blank patch in Richmond’s otherwise detailed map of the Palatine. Matching the description given by Suetonius of a ‘modest house’, and that of as consisting of ‘apartments both private and public’, it consists of two levels, an upper level consisting of an atrium and apartments, and the lower, main level a complex of some twenty rooms which were split between living quarters and official rooms.322 Second Pompeian style wall decoration dates the building to the 3rd and 2nd last decades of the 1st Century, which is in keeping with the ancient sources.323 At the centre of the lower level is a spacious room, described by Carettoni as an oikos, or sitting room with a ledge around the wall supporting columns and pilasters.324 Flanking this room are two rooms most notable for having a series of niches in the walls. It is these rooms that have been identified as the libraries. Each of the rooms has dimensions of 9.6 x 4.3 metres, giving a total space of 44.16 square metres in each room, making each room almost four times larger in area than the library in the House of Menander. In total the space of the libraries is some 88 square metres, or ample space, as Bruce notes, for desks and chairs.325 Yet given the proximity of the libraries to the oikos, whose tight unity with them strongly suggests an integrated design, it would appear more likely that reading was done here. The space in the libraries, instead of being occupied by desks and chairs, may have held statues and free-standing cupboards. As noted, the main features of the rooms are the niches, six in each, three on each wall. They are all 2.08 metres in height and 35 centimetres deep, while differing slightly in width between the rooms, the niches in the western room being 99 centimetres wide, and in the eastern room 91 centimetres. Each set of niches flanks a central niche, slightly smaller, 1.73 x 2.08 x .35 metres in the western room, 1.73 x 2.39 x 0.41 metres in the eastern room. As Bruce notes, these niches were almost certainly meant for decorative statuary.326 These niches are a completely new development in the history of Roman libraries. There is no hint of them in the libraries in either Herculaneum or Pompeii. They seem to be almost entirely a product of the Augustan era. It is reasonably certain that the first use of them was in the Library of

321 Richmond (1914: 196 ff) identifies the House of Augustus with the House of Livia. Blake (1947: 168) places it on territory later occupied by the much grander of Domitian. 322 Suet. Aug. 72.1; Dio 55.12.5. 323 Cf. Barton 1996: 91 ff. 324 1969: 24. 325 1986: 519. 326 1986: 520. 56

Palatine Apollo adjacent to this house, since the only remains of that library surviving, although those remains are believed to be a Domitianic rebuilding, contain them.327 The niches in the House of Augustus probably contained either cupboards (armaria) or shelves (pegmata). Thus it can be seen that the shelves from Herculaneum and the cabinets from Pompeii have been shifted from protruding into the floor space of the library into wall niches, thus allowing more space within the library. As previously noted, this space could have been utilised for desks and chairs, or for decorative statues. The origin of these niches, springing completely unheralded into the development of Roman libraries, is a subject of some debate. Wendel argues that the inspiration came from Egypt, brought back by Caesar from the Library of Alexandria, which in turn had taken the idea from New Kingdom architecture, with rock-cut recesses found in buildings at Tell el-Armarna.328 There is no real supporting evidence for Wendel’s hypothesis, however, since no remains of the Library of Alexandria or any other Egyptian library have been found. In fact, it seems likely that the Library of Alexandria was built on a Greek model, not an Egyptian one, and there is no evidence of niches in any known Greek library, as Makowiecka points out.329 Her theory is that the niche was a natural development of the Romans’ increased use of concrete, masonry and brick, which makes the installation of such innovations as wall niches much easier. There had already been a tendency towards the usage of half- columns, pilasters and ledges to break up walls, and from there the idea of fully-recessed niches to free up floor space was a quite logical development. Another development is the presence of decorative portraits within some of the niches in the library rooms. As has already been shown, this feature was present in the private library excavated on the Esquiline, but as that structure was constructed almost certainly a century or more later, the library in the House of Augustus is the first known use of this idea. As Bruce notes, these portraits probably depicted favourite authors.330 This will become a continuing feature within the Imperial private libraries, either as decorative medallions, or as frescoes. It is another step along the path from the bare storage-room libraries of Herculaneum and Pompeii, where books were stored and then removed for reading elsewhere, to rooms, which, even if books were read elsewhere, were clearly intended to be visited and admired. In keeping with this development are the presence of ornamental pilasters and ledges and inlaid marble floors.

327 It is reasonably certain that the Atrium Libertatis did not have such niches. As it was an existing building, probably adapted from an archival function, it is unlikely that it would have been rebuilt to contain wall niches, although as a library it is almost contemporaneous with the Library of Palatine Apollo. 328 Wendel 1949: 413. 329 Makowiecka 1978: 33. 330 Bruce 1986: 520. 57

The other obvious innovation in this structure is the use of paired libraries, again a feature which is completely unique to Roman libraries. These are the earliest physical remains of such a paired library found, although it is known from the sources that the Library of Palatine Apollo was also such a paired structure. The most likely reason for the pairing is the well-attested division of libraries into Greek and Latin sections, which became standard practice in public libraries. Bruce argues that the division here might be because of the tight structural unity with the oikos, the paired libraries creating an architectural symmetry.331 However, the close association between this building and the Library of Palatine Apollo, which is definitely attested in the ancient sources as being divided into separate Greek and Latin sections, makes it seem most likely that this library followed the lines of the public library, which certainly seems to have been built contemporaneously or at most only a brief time before. The library in the House of Augustus, being in the part of the building which seems to have been designated for official functions, may itself have had at least a semi-public nature. It seems possible that it may have been used not only by the Imperial household, but also by friends and guests of the Emperor. This raises the intriguing possibility that there may have been another library elsewhere in the building for the private use of the Emperor and his family. Given Augustus’ well- attested interest in literature and intellectual pursuits, it would seem almost certain that he would have had a library of some type close to his private quarters. Suetonius relates that Augustus had a private study, a retreat known either as ‘Syracuse’ or the ‘technyphion’ at the ‘top of the house.’332 This would seem to indicate that it would have been in the upper terrace of the house, although it has never been definitely located. Given the distance between the upper terrace and the official library, it seems almost certain that Augustus would have had a more private library located somewhere near his retreat. The library in the House of Augustus represents a radical break from private libraries of the Republican era. Obviously the Emperor had more resources to spend and build than the aristocrats of the late 1st Century, so the improvements in size, decoration and functionality are not surprising. However, over and above the extra resources allocated to this structure, it represents conceptually a significant departure from previous library models. These developments appear to have followed the lead of the early Imperial public libraries, but they would continue to appear and to be improved upon in subsequent Imperial palaces and villas. It is known that Augustus built villas outside Rome as well as on the Palatine. Two of these, for which remains have been found, are at Antium and Ostia. In neither have libraries been definitely

331 1986: 519. 332 Suet. Aug. 72.2. 58 identified archaeologically, but it is known that libraries existed at both because of inscriptions detailing staffing. The villa at Antium yielded an inscription, known as the Antiates Ministrorum domus Augustae listing the freedmen and slaves employed at the villa between 31 and 51 AD.333 Among these are four individuals, two slaves and two freedmen, listed as a bybliotheca. Similarly, for the villa at Ostia, a first century inscription lists a vilicus a bibliotheca.334 The employment of these individuals shows that the imperial private libraries were professionally staffed, and that in fact their staffing was probably not much less sophisticated than that of the public libraries, whose staffing structure is well-known from numerous inscriptions.335 It has been suggested in fact that the staffing between the public and private Imperial libraries was interconnected, with staff receiving their initial experience in the Imperial households before moving on to the public libraries.336 In fact, it seems likely that control of both the public and private libraries was held by the procurator bibliothecarum, the senior administrator, who is well-attested in numerous inscriptions as the public servant ultimately in charge of public libraries.337 It seems likely that this arrangement continued until the reign of Trajan, when the post of procurator bibliothecarum was replaced by a lower-ranking position, with the administration of public and private libraries divided.338 After his death, Augustus’ modest residence in Rome was replaced by increasingly grandiose places built by his successors. The residence which succeeded Augustus’ house was known as the Domus Tiberiana, although most of what remains today is actually a product of the reigns of Caligula, Domitian and Hadrian. The portion that was actually built by Tiberius remains largely inaccessible beneath the 16th Century Farnese gardens, and no archaeological remains which can definitely be identified as a library have yet been found.339 That there was a library in the Domus Tiberiana is confirmed by the literary sources, though since all of them are from the mid-2nd Century or later, it is possible that the library to which they are referring is not one established by Tiberius himself, but by his successors. Although there are no confirmed remains of the Domus Tiberiana library to extend the knowledge of the structural development of Imperial private libraries in the Early Empire, the literary

333 CIL 10.6638. 334 CIL 14.196. 335 Cf. Boyd 1915; Houston 2002: 138 ff. 336 Bruce 1986: 541. 337 2002: 139 ff. 338 Bruce 1983: 155. 339 Some foundations which were excavated around the atrium of the Domus Tiberiana in the 1860s have been tentatively identified as those of a library, but confirmation is lacking. Cf Lanciani 1897: 146; Van Deman: 1924: 368 ff; Blake 1959: 16. To further complicate matters, the structure which later became the original church of S. Maria Antiqua has also been identified as a possible location of the Domus Tiberiana library. Cf Rushforth 1902: 19 ff; Coarelli: 2007: 77-78. 59 references do contain important information regarding how these libraries operated. , in his correspondence with Fronto, records that some sort of a bribe is needed to obtain books from the librarian (bibliothecarius) of the Domus Tiberiana.340 This is the only record that indicates that the Imperial private libraries had a professional staff, though there is no reason to doubt that they did. The well-developed staff hierarchy of the Imperial public libraries is well-attested by inscriptions, and it can be assumed that the Imperial private libraries were similarly well-staffed. Gellius records that, when he was sitting with friends in the library of the Domus Tiberiana, he was brought a book written by Marcus Cato Nepos.341 This has been used as evidence to indicate that the Domus Tiberiana library was actually a public library.342 However as Bruce indicates, this only indicates that the library was open to guests and clients of the Emperor.343 This further confirms the evidence of the library of the Domus Augusta having had a semi-public function, within the official section of the house, and being open to guests of the Emperor. It demonstrates the significance of Imperial literary patronage, and indicates that these libraries existed not only for the use of the Imperial household, but for the writers patronized by the Emperors. This in turn is a continuation of the function of libraries of the Late Republican aristocrats, such as those of Sulla and Lucullus, which were open to the use of guests and friends. It indicates that the Emperors had, by and large, taken over the role of literary patronage from these aristocrats, and that this included maintaining libraries for their literary clients, their guests and their friends. This is in fact quite possibly one reason for the Emperors deciding to spend considerable public resources upon building public libraries for what may have been an admittedly small clientele, as discussed in Ch. 2. Tiberius may also have been responsible for the construction of another library, this time outside Rome. The Villa Iovis, on the Isle of Capri, was one of twelve villas that Tiberius erected on Capri, one named for each of the major Roman gods.344 Predictably, the villa named for the chief god is believed to have been his favourite retreat. Discovered in 1827, serious excavation did not begin until 1935, but as early as 1900, there was speculation that the villa contained a library.345 However, excavations of the 1930s proved that Weichart’s contention for a library placed above the vast water cisterns was untenable.346 Maiuri, who directed these excavations, in turn proposed that two identical rectilinear rooms on either side of a hall that led to a gallery overlooking the sea were archival rooms,

340 Fronto, Ep. 4.5. 341 Gell. NA 13.20.1. 342 Boyd 1915: 59. 343 1986: 521 ff; cf. Hall 1932: 31. 344 Tac. Ann. 4.67. 345 Weichardt 1900: 113. 346 1958: 40. 60 understandable since Tiberius was directing the affairs of the Empire from here during his self- imposed exile on Capri. He later withdrew this suggestion and never proposed an alternative archival room or a library.347 Yet these two rooms have been adopted as possible libraries by later scholarship.348 The two rooms are square, each about 7.8 metres long, with a combined area of about 122 metres square. Each has a small rectangular anteroom facing onto the hall, the purpose of which is unknown.349 The hall leads to a large gallery with sizeable windows overlooking the sea. The gallery contains two identical alcoves which project diagonally north-east and south-east, and which are believed to have contained couches, possibly for reading. These alcoves, together with the paired nature of the rooms, have led most scholars to believe that the rooms were Greek and Latin libraries. This cannot be confirmed, however, because there is not enough left of the walls to indicate whether the rooms had niches, or shelves for that matter. There is no trace of decoration remaining in either room, although decorative plaster survives in the gallery and alcoves.350 It is probable that they were libraries, given the proximity to the gallery, which seems to have been specifically designed for reading, but the lack of any significant architectural remains makes it impossible to tell whether it followed the model of the existing Imperial libraries in Rome. The total area of the rooms and the gallery is much larger than the equivalent area in the Domus Augusti, which is interesting given that this library could not have had the semi-public function of Augustus’ library, and could only ever have been intended for the use of the Emperor and his most intimate circle. In 64 A.D., the heart of Rome was destroyed by a fire that lasted a full nine days. The centre of the city was almost completely obliterated, including the residence of Nero, who either made desperate attempts to contain the fire, or ascended a tower and sang songs about Troy, depending on whether Tacitus or Suetonius is to be believed.351 Regardless of his behaviour at the time of the fires, his actions afterwards are beyond doubt. With his own palace destroyed, he set about levelling the area to build an inconceivably more vast and glorious residence, stretching from the Palatine to the Esquiline, which came to be known as the Domus Aurea. Estimates of its size vary, from 125 to 370 acres.352 Successive destruction by the Flavians and by Trajan, however, have ensured that all that remains today of any significance is the residential wing cut into the . Not surprisingly for an

3471958: 48 ff. 348 Kahler 1965: 89; Andreae 1977: 535. 349 Bruce (1986: 537) argues that they were probably waiting rooms. 350 Bruce 1986. 537. 351 Tac. Ann. 39; Suet. Aug. 38. 352 Ward Perkins 1956: 212. 61

Emperor whose literary, musical and theatrical interests are well-attested, it is believed to have contained at least one library. One of the most distinctive features of the remains of the Domus Aurea is the enormous hexagonal courtyard that divides the western part of the building, believed to be the private quarters, from the official, eastern section. Opening onto the eastern side of the courtyard is a suite of three interconnecting rooms. The central room is larger, apsidal, with a substantial recess at the rear. There is a large central niche in this recess, flanked by two smaller niches. In the main body of the room are 4 more niches. The main portion of the room measures 5.8 by 3.9 metres, with a total area of 25 square meters. The recess is 2.9 by 1.4 metres, with a total area of 4.1 square metres. The central niche is 1.8 by 1.2 by .33 metres, and the other niches are 1.8 by 1 by 0.46 metres. The niches are 1.1 metres above the floor.353 Not surprisingly, in view of these niches, the room has been identified as a library.354 No trace of decoration remains in the room today, but there is reason to believe that it was once decorated with marble panels, gilding and stucco painted in Fourth Pompeian style.355 An 18th Century observer noted that the niches were fluted and painted with musical themes.356 Adjoining the room are two smaller rectilinear rooms, which have been identified as sellaria or sitting-rooms.357 If this is the case, then the structure is following a seemingly established pattern in these libraries of including the reading room as part of an integrated unit, replacing the peristyle with exedrae, which were features of Late Republican libraries. Both library and sellaria would have been well-served for light, with the library having two rectangular apertures in the ceiling. If the structure is indeed a library, then it represents a small, but significant break from the previous pattern of libraries. It is the first known library to dispense with the purely rectilinear shape that distinguishes all known libraries up to this point. It is only a small break, however, not a radical one. The library is not a huge departure from the traditional rectilinear shape.358 Yet in view of later,

353 Bruce 1986: 526 354 Not without dispute, certainly. (1966: 117) is so certain of its status that he declares it the only room in the Domus Aurea whose function is certain. However, Strocka (1981: 309) states that the niches are too far above the floor to be of use in a library. Segala and Sciortino (1999: 80) are doubtful because of the lack of pegmata in the niches and also because the south-facing aspect of the room defies Vitruvius’ prescription. Ball (2003: 92) maintains that the niches could have been for statues as easily as for books, for which viewpoint some support may be derived from Pliny’s assertion (NH 34.84; 35.120) that the Domus Aurea was laden with looted art. 355 Bruce 1986: 522. 356 Cameron 1772: 54. 357 McDonald 1982: 38 n. 46. 358 Indeed, by the standards of the Domus Aurea, it is quite conservative in design, given that the building features such radical departures as a hexagonal courtyard and an octagonal room. 62 far more radical designs in Imperial libraries, this appears to have provided the chink of light that led to a complete break from the rectilinear form. It is possible that this was not the only library in the Domus Aurea. In fact, due to its vastness it seems inconceivable that the palace did not contain more than one library. As noted, it is believed that this library was part of the official section of the residence, which may mean that another library, for the personal use of the Imperial family alone, was located in the western, private wing. In addition, given the usual pairing of Roman libraries, it seems likely that this library had a twin elsewhere in the complex. In fact, just across the courtyard, is an almost identical suite, a set of three interconnecting rooms. The only difference is that the main room is rectilinear and lacks niches, although it does possess a recess at the rear. A possible reason for the lack of niches in this room is raised by Ball, who noted that the niches in the eastern library were added comparatively late in the history of the complex, either in the late Neronian or Othonian period.359 If this is the case, it is possible that these rooms only became libraries quite late in the history of the residence. Thus, while the niches in the eastern room were completed, work might not have begun on those in the western room when the Flavians came to power and stopped all further work on the Domus Aurea. The two rooms are not exactly identical, therefore, because they were not originally intended to be paired. After the downfall of Nero and the chaos of the year of Four Emperors, neither Vespasian nor Titus made significant additions to the Imperial complex on the Palatine, though they did turn Nero’s vast lake into the Flavian Ampitheatre. Vespasian chose to live in a modest residence on the , while it is believed that Tiberius lived in the Domus Tiberiana.360 When Domitian came to power, he set about constructing a palace on the Palatine that would literally become the last word in Imperial residences on the Palatine, remaining in use until the end of the Empire and beyond. Although Suetonius disparaged Domitian’s lack of literary interest, he does concede that the Emperor went to a great deal of effort to rebuild and restock the public libraries which had been destroyed by fire, even sending scribes to Alexandria to make copies of books.361 The particular incarnation of the Library of Palatine Apollo, which has been found adjacent to his palace, is believed to have been largely his work. Despite Domitian’s supposed lack of interest in things literary, it can be expected that the Domus Augustana he constructed would contain a library, as had every other Imperial palace on the Palatine. Yet one has never been positively identified. He essentially constructed the palace in two sections. To the north is the section reserved for official duties, known

359 2003: 92. 360 Cass. Dio 65.10.4. 361 Suet. Dom. 20. 63 as the Domus Flavia. No library has been identified here, although it was perhaps seen as unnecessary because the Library of Palatine Apollo is literally outside the front door. The larger, southern section, comprising the Emperor’s private quarters, the Domus Augustana proper, does, on the other hand, contain at least one candidate worth considering. From the south, the Domus Augustana can be entered by a large curving pulvinar or balcony, from which, it is believed, the Imperial family viewed events in the .362 An entry hall with colonnades leads to a large peristyle. On the northern side of this peristyle is a suite of three rooms. The central room is polygonal, domed, with a tall rectangular niche flanked by smaller rounded niches. Of most interest, however, are the two adjoining rooms. These are octagonal, with alcoves in each of the northern, eastern and southern walls. Each of these alcoves has rectangular niches in the side walls, and a semicircular recess between them. The rear wall of each has a large curved niche.363 The potential for these rooms to be considered libraries is obvious, though no scholar has yet admitted the possibility. The rooms are paired, with both rectangular niches suitable for cabinets or shelves and circular niches for statues, features already well-known in previous libraries. In fact only a couple of scholars have put forward possible uses for these rooms other than libraries. McDonald conjectures that the rooms are nymphaea.364 This seems unlikely, for the rooms do not contain water features and there are other, more obvious nymphaea in the structure very close by.365 McKay comes closer, by labelling the rooms salons.366 In the structural sense, a salon is usually defined as a drawing or sitting room, or a reception room. In a less physical sense, it has also come to mean a gathering of people by a host for conversation, amusement and literary or artistic stimulation. It is possible that these elaborate rooms were intended for just such a function, for private gatherings of the Imperial family and selected guests. In fact, it is possible to see these rooms as combined libraries and sitting rooms, with the books tucked away in alcoves, accessible, but not usurping space in the main part of the room. The location of the rooms, adjacent to the peristyle, harks back to the days of the Republic, and further back, the Greeks, when discussion was carried out in the exedrae of the peristyle. This association is strengthened by the fact that each room presents a very open face to the

362 Lanciani 1897: 143. 363 Blake 1959: 121. 364 1982: 65. It should be noted that Blake (1959: 136) describes rooms with similarly alternating rounded and rectangular niches at Domitian’s Alban villa as nymphaea. This pattern also appears in the so-called Villa of Clodius at Albanum, which Domitian modified, and another of Domitian’s villas at Circeo. It is, as Blake notes (1959: 139), clearly a Domitianic trademark. 365 Coarelli 2007. 366 1975: 76. 64 peristyle, as Blake notes.367 Admittedly this sort of gathering seems not to fit with the character of Domitian as presented by Suetonius, but the brickwork seems to confirm that it was constructed during his reign, though brick-stamps suggest later work as well.368 Given the evidence that Suetonius himself presents regarding Domitian’s interest in the public libraries and this carefully designed complex within his private dwelling, it must be said that the slander about Domitian’s supposed lack of interest in literature looks decidedly shaky If these rooms are indeed libraries, and there are compelling reasons to believe they might be, then they represent a major break with the past. After the small breakthrough of Nero’s non-rectilinear library, these are extremely sophisticated designs, with careful attention being paid to both the efficient use of space and the radical use of shapes. Yet the basic features of libraries that have been seen in all the Imperial structures to date are present, most notably the pairing and the niches. Future developments of the Imperial private libraries would show how these features would be maintained. With Domitian’s construction of the Domus Augustana and its subsequent usage by Emperors until the end of the Empire, new developments in libraries must of necessity be looked for outside Rome. Neither of Domitian’s immediate successors, Nerva or Trajan, is known to have made any significant arrangement for libraries in their dwellings, although Trajan, it must be remembered, did create the largest of Rome’s public libraries in his Forum. It was left to Hadrian to make the next major development in private library construction, at his massive villa complex in Tivoli. Previously thought to have been built during the later years of Hadrian’s reign, the villa at Tivoli is now known to have been a project of his first decade. The only significant description of it in the ancient sources comes from the author of the Historia Augusta (‘Aelius Spartianus’), who informs the reader that Hadrian gave the names of provinces and famous places to parts of it, including Lyceum and Academia.369 Names such as this, taken from well-known Greek places of learning, are testament to Hadrian’s well-attested Hellenophilia and love of knowledge, culture and literature.370 This is emphatically supported by an examination of his villa, which may have contained as many as four libraries. To construct his edifice, Hadrian chose a site already occupied by a villa from the Late Republic. His architects incorporated the structure of this villa into the new design and it became the centre of the Imperial villa, from which subsequent developments and additions spread outwards.371

3671959: 121. 368 Platner 1965: 162. 369 Hadr. 26.5. 370 Hadr. 1.5, 14.8, 16.1; Tert. Apol. 5.7; cf. Syme 1991: 103 ff. 371 Coarelli 2007: 483. 65

Among the adaptations was the transformation of one of the Republican villa’s rooms into the new complex’s first library. From what had been the entry area of the old villa, above what is now known as the Courtyard of the Libraries, a small porticoed peristyle with pilasters adjoins a group of three rooms, the central room of which was given nine niches, eight of them rectangular with the central niche in the back wall being rounded.372 The rectilinear room is 8.3 metres by 8.2 metres, with an area of 68 metres square. The rectangular niches are 1.6 metres wide with a depth of 63 centimetres, while the rear rounded niche is slightly deeper (75 centimetres). In what had now become a well-established pattern, this niche was almost certainly occupied by a statue, possibly of Hadrian himself. The floor below the niches is surrounded by a low podium, 28 centimetres above the floor and 35 centimetres wide. This is a significant development, because permanent podiums are unknown in any other private library in the Roman world.373 On either side of the library are rooms which are 9 metres by 6.15 metres. As Bruce notes, the complex is similar to the three-room library suite in the Domus Aurea and can be considered an integrated whole.374 The two adjoining rooms, used for reading or reception, form what has now become a commonplace with libraries. The location of the library suite adjoining the porticoed courtyard echoes the early library model of exedrae opening off a peristyle, as does its conservative rectilinear construction. How much of this conservatism was due to the necessities of adapting an earlier structure and how much to the desires of Hadrian and his architects is unknown. Yet it represents a clear step backwards from the more radical library designs of Nero and Domitian. Coarelli argues that this library, in the heart of what would become the Imperial living quarters, would have been Hadrian’s private collection, though probably open for the use of his family and guests.375 If this is the case, then almost certainly there would have been a more public and official library elsewhere in the villa, as has been the pattern in other Imperial palaces. Some scholars have been bothered by the fact that only one library has been identified in the complex described above. Given that by this time the pairing of libraries seems to have become almost mandatory, as Makowiecka argues, this library should have had a mate close by.376 A possible candidate has been identified. Below the terrace containing the library is the large Courtyard of the Libraries. On the opposite side of this courtyard is a rectilinear exedra opening directly onto it. This has been tentatively identified as a library, although the identification is hotly disputed. Makowiecka

372 Blake (1973: 238) argues that this room had been the tablinum in the Republican villa. 373 The Library of Celsus in Ephesus, the Bibliotheca Ulpia and, significantly, the Library of Hadrian in Athens are known public libraries with permanent podiums. Cf. Makowiecka 1978: 17. 374 1986: 529. 375 2007; 484. 376 1978: 75. 66 identifies it as a Latin library, the partner to the Greek library already identified, and states that the libraries can be respectively identified as Greek and Latin because of their construction.377 This theory argues that, since the terrace library was built upon a peristyle in the Greek fashion, it must have contained Greek texts. However, as has already been shown in respect of Republican villas, early Roman library design followed closely on the Greek model. Hence this style could as easily be classed as early Roman as it could be classed as Greek. There is thus no necessity for thinking that Greek texts were stored in a library of this design. The so-called Latin library is a rectilinear structure, 7 metres by 5 metres, with an apse and seven niches, six on the side walls and one in the back wall. Its most distinctive feature is the construction of double walls, with a narrow passage running between them. Even the overhead barrel- vaulting is doubled. This led to its identification as a nymphaeum, with the overhead space between the vaults thought to have been used for water storage and pipes running between the walls.378 The main problem with this theory is that no trace of piping has been found in the structure, although Makowiecka concedes that the initial purpose of the building may have been changed during the life of the complex.379 Clearly, the design of the building with the side and central niches resembles a library, and it is entirely possible, even probable, that it did serve as a library at some stage during the villa’s life. It seems likely that its initial usage, at the time the terrace library was constructed, was as a nymphaeum, and therefore it is unlikely to have been a pair with that library. There is also the fact that it is a considerable distance away from the terrace library, farther than any known paired library is from its mate. Makowiecka argues that it could have been the pair because it is on an exact line with the terrace library, even to the extent of being off-centre in the axis of the courtyard, implying a connection between the two. This is certainly possible, though the distance and the fact that the two libraries appear to belong to different complexes of buildings tend to argue against it. Makowiecka also stresses that the double wall structure has been seen in other libraries, as a means of keeping humidity away from the collection.380 Clearly, the most that can be said with any degree of certainty is that at some stage in its existence, this building was used as a library. Whether or not it was the pair of the terrace library, however, remains unproven. The Courtyard of the Libraries derives its name from the two structures which lie in front of it, flanking the double-walled structure described above. Oriented to the north and facing onto a terraced

377 1978:76. 378 Gusman 1904: 100; Bruce 1986: 534; Adembri 2000: 63; Coarelli 2007: 484. 379 1978: 77. 380 1978: 77. The Library of Celsus at Ephesus is the best-preserved example of this; cf. Johnson 1980: 651 ff; Casson 2001: 117. 67 garden, the two apsidal structures were identified as libraries in the 19th Century, and have been signposted ever since as the Greek and Latin libraries.381 They are respectably large, the westernmost Greek Library being 28 metres by 10 metres, the Latin library somewhat smaller at 22 metres by 7 metres. Both structures have apses, however the Greek Library apse is rectilinear, while the Latin Library is rounded. It is now universally believed that they were not libraries, but either seasonal triclinia, a tower belvedere or, more recently, the monumental entrance to the palace. The discovery of a heating system in the upper storey of the Greek library has been interpreted as confirming that the buildings were residential.382 It is not hard to see, however, how the buildings could have been interpreted as libraries, since each has prominent rectangular niches and possesses adjoining rooms similar to the reading rooms known from other Imperial dwellings. It is not impossible that in fact they were a variation on the salon theme, dining rooms with book storage that could be used for literary functions and readings. Such a purpose has in fact been proposed for the large hall in the Sala dei Filosofi, which will be dealt with later. Immediately to the west of the Courtyard of the Libraries is a quite remarkable structure called the Teatro Marittimo or Maritime Theatre. It is a completely circular structure, with a moat surrounding a small island. The island is not in fact a theatre, but a completely self-contained, miniature villa. As Kahler notes, the only thing lacking is a kitchen.383 It seems likely that this extraordinary structure was Hadrian’s private retreat, a more advanced variation on Augustus’ Syracuse.384 The eastern section of this miniature villa contains three small rooms, with the central room containing an I-shaped wall oriented lengthwise. Above and below the arms of the I are the rooms which have been identified as small libraries. The rooms are cruciform in shape, with recesses in each wall of slightly differing sizes, between 2.2 and 2.3 metres wide and between 45 centimetres and 1.3 metres deep. It is believed that the rooms contained armaria rather than shelves as no anchor holes are visible. No decoration now remains. It is likely that the alcoves formed by the side of the I contained couches, with the eastern alcove most suitable for reading in the mornings and the western best in the afternoon.385 There are obvious similarities to the retreat in Tiberius’ Villa Iovis on the Isle of Capri, with the two small libraries as part of a suite with a room set up for reading comfort. The pairing of the libraries suggests a division between Greek and Latin libraries, although whether such a

381 Gusman 1904: 93 ff. 382 Kaehler 1950: 106 ff; Bruce 1986: 528; Adembri 2000: 63 ff. 383 1965: 150. 384 Coarelli (2007: 486) notes that it may have been inspired by a similar structure in the palace of Dionysius of Syracuse, which was mentioned by Cicero (Tusc. 5.59). Cf Adembri 2007: 62. 385 Kaehler 1950: 118 ff; Blake 1973: 242 ff; Bruce 1986: 533 ff. 68 scheme was followed in a library intended for the use of one person is unknown. It is possible that the division was a purely personal one devised by Hadrian himself, perhaps by genre for example. This sort of personal library is noted by the sources, with Pliny the Younger describing a library in the living quarters of his Laurentine villa, consisting of one set of shelves containing books he likes to ‘read and read again.’386 It is quite possible that this was the sort of library Hadrian utilised in his retreat, containing his favourite works. Immediately to the west of the Maritime Theatre is a large apsed hall known as the Sala dei Filosofi or Philosopher’s Hall. Its dimensions are impressive. The rectangular hall is 17.5 metres by 14.4 metres, with the apse measuring 11.5 metres by 6 metres. Its total area is almost 300 metres square. As Makowiecka notes, its size puts it into the class inhabited by the largest public libraries in Rome, and its apsidal structure has in fact been compared with the remains of the Library of Palatine Apollo.387 The niches, seven in all, are contained only in the apsidal area. They are correspondingly large, 3 metres by 1.4 metres and 75 centimetres deep. Although sometimes identified as a library, the general attitude of modern scholarship is that it was not. Makowiecka relates the main objections as being its size, too big for a library within a private residence, and the presence of four large entrances in the side walls, which would admit too much noise. She also argues that the recesses are too large and deep for bookshelves, as well as being too far off the floor and more likely intended to contain statues.388 Bruce, on the other hand, does concede the possibility that it is a library, arguing that the hall may have been a literary salon where readings took place. He believes that the height of the niches would not have been a problem, given that wooden steps could have been used.389 As he points out, noise would have been a factor in most of the public libraries as part of their nature, and ways may have been found to deal with this. That this building may have been some type of salon or hall for public reading is certainly a possibility, given that there may have been a salon structure in the Domus Augustana, albeit on a much smaller scale. Hadrian’s villa represents something of a paradox. In a structure which is noted for its radical interpretations of architecture and described as the ‘culmination of architecture’, its libraries are surprisingly conservative.390 While there are some small developments, none of Hadrian’s libraries build on the initial steps taken by Nero and Domitian. The first library was extremely

386 Plin. Ep. 2.17.9. 387 1978: 74; Bruce 1986: 531. Blake (1973: 243) and Coarelli (2007: 486) identify it as a library, but make no detailed arguments. 3881978: 75. 3891986: 533. 390 McKay 1975: 131; Ward-Perkins 1981: 109. 69 conservative, purely rectilinear and looking back towards the libraries of the Greeks and the Republic. The one significant step forward was the permanent podium, which is unknown in any other private library in the Roman world. It is possible that such podia were utilised in Greek libraries. The one preserved Hellenistic library, the royal library of , possesses one, although there is dispute about whether it served as a podium proper or a support for the bookshelves.391 It is possible that the Roman belief in practicality above all militated against permanent podia. In the limited space available in private libraries, a permanent podium takes up space, and it may have been seen as more practical to have temporary and portable access to the shelves. Why this library alone has a podium is unknown. It is possible that it was an existing structure in the building from which the library was adapted. It is significant that one of the only other libraries in the Roman world with a podium is the Library of Hadrian in Athens. It is certain that Hadrian had a major say in the design of his buildings, but exactly why this building alone has such a structure is unknown. With regards to the supposed pair of this library, located beside the Courtyard of the Libraries, its major development would seem to be the double wall system. As noted, a double wall system has been found in other libraries in the Roman world, most notably the Library of Celsus in Ephesus. However, it seems unlikely that a double wall system was installed here for the purposes of a library. The library simply seems too small to justify it. It is more likely that the double walls were indeed part of the structure’s earlier nymphaeum phase, and were of necessity still in place when the building was later converted into a library. Otherwise the library’s structure is, again, quite conservative. The Sala dei Filosofi does seem to expand on the salon idea initially seen in Domitian’s palace, and there is little doubt that it would have functioned effectively as a large space for public readings. However, the question must be considered whether it actually functioned as a library. The niches in the structure are very large and only located in the apse. It seems more likely that they were intended for statues as a backdrop to the speakers who were reading to the audience. It is possible that books were stored elsewhere and brought here for readings. This is impossible to prove, but at this stage it must be said that this structure’s status as a library is unproven. The most radical structure in the Villa is undoubtedly the Teatro Marittimo. This miniature, self-contained villa was obviously intended as the Emperor’s personal retreat. While the design of the structure is radical, the library and its associated rooms are a variation on a theme already noted in Imperial dwellings, most notably the library and reading complex from Tiberius’ Villa Iovis on the Isle

391 Makowiecka 1978: 17. 70 of Capri. There is no doubt that this represents the sort of Imperial retreat which was attested in Augustus’ palace, but it does not seem to represent a significant step forward in library design. It is regrettable that there are so few remains of non-Imperial private villas left, so that it can be seen whether or not they followed the lead of the Emperors in making such major advances in private library design. The only real known example is the Esquiline villa, which, at least in terms of decoration, with its stucco portraits, follows the example of the Domus Augusta and the Domus Aurea. It could also be argued that the pilasters which divide the walls are at least a partial attempt to imitate the niches which became such a ubiquitous feature of both the public and private Imperial libraries. It is only in terms of breaking away from the rectilinear design that it does not follow these libraries, and then, as has been seen, even the Emperors could not bring themselves to abandon fully the rectilinear structure. As it stands, however, an archaeological examination of private libraries gives a very one- sided picture of library development during the early Empire. The Emperors obviously had far more resources to devote to library design and construction. In addition, they had the full resources of the Imperial public service to devote to staffing them. Their libraries were also at least partly public, with most of them fulfilling a function as part of the public part of the palace, and in that sense they cannot really be considered private. To examine real private libraries, there is only the testimony provided by the writers who were by and large the owners of these libraries. The next chapter will examine the work of one of these writers, Pliny the Elder, in an effort to determine not only the nature of his library, but also how he used library resources, his own, as well as those of the public libraries.

71

Chapter 4: The Library of Pliny the Elder

Previous chapters have examined the development of Roman libraries from primitive beginnings to the sophisticated private libraries in the villas of the Late Republican elite. They have charted the subsequent development of public libraries in the Early Imperial era, and asked whether these libraries were of any great relevance to the writers who seem to have been intended as their main clientele. Then followed an examination of the development of private libraries in the Imperial era, where almost all the surviving evidence relates to the grand villas and palaces of the Emperors. This chapter will now move from the general to the particular, and examine the library of one particular Roman writer, Pliny the Elder. There are a number of reasons for choosing Pliny. First, thanks to his nephew, there exists a detailed account of how he used his library, arguably unique in the history of Roman literature. Second, Pliny himself, through his unprecedented index to the Natural History and his diligent attention to naming his sources, provides an unparalleled opportunity to study the contents of his library and his pattern of usage in writing the book. Third, as will be shown in the course of this chapter, Pliny is the most librarian-like and library-minded of all Roman writers, through his highly- ordered thinking and desire to control the retrieval of information. As noted, Pliny the Younger left an extremely detailed account of his uncle’s study habits. Thanks to him, more is known about his uncle’s work habits than about those of most other ancient authors, as Murphy notes.392 The Younger Pliny’s indicates that his uncle devoted his entire waking life to study, literally, for he would continue working by lamplight and even wake during the night to work.393 This nocturnal study habit is not entirely unusual, as it is mentioned as a habit by other Roman writers.394 However, Pliny’s adherence to his study routine throughout the day is, as far as is known, unique. His nephew records that he had books read to him while he was sunning himself, and made notes and extracts.395 Similarly, he would continue working through meals, and even while bathing.396 While travelling in a chair, he would continue working.397 However, as impressive as his dedication was, far more interesting is the information that this reveals about his library. The very fact that he was prone to use books in such unusual places indicates that, almost wholly and solely, the books he was using must have come from his personal library. There is no record of Pliny using a

392 2004: 3. 393 Plin. Ep. 3.5.8. 394 Cic. Parad. Stoic. 5; Stat. Theb. 8.112; Juv. 1.51, 7.99. 395 Ep. 3.5.11. 396 Ep. 3.5.12, 14. 397 Ep. 3.5.16. 72 public library, and it seems unlikely that personally visiting one would have fitted in with his peculiar study habits.398 If he did have need of a book that could perhaps only be obtained from one of the public libraries, it seems possible that he might have sent a slave to make a copy of the required text, which would then have become part of his personal library.399 Similarly, it is almost certain that, if Pliny borrowed a book from a friend, he again would have had a copy made and added to his library.400 Pliny’s letter records even more information about his uncle’s usage of his personal library. Hidden behind the details of the Elder Pliny’s obsessive study habits there possibly exists a sophisticated information recording and retrieval system. Pliny specifically records, not once but twice, his uncle’s method for collection information from the books he used. First, he notes that, while sunning himself, the Elder Pliny would have a book read to him while he ‘made notes and extracts’ (adnotabat excerpebatque).401 He uses the same terminology soon afterwards, stating that his uncle ‘made excerpts from everything he read’ (excerperet), and during meals ‘he took rapid notes’ (adnotabatur).402 Locher has noted the careful distinction made between adnotare and excerpare, and argues that, while the meaning of excerpare, ‘to make extracts from a text’, is clear, Pliny’s use of adnotare has been misinterpreted.403 Although the word’s original meaning was ‘to add notes to something already written’, it later came to refer simply to the taking of notes, which is what Locher believes is the conventional view of what Pliny was saying. Locher argues, however, based on Pliny’s use of the term elsewhere in his letters, that the original meaning was intended, and that the Elder Pliny was in fact adding notes to something already written, either the text itself, or adding notes to his extracts.404 Locher takes this further to argue that, if the Elder Pliny was adding notes to either extracts or to the texts he was using, in fact the notes were retrieval aids, designed to facilitate rapid recovery of information when he came to the point of actually writing his books.405 Locher envisages a pile of slips or scraps of , each containing an extract from a text, ordered in such a way, by whatever

398 It seems very unlikely, for example, that he used the libraries in the same way that Galen describes in the Peri Alupias. Galen appears to have spent considerable time in the public libraries himself, going so far as to store his notes and copies in storehouses near the Palatine, probably to avoid having to carry them back and forth from home. See Gal. Per. Al. 12-13; cf. Nicholls 2011: 127-128. 399 Although there is no record of him using the public libraries, they are mentioned a number of times in the Natural History (7.115, 210; 34.43; 35.9, 10) and Pliny is quite positive about their usefulness, which tends to support the idea that he utilized them in some fashion. For copyists, cf. Cic. Q Fr. 3.4.5; Sen. Ep. 27.6; Hor. Ep. 2.2.5; Plut. Crass. 2. 400 Cicero in particular records numerous instances of borrowing books and making copies for his own library: Att. 2.4.1, 2.20.6, 8.11.7, 13.8, 13.31.2, 13.32.2, 13.33.2. 401 Ep. 3.5.10. 402 Ep. 3.5.11, 12. 403 Locher 1986: 25. 404 Ep. 10.96.4. 405 1986: 26. 73 notes Pliny added to them, that they could be rapidly assimilated into whatever he was writing, creating what was in effect an early version of a computerised document retrieval system.406 This would certainly fit in with the encyclopaedic structure of the Natural History, which is, as the Elder Pliny himself attests, less a book than a storehouse of facts.407 Locher admits that there is no way to prove that this is what Pliny actually did, and there has been little if any critical discussion of his theory.408 However, if indeed Pliny did create some mechanism like this in order to catalogue and retrieve the data he extracted from his mountain of reading, in effect he created the first known database. Pliny’s letter also reveals significant information about the staff that his uncle used to assist him in his research. The titles of two of the servants who assisted him are divulged in the letter. The reader (lector) was employed to read books aloud during meals while the Elder Pliny took notes.409 A secretary (notarius) travelled with him carrying a book to be read and notebooks for recording.410 It seems likely from copious evidence elsewhere that staff of this kind were commonly used. A lector is most commonly noted for given the task of reading as entertainment, where the host would have literary works read aloud for his guests.411 Unusual as it may seem, this is what the Elder Pliny was doing while he was researching. As Johnson notes, while he was taking notes from the book the lector was reading, a guest asked the servant to repeat a line. In effect, then, Pliny was combining research with entertainment.412 It can be assumed, however, that Pliny also used his lector in private to read to him. There are a number of examples of Romans having books read aloud in lieu of reading themselves.413 The advantages of having books read aloud included coping with declining eyesight in an era before spectacles, as Horsfall notes.414 The difficulties of dealing with lengthy and burdensome book rolls, which had to be unrolled and kept unrolled, would have been made easier by having a lector read from the text, and this also avoided the necessity of having to read oneself from what might be a difficult text, written as was the custom without punctuation or spacing.415 In a situation such as the Elder Pliny regularly put himself into, rapidly and intensively reading books and noting their

406 1986: 27. 407 Plin. NH pref. 17. 408 The Oxford Latin Dictionary lists only one other usage of adnotare in a sense similar to that employed by the Younger Pliny, and a reading of that passage would seem to indicate that Pliny the Elder annotated the books rather than the excerpts (Fronto, Ad. Am. 2.2). For critical reaction to Locher’s essay, see Hankinson 1988: 340; Riddle 1988: 398. 409 Ep. 3.5.12. 410 Ep. 3.5.15. 411 Nep. Att. 14.1; Plin. Ep. 9.36.4, 9.40.2; Gell. NA 2.22.1; cf. Starr 1991: 341 ff. 412 2000: 616. 413 Cic, Fam. 7.1; Suet. Aug. 78.2; Fronto, Ep. 3.1. 414 1995: 49. 415 1991: 343. 74 contents, it is easy to see that just such a well-trained individual would have been an essential part of Pliny’s research. In fact, it is likely that Pliny’s research methods would have been nearly impossible without one. The other member of the Elder Pliny’s research staff who is identified is the notarius. The Younger Pliny only uses this term one other time in his letters, and the servant in that instance is clearly identified as someone who receives his master’s dictation of his literary works.416 In the case of the Elder Pliny, the servant is given the task of receiving his master’s thoughts on the books he is having read to him. Elsewhere in Roman literature, a notarius is usually identified as a stenographer or shorthand writer.417 It seems extremely likely that Pliny’s notarius also took down his master’s notes in shorthand, given the rapid pace that Pliny probably set. The second source of information about Pliny’s library is the extensive preface he provided for his Natural History. Addressed to the Emperor Titus, in his preface Pliny sets forth a detailed explanation as to why he wrote the book, what he hoped to achieve, and, to some extent, how he went about it, including an explanation of the origins of encyclopaedias, described by Howe as unique in the encyclopaedias up to and including the Medieval era.418 Of primary interest for the purposes of this discussion, the preface also includes valuable insights into Pliny’s library and his attitude towards libraries in general. While direct information about Pliny’s own library is limited in his preface, some interesting details are recorded. Most notable is the rather obvious fact that the bulk of Pliny’s library would have consisted of books on nature, which he declares simply to be the subject of the Natural History, clarifying it somewhat by stating that in actuality he means the whole of life.419 This is confirmed by even a cursory study of the massive summarium or table of contents that makes up the whole of Book 1. He also records that he used 2000 books in the writing of the Natural History, a statement which gives some idea of what must have been the impressive size of his collection. Very few details of the size of private libraries are given in the sources. Considering that the 2000 volumes Pliny mentions almost certainly do not constitute the entirety of his collection, since he had, after all, according to his nephew, written other books on a variety of subjects, it can be argued that his personal library must been of impressive size.

416 Plin. Ep. 9.36.2. 417 Sen. Apoc. 9.2; Mart. 5.51.2; Quint. Inst. Or. 7.2.24. 4181985: 562. 419 NH pref. 13. 75

In his preface, Pliny mentions the authors and titles of some of the works that must have been in his collection. Some of the authors are quite to be expected, for example, Cato, Cicero, Livy, and Virgil. Others, however, are something of a surprise, for example Catullus, who is actually the first author mentioned in the preface.420 Pliny mentions a large number of titles of books that he does not approve of, predominantly because of their titles. He launches a diatribe against authors, predominantly Greek, who give their books fancy titles that do not indicate anything of the contents, or worse, give promise of content that is actually lacking.421 However, it is possible that despite his dislike of these types of books, they were still part of his library. This can be surmised because his nephew indicates that Pliny considered no book bad enough that he could not extract something from it.422 Although the information about Pliny’s library indicated in the preface is limited, certainly compared to that available in the summarium and the book itself, what is revealed nevertheless is Pliny’s attitude towards libraries. The preface is permeated with Pliny’s desire to see knowledge preserved and made available to the widest possible readership. As Howe notes, Pliny argues that writers have a duty to produce books which will advance the common good.423 Most notably he echoes the statement by Domitius Piso that what is needed are not ‘books, but storehouses.’424 This has usually been interpreted as meaning reference books, especially encyclopaedias, rather than books devoted to a single topic.425 However, he might also have been referring to libraries. It can be argued that if encyclopaedias as storehouses of data were to be created, then libraries are the logical and ultimate development of these storehouses. This is supported elsewhere in the Natural History, where Pliny gives his endorsement to the idea of public libraries.426 What is more, Pliny clearly indicates that he desires this information to be available to the widest possible audience, stating that his book is written for ‘the common herd, the mob of farmers and of artisans.’427 While there is an element of sarcasm and exaggeration in Pliny’s statement, there is little doubt that he intended the knowledge he provides to be available to those who, as Beagon notes, were educated but not intellectual.428 Looking back to Chapter 2, wherein the potential clientele for Rome’s public libraries was discussed, it is clear that Pliny envisaged a readership for his book that was considerably wider than the narrow aristocratic

420 NH pref. 1. 421 NH pref. 24. 422 Plin. Ep. 9.36.2. 423 1985: 563; NH pref. 16. 424 NH 17. Cf. Too 2010: 59 4251999: 36. 426 NH 7.115; 35.10. 427 NH pref. 6. 428 1992:13; cf. 1999: 40. 76 circles in Rome which are generally assumed to be the only ones who could read. As Goodyear notes, Pliny had a practical vision for his book.429 He sought to instruct and pass on practical and useful knowledge. It seems unlikely that this type of knowledge was intended for upper-class literary circles, but was meant as a measure to preserve knowledge that Pliny feared was disappearing as Romans became more and more lazy and addicted to luxury. With this attitude, it seems reasonably certain that Pliny would have whole-heartedly endorsed the idea of libraries as a key measure in this preservation of knowledge. Certainly, his personal library, which, as was seen earlier, would have been quite extensive, was a tangible monument to this belief. One of the most remarkable features of the Natural History is Pliny’s massive summarium, which takes up the whole of Book 1. As Conte notes, it is quite unique in the ancient world.430 Pliny’s purpose in creating it is made clear in his preface. In keeping with his idea of treasure-houses rather than books, he intended it to be used as a retrieval device for specific facts.431 Although usually described as a table of contents, it is in fact more than simply this. Its sheer size alone shows this, as it dwarfs any sort of contents table produced today. In fact, it almost qualifies as a work in its own right. In addition, it is more complex than a modern table of contents and contains some features of an index and a bibliography. Its depth and complexity certainly make it an invaluable resource in examining Pliny’s library and how he used it. The summarium consists of three principal parts. For each book, a table of contents, a total of the number of facts, investigations and observations, and a list of the sources are given. The contents table gives a numerical listing of each section of the book with a short phrase or sentence describing the content of the section. There is no common structure to the descriptive comment. Some are single-word or two-word descriptors. Others pose questions, which are presumably to be answered in the text. Some have three or more short sentences or phrases describing the content in slightly more depth. As Calvino notes, there is nothing systematic about the descriptions and some of them are quirky and individualistic.432 However, there is a structure within Pliny’s table of contents that enables it to function, as he intended, as a retrieval aid. Each keyword, or keywords within the descriptive phrase or sentence in the table, is repeated in the first sentence of the chapter described. It is, in fact, a less rigid version of a sophisticated table of contents known from an earlier Roman medical writer, Scribonius Largus. Largus flourished during the Claudian era, and his work De Prescriptione has

429 1982: 670. Cf Doody 2010: 92 ff 430 1994: 69. 431 NH pref. 33. 432 1986: 316. 77 survived. His table of contents also used descriptive phrases and sentences, though, unlike Pliny’s, his descriptors then appear as distinct chapter headings. It is possible that Pliny also had something similar at one stage. As Doody notes, the text of the Natural History has been much amended and altered during its sojourn from its composition through the Middle Ages to the present.433 At times, the summarium has been broken up and its parts attached to the head of the books to which they refer. It is remotely possible that, if Pliny did have chapter headings like Scribonius, they were inadvertently removed at some stage during this process. There is, however, no firm evidence to suggest that Pliny was influenced by Scribonius in the structure of the table of contents. Somewhat surprisingly perhaps, although medical writings form a key part of Pliny’s sources for the Natural History, Scribonius is not listed as a source, although he was certainly contemporaneous with Pliny while the latter was a young man. The table of contents may also present another clue to how Pliny used his sources. As has already been noted, Locher argues that Pliny made notations and comments on the excerpts and notes that he made from his reading to act as retrieval aids while writing the Natural History. He takes this further by arguing that these notes and comments subsequently became the descriptors that Pliny used in his table of contents.434 The seemingly arbitrary nature of the phrases and short sentences that Pliny makes use of in his table would then make perfect sense. They are in fact the very comments and notes Pliny made while reading, which is exactly what they read like. If this is true, then Pliny has made his own retrieval aids into retrieval aids for the reader. As Too notes, he has made it possible for the reader to look for specifica facts within the Natural History without having to read the whole thing.435 It is in fact an ingenious use of an existing resource and would have been a tremendous time- saver. In fact, given the massive task Pliny set himself, it seems certain that he would have been in search of any time-saving methods possible. While this cannot be proved, it certainly seems logical in light of the evidence. The second part of the summarium gives a number for all the facts, investigations and observations recorded in the book. This is quite unique and is known from no other writer in the ancient world. It is this part of the summarium that Pliny appears to mention when he states in his preface that the book contains 20,000 facts. However, as with so much of Pliny’s arithmetic, this does not appear to add up. Addition of all the totals listed in the summarium gives a staggering total of

433 2001: 2. 434 1986: 27. 435 2010: 59. Doody (2010:94) notes that Martial (10.1) jests that readers should not bother reading the whole book, but select the poems they want to read. 78

33,707 facts, investigations and observations. As Pavord notes, it is doubtful that anyone has ever added up the total of facts in the book itself to determine whether that total matches Pliny’s total of 20,000.436 However, it is the third part of the summarium that is of most interest in regard to discovering information about Pliny’s library and how he used it. It is also the part of the summarium that has aroused the most interest from scholars over the years. This is because, as Goodyear notes, it is so ‘unusual and noteworthy’ in an ancient text.437 It is, in short, a list of Pliny’s sources. However, in terms of Pliny’s library, it is of greater significance than that. It is, in effect, a catalogue of at least part of Pliny’s collection, something completely unknown from any other Roman writer.438 It is far from being a perfect catalogue, because he generally lists the authors only, detailing the titles only rarely. There is also the problem that, in all likelihood, many of the authors listed were not used in their original texts. It seems probable that Pliny made extensive use of secondary sources, compilations of the sources he used. In fact, as Lloyd notes, the sheer size of the task he set himself would have forced him to do this, as it would have limited the number of texts he could physically consult.439 As imperfect as this catalogue is, it nevertheless yields useful information about Pliny’s library. First, in this bibliography, Pliny divides his sources into Roman and foreign (mostly Greek) authors.440 This can probably to some extent be put down to Pliny’s ideas of Roman cultural superiority and his expressed disdain for Greeks. However, it could also indicate that, like the Roman public libraries and also the private libraries of the Emperors, Pliny’s library was divided along linguistic lines. There is, of course, no independent evidence to verify this, though it seems logical from the basis of what is known about other Roman libraries and Pliny’s own mania for order. As already noted, Pliny indicates in his preface that he consulted 2000 books in the course of writing the Natural History. If it is assumed that his collection contained books that he did not utilize in the writing of this book, then it is a fair assumption that his library was probably larger than this. It would seem logical for a library of this size to be divided linguistically, though of course hard evidence is lacking. There is some evidence of a logical systematisation within Pliny’s listing of his sources. In general, the listing of the authors follows the order in which they are referred to within the book in

436 2005: 63. 437 1982: 670. 438 Inscriptions which seem to have been the catalogues of public libraries in Rome have been discovered, most notably a fragmentary example from the 2nd or 3rd Century, which may have been from the Bibliotheca Ulpia in the Forum of Trajan. It seems likely that catalogues in public libraries were engraved either on panels or directly on the walls. See IG 14.1097, 1098, 1098a; cf. Blum 1991: 127. 439 1983: 147. 440 Cf. Too 2010: 63 79 question. There are exceptions, with some of the lesser authors being mentioned in the summarium, but not referred to by name in the book. However, the general trend lends some tentative support to Locher’s argument that the summarium consists of Pliny’s own notes and commentaries being used as retrieval aids. There is also some evidence of subject subdivision within the source listing. Pliny groups together authors who have produced books on similar subjects. For instance, in the listing for Book 2, he lists Caecina’s text On the Etruscan Disciplines, and then immediately afterwards two other authors qui item, ‘who are similar.’ Later on in the index, he begins grouping the writers of medical texts together, first for Book 12 as physicians, then from Book 13 describing them as medical writers. While it is a stretch to claim that this indicates that Pliny’s library was subdivided along subject lines, although logically it may well have been, it does indicate what Conte describes as Pliny’s ‘demon of bibliography.’441 Pliny’s summarium is such a unique piece of work that an examination of where the author derived the inspiration and the model from is worthwhile. Pliny himself says in his preface that he took the idea from Valerius Soranus and his book Epoptides.442 Unfortunately, as this work has not survived there is no way of knowing how closely Pliny followed this model. There is, however, a surviving work which may have influenced the format of his listing of sources. Varro, in De re rustica, provides a list of some fifty sources in his preface. Although obviously nowhere near as detailed as Pliny’s long list, the similarities are obvious, and, in addition, a large number of the authors Varro cites appear in Pliny’s list too. Varro was a huge influence on Pliny. He is cited by name a massive 82 times in the Natural History. Another possible, silent influence upon the structure of Pliny’s summarium may have been the scholar, poet and librarian Callimachus, whose massive Pinakes, the 120-volume catalogue of the Library of Alexandria, established, as Pfeiffer notes, the model for cataloguing and bibliography in Antiquity.443 It was still in use by scholars in the 1st Century AD and there seems little doubt that Pliny would have had access to a copy. Callimachus is referred to 14 times in the Natural History, an eminently respectable total. Possible evidence that Pliny used the Pinakes comes from the fact that he cites Callimachus a number of times in the summarium as a medical writer. The only medical text that is known to have been written by Callimachus is one of the pinakes which is believed to have dealt with medical writers and their books.444

441 1994: 69. 442 NH pref. 33; cf. Rawson 1985: 51; Holford-Strevens 1988: 23. 443 1968: 127. 444 Callimachus, fr. 429; cf. Witty 1958: 135; 1968: 126 ff. 80

Of course the richest source of information about Pliny’s library and how he used it can only come from the book itself, the thirty-six volumes that comprise the actual contents of the Natural History. This is possible because, as Conte notes, the Natural History is the book which is richest in internal references of all known works from the ancient world.445 Pliny cites 473 writers, 146 Roman, and 327 Greek and other foreign writers. The fact that he cites these writers by name and also in some cases cites the title of the work in question is a massive resource for the study of his library and how he used it. For the purposes of this study, five of these authors have been chosen as a representative sample, the five authors who have the most citations in the Natural History. As it happens, these authors form an excellent cross-section of the sources used by Pliny. Two of them, far and away the most-cited sources in the book, are Roman, Marcus Porcius Cato Censorius and Marcus Terentius Varro. Two of them are Greek, Homer and Theophrastus. One is neither Greek nor Roman, but an author who straddled both worlds, an African who wrote both in Greek and Latin, King Juba of Mauretania. As well, each of the five represents essentially a different genre of writer: Cato the practical writer of agricultural handbooks, Varro the polymath, bibliophile and encyclopaedist, Homer the semi-mythical epic poet, Theophrastus the Greek zoologist, and Juba the African king and travel- writer. Between them, these five writers are cited by name 311 times in the Natural History. Discovering just why Pliny relied on these particular authors so much and how he used them gives an insight into how he used his library to construct the Natural History. Cato the Censor has the distinction of being the most cited author in the Natural History. He is cited by name more than 80 times in volumes II-XXXVII, in addition to 15 citations in the summarium. As is perhaps to be expected from an author who has been frequently accused of inconsistency, the citations in the summarium do not exactly match those in the books themselves.446 There is no doubt that Cato is not only one of the most important writers used as a source in the Natural History, but he has, as a person, far and away the most important influence on Pliny of any of the sources used in Pliny’s work. Cato is Pliny’s role model, the ideal Roman, stern and disciplined, wedded to simple antique rustic values, disdaining luxury, and despising the all-pervading Greek influence that has overtaken Rome. His praise for Cato is exhibited throughout the Natural History. Cato is, in Pliny’s words, ‘a supremely competent and unrivalled agriculturalist’, ‘a leading authority’, and ‘the master of all excellent crafts.’447 In fact, for his own purposes, it appears that Pliny has

445 Conte 69. 446 Cato is cited in the summarium as a source for volumes 3, 4, 8, 11, 14-22, 28, 36. He is cited by name in volumes 3, 8, 14-21, 25, 26, 29, 36. Curiously, the summarium cites him as a source for 15 volumes, and he is cited by name in 15 volumes. As can clearly be seen, however, the volumes cited do not always match. 447 NH 14.48, 16.193, 25.4. 81 idealized Cato to fit the image he wants. As is now realised, Cato was not as fanatically anti-Greek as he has been painted by Pliny and other authors. As has been pointed out by a number of scholars, while Cato’s denunciations of the Greeks are well-known, he himself acquired an early knowledge of the , was well-educated in Greek literature, and to some extent used Greek models for his works.448 As Astin notes, he was in fact more condemnatory of the rampant and uncritical Philhellenism that was taking over Rome than of the Greeks themselves.449 In addition, Cato was not the champion of the simple antique Roman farmer that Pliny idealized. As White points out, Cato was writing for a new breed of farmer, businessmen owning large properties and relying on slave labour to work their properties.450 Nevertheless, although it appears that Cato was not quite the role-model Pliny desired or envisaged him to be, there is no doubt that his books were heavily used by Pliny in the writing of the Natural History, and his use of these books furnishes a good deal of information about how Pliny used his library in general. Cato is believed to written around a dozen books, although some were probably never intended to be published.451 Of these, only his agricultural handbook, De agricultura, survives in full. It has the distinction of being the earliest known work of Latin prose to survive intact. Fortunately, it is also the source that Pliny relied on the most in writing the Natural History. Fortunate, because as it survives in full, it is easy to examine the work and to see how Pliny used it. While Pliny frequently cited his sources’ names, he rarely noted which of their books he was citing from. With De agricultura, however, it is relatively easy to place virtually every Catonian reference in the Natural History, more than 70 of them, in its proper context.452 One of the most immediately noticeable things about Pliny’s use of De agricultura, once his references have been tabulated and identified with the corresponding passage in Cato’s work, is that Pliny made a habit of mixing and matching different passages that deal with the same or similar subjects, often from parts of the book very remote from one another. A good example of this is Pliny’s reference to Cato’s uses for olive-oil lees.453 He unites two references from Cato that come, respectively, from De agricultura 91 and 130. Similarly, a reference to the difficulty in rearing cypress trees, which is taken from De agricultura 47 and 151.454 Part of this may be due to Cato

448 Astin 1978: 168. 449 1978: 169. 450 1973: 456. 451 1968: 182 ff. 452 Although De agricultura is referred to more than 70 times, the references are almost wholly confined to just six volumes, 14-19. 453 NH 15.33-34. 454 NH 16.139. 82 himself, and his practice of using what have been identified as ‘doublets.’455 This is where Cato repeats himself, sometimes word for word in different parts of the book. However, it may also be evidence of Pliny’s working method. As was noted earlier on, his nephew records that Pliny made notes and excerpts from the books he was reading. It appears likely that it was these notes and excerpts, rather than the books themselves, that he utilised while writing the Natural History, and that, as Locher argued, he arranged these notes into some fashion that enabled him to retrieve data quickly. If these notes were arranged according to some subject-based keyword system, then it would have been relatively easy to identify two or more passages dealing with a similar subject and put them together in the Natural History. It would have been much more difficult, even though De agricultura is a relatively short work, to identify similar passages straight from the book. Pliny’s method sounds like a recipe for chaos, and one can imagine massive piles of notes and excerpts, but if Pliny used an effective retrieval system, then it would ultimately have turned out to be every efficient and time- saving. Certainly efficient and time-saving, but definitely not error-proof. Pliny makes mistakes in his transcriptions from De agricultura. The wonder is, given the size of the task he set himself, that he did not make more errors. In one place, he misquotes Cato as saying that Coan wine should be left for four years in the sun, rather than, as Cato says, two or three days.456 In another place, despite quoting Cato word for word, he uses ‘sappino’ (black fir), instead of ‘carpino’ (black hornbeam).457 Presumably, he made errors like this either through misreading the original text while dictating his notes, or through transcribing them correctly but misreading his notes when writing the Natural History.458 This is something of an indictment of the frenetic methods of reading that Pliny used, as described by his nephew. Then again, the error might not rest with Pliny, but with corrupt copies. There is a strong possibility, given the sheer number of books that Pliny had to work with in writing such a vast book, that some of the copies were corrupt. Cicero complains about the poor quality of bookshop copies during the Late Republic, and there seems little reason to believe that things were much better during the Empire.459

455 Astin 1978: 192. Astin gives as examples, among others, De ag. 51-52 being repeated in 133, and 63 being repeated in 135.4. Astin examines all the theories offered for these doublets, and ultimately puts it down to a lack of systematic planning on the part of Cato, due mainly to his having no previous models to draw on. Cf. Astin 1978: 198 ff. 456 NH 14.79; De ag. 112. 457 NH 16.193; De ag. 31.1. 458 Another possibility is that the secretaries to whom he dictated misheard or made errors through ignorance of the subject material. 459 Q. Fr. 3-.4.5, 5.6. 83

Pliny makes fairly extensive use of direct quotes from Cato. Some of these are brief, consisting of one line or even a couple of words or a phrase. Others, however, are large, meaty, verbatim chunks from De agricultura, whole paragraphs or even more. An excellent example of this is where Cato is dealing with varieties of grape vines, and emphasizes the importance of his opinion by stating that it ‘deserves to be set out separately and handled at full length.’460 At another point Pliny sets out clearly his reasons for using direct quotes, saying that Cato’s thoughts are best expressed in his own words.461 Of greater interest for the purposes of this study is the fact that, in the distinction between the short quotes and the paraphrasing of Cato’s words on one hand, and the meaty chunks on the other hand, we can apparently see the difference between Pliny’s notes and excerpts. As was noted earlier, the Younger Pliny records his uncle taking notes (adnotare), and making excerpts (excerpare) and clearly makes the distinction between them. It appears that Pliny, for his own purposes, either just dashed off quick notes, or made very short excerpts, or, at whim, made longer full-length verbatim excerpts. In the Natural History, therefore, when Cato is cited, if Pliny paraphrases his words, or gives a very short quote, we are seeing Pliny’s notes. However, when he delivers a large verbatim quote, he would seem to be reflecting one of his excerpts. In addition to De agricultura, Pliny makes use of at least two other works by Cato. The Origines, which was probably Cato’s last work and may even have been unfinished at his death, is of high significance in being the first known Latin history. It has not survived intact, but some 125 fragments give a reasonably complete picture of a total of seven books. Pliny makes use of Books 2-4, which primarily deal with the origins of Italy and its tribes. All of these references are found in Books 3 and 8 of the Natural History. One of the more interesting aspects of Pliny’s use of Cato here is that while he is unstinting in his praise of Cato when using De agricultura, when using the Origines he is prepared to show that the master had feet of clay, so to speak. In one reference he gently refutes Cato’s insistence that Vertamacori was a Ligurian village. He then records a bashful admission from Cato that he does not know the origins of the Orumbivians.462 A little later on, Pliny records that Cato is almost alone in believing the Salassi and Lepontii to be of Tauriscian origin.463 Considering there are only around twelve references from the Origines in the whole of the Natural History, it is somewhat amazing that Pliny has found so much that is debatable in Cato’s work. To a certain extent it belies the oft-stated argument that Pliny was an uncritical compiler.

460 NH 14.45-46; De ag. 6.4-7.2. 461 NH 17.125. 462 NH 3.124. 463 NH 3.134. 84

The other work of Cato, of which Pliny is definitely known to have made use, is the Ad filium, ostensibly a collection of exhortations, maxims and advice that Cato compiled for his son. It was at one stage thought to have been one of the earliest Roman encyclopaedias, but it is now generally believed that the fragments on medicine, warfare and farming which have been attributed to this work in fact came from other books by Cato.464 There are only two definite fragments of Ad filium in the Natural History, at opposite ends of the work.465 The latter reference, however, is one of the most interesting and most studied quotes in the work, because it deals with Cato’s attitudes towards two of Pliny’s pet dislikes, namely Greeks and doctors.466 ‘When that race gives us its literature it will corrupt all things, and even all the more if it sends us our physicians.’ There seems little doubt that Pliny chose to reproduce this quote at length because it so clearly follows his own thinking, and fits the image of Cato that he wants to reinforce.467 Clearly, Cato did not like Greeks, and he liked Greek physicians even less. However, an examination of the sources reveals that he was well-read in Greek literature, and was acquainted with Greek philosophy, although not necessarily from early in his life.468 Astin makes the point that Cato’s dislike of Greeks appears to have arisen from a belief that, while they had once been a decent race, they had descended into corruption and deceit.469 In the extract from Ad filium that Pliny uses, Cato tells his son to ‘dip into their literature, but not to make too close a study of it.’470 Clearly, he thought that while there was benefit to be gained from Greek culture, too close an association was dangerous and could lead Roman society into the same morass as the modern Greeks. However, distinctions of this kind appear not to have mattered to Pliny, who eagerly seized on this extract at face value, because it said exactly what he wanted to believe about Cato, and gave the image of Cato that he wanted to reinforce. Much the same could be said about Pliny’s use of the works of Marcus Terentius Varro. French argues that Pliny uses Varro very selectively to create an image of Varro that suits Pliny’s interests, a practical man with personal experience of the extension of Roman power.471 In truth, however, simply because of the vast range of Varro’s interests and writings, he cannot be contained in such a mould, and emerges in the pages of the Natural History as his contemporaries and successors regarded him, even hostile Christian authors who held him to be the epitome of learned , as

464 1973: 332 ff. 465 NH 7.171, 29.13-14. 466 It is also quoted by Plutarch at Cato Mai. 23.2. 467 1973: 171. 468 Cic. Sen. 1.3; Nep. Cato 3.1; Plut. Cato Mai. 2.3-4. 469 1973: 172 ff. 470 NH 29.14. 471 French 229. 85 an individual of prodigious learning and scholarship.472 Perhaps it is because Varro cannot be the author that Pliny wants him to be that Pliny is so stinting in his praise. Although he may have cited Varro nearly as many times as Cato, nowhere does he praise Varro as he does Cato. In fact, while he gives fulsome praise to Homer and Theophrastus, who are Greek, and King Juba, neither Greek nor Roman, Pliny gives nothing in the way of praise to Varro, despite the fact that in writing the Natural History, Pliny used Varro, almost certainly more than any other author, as a source for an incredible number of subjects, including geography, agriculture, medicine, art, theatre, mirabilia and biography. That he was able to find such a broad coverage of material in the books of one author is not surprising, because it is believed that Varro may have written as many as six hundred works, of which unfortunately only one survives completely and another in part. It is likely that Pliny used Varro’s massive Antiquitates, which ran to 45 books, for the bulk of his geographical knowledge, concentrated in Books 3-6. His material on theatre almost certainly came from Varro’s numerous works on matters theatrical.473 Material on medicine was most probably drawn from the Disciplinae, the encyclopaedic work that prefigured the liberal arts curriculum of the Middle Ages. For the purposes of this study, closest attention will be paid to Pliny’s use of Varro’s writing on agriculture and on art. Varro’s only surviving complete work is De re rustica, which may have been the last work of his long life.474 It is significant that, with its preservation, an unbroken sequence of Roman agricultural writing from Cato through Varro and to Pliny himself survives virtually intact. This makes comparison of the various approaches to writing about agriculture quite straightforward. Certainly Pliny made a great deal of use of his predecessors in agricultural writing, though Columella is only rarely cited in comparison to his predecessors. To some extent Varro is used as a support act for Cato, despite the fact that Varro’s coverage of agriculture was much broader than that of Cato.475 While Cato largely concentrated on vines and olives, Varro dealt with a wide spread of agricultural concerns. Yet Pliny largely ignores this, and virtually all of his citations from De re rustica come from Book 1. The reason for this is not hard to see. As White points out, Book 1 deals with basic agriculture, while Book 2 covers animal husbandry, in which Pliny has little interest, and Book 3 with specialised, luxury agricultural products, which are anathema to Pliny.476 This is perhaps one reason why Pliny uses Varro only as support for Cato in agricultural matters. Another reason might be the

472 Cic. Brut. 205, Acad. Post. 1.3.9; Quint. Inst. Or. 10.10.95; Gell. NA 4.16.1; Lactant. Div. Inst. 1.6.7; Aug. Civ. Dei 6.2. 473 De Scaenicis Originibus, De Actionibus Scaenis, De Personis, De Descriptionibus, as well as his specific works on the plays of Plautus, Quaestiones Plautinae and De Comoediis Plautinis. Cf. Conte 1994: 214. 474 Conte 1994: 214. 475 Some examples of Varro being used to support Cato: NH 17.54, 28.21. 476 1973: 482 ff. 86 literary nature of Varro’s text, which is not a practical agricultural handbook, but a set of dialogues amongst various characters, including Varro’s wife Fundania and the agricultural writer Tremelius Scrofa. As was noted earlier, Varro refused to be the writer of practical instruction that Pliny wanted him to be, and this is certainly clear from De re rustica, which has been described as the work of a ‘garrulously entertaining old scholar.’477 An examination of Pliny’s use of De re rustica tends to support the idea that, for Pliny, Varro was secondary in agricultural matters to Cato. Apart from being used to support Cato’s ideas, Varro is almost entirely referred to via single quotes from De re rustica, unlike the numerous instances where Pliny went to the trouble of identifying widely-spaced parts of De agricultura that dealt with similar subjects. It appears clear from Pliny’s use of Varro that De re rustica does not rate that kind of attention. Similarly, while Pliny makes extensive use of direct quotes from Cato, sometimes emphasizing the value of reading them in the author’s own words, direct quotes from De re rustica are much rarer.478 It is clear that Pliny regards Varro as being secondary to Cato in agricultural matters, even though Varro is closer in time to Pliny and has a wider coverage of agriculture. Some of this might be put down to Pliny’s belief that Cato represents the ideal of the antique Roman farmer, despite the fact that Cato and Varro both favour the same type of farm enterprise, large slave-owning latifundia. Support for this idea might be found in the fact that Pliny cites Virgil’s , a poem which extols antique rural values and traditional Italian farming, many more times than he does Varro’s work, despite the fact that, ironically, De re rustica is recognised as being a major source for Virgil. The picture is completely different when it comes to looking at Pliny’s coverage of art in the Natural History. It is now almost universally believed that Varro is Pliny’s main source on matters related to art, and Pliny relied on Varro’s works to access more obscure Greek sources.479 Pliny uses a core group of Greek authors, Xenocrates, Antigonus, Duris, Heliodorus and , who are now recognised as constituting antiquity’s first practitioners of professional art criticism.480 The evidence indicates, however, that he largely drew on them via the mediation of Roman authors, the chief authority being Varro. It is not known from which of Varro’s works he drew the information, as none of Varro’s known works appear to deal specifically with art.481 Much of the debate now centres on

477 Rose 1954: 222. 478 Some examples: NH 14.96, 18.165 ff, 18.348. 479 Varro is certainly the most frequently cited writer on art in the Natural History. Cf. Sellers 1896: lxxii; Pollitt 1974: 80. 480 Pollitt 1964: 317 ff. 481 It is likely, assuming that he did not write a specialist text on art which has not been mentioned by ancient sources, that the source was either or both of the Antiquitates or the Disciplinae. 87 whether Pliny relied entirely on Varro or used him for the core of his information, but still accessed the Greek sources directly.482 It has long been argued that Pliny did not directly access all of the sources he lists, but used compendia and other collections gleaned from the original sources. It could be argued that this was necessary given the sheer size of the task he set himself in writing the Natural History, as well as his own frenetic research methods. The problem with using intermediary sources is that the text which comes down from the original may have been altered by the intermediary, and this certainly seems to have been the case with Varro’s use of the Greek art historians. It is now believed that Varro doctored the Greek works to suit his own purposes. For example, while using the work of Xenocrates, Varro elevates the sculptor Praxiteles, always a great favourite of the Romans, out of a mere alphabetical listing of Greek artists and gives him a more prominent position than did Xenocrates, who apparently did not rate his work in marble highly.483 Varro also, it seems, gave a more favourable assessment of the writer Pasiteles, with whom he may have been personally acquainted, than Pasiteles perhaps deserved, since it seems likely that he may have simply rehashed Xenocrates and Antigonos.484 This assessment Pliny duly followed, assigning Pasiteles a more important position in the indices than his work may have merited. It is perhaps surprising that Theophrastus is cited more times than Aristotle in the Natural History. There is no doubt that both writers were hugely important sources in the writing of Pliny’s work, and that works written by both were used much more frequently than Pliny acknowledges openly.485 However, Theophrastus has the advantage of being a major source for two subjects, botany and mineralogy, while Aristotle is primarily used for zoology.486 Whatever the reason, Theophrastus remains one of the mostly highly esteemed Greek sources in the Natural History, with Pliny praising him on a number of occasions, describing him, for example, as the ‘first foreigner to write with care about the Romans’, ‘a most esteemed Greek author’, and ‘a weighty authority.’487 It is possible that Theophrastus appealed especially to Pliny because of his desire to bring order to the world of botany, as Pavord notes.488 This might well have resonated with Pliny’s own desire, expressed in the preface, to bring order to the whole broad world of knowledge.489

482 Sellers (1896: xiv) epitomises the former view, Coulson (1976: 364) the latter. 483 NH 69-71; cf. 1896: xxii ff. 484 NH 36.40; cf. 1896: lxxviii ff. 485 Lloyd (1983: 141 ff) notes a number of instances where Pliny cites Theophrastus virtually word for word without acknowledgement, e.g. 25.26, 25.48, 25.92. 486 Theophrastus is a major source in Books 12-27 and 33-7, Aristotle in Books 8-11 and 28-32. 487 NH 3.57, 15.1, 26.99. 488 2005: 21. 489 NH pref. 14. 88

Pliny makes use of at least three of Theophrastus’ works in the Natural History, the De Causis Plantarum, Historia Plantarum dealing with botany, and De Lapidibus dealing with minerals.490 Of these the latter two are the most heavily used. Historia Plantarum is the major source on botany used in the Natural History, supplemented by a few references from De Causis, while De Lapidibus is one of a number of sources on minerals. Pliny’s use of these sources largely follows the pattern uncovered with the sources already examined. What is perhaps most striking is the number of errors and differences between Theophrastus’ version and his own. For example, Theophrastus describes a plant that contains wool in its bulb and which grows near the sea.491 Pliny copies this quite literally, except that he states that the plant grows near rivers.492 Similarly Theophrastus describes a plant called krataigos, which Pliny says Theophrastus also calls aquifolium. Yet Theophrastus makes no mention of such a name, and indeed it seems to have been quite a different plant anyway.493 Perhaps the most significant error which Pliny made, and one which has even caught modern translators out, is his belief that Theophrastus said that a method of extracting cinnabar was found in the silver mines of Attica by an individual named Kallias of Athens.494 In fact, Theophrastus, who had earlier been discussing mining in Asia Minor, was continuing to discuss mines near Ephesus, and seems to have indicated that Kallias, who had been involved in mining in Attica, went to Ephesus and discovered his cinnabar extraction method there.495 Errors such as these raise questions, first about Pliny’s research methods, and second about the quality of the texts he was using. One important question is whether, in the case of Greek authors, Pliny was reading directly from the Greek or using a Latin translation. It could be argued that the size of the task that Pliny set himself and the frenetic nature of his research methods would tend to suggest that, for simple ease, he would have used a Latin translation. Possibly, as has already been noted, he was not using Theophrastus in the original, but as part of a compendium or through the aegis of some intermediary source. If he was using an intermediary source, the quality of his transcription would depend by its very nature on the quality of the intermediary’s transcription. Even if the quality of the

490 Books 3-11 contain a selection of material on animals, geography, art and mirabilia, which cannot be credited to any surviving work of Theophrastus. Diogenes Laertius (Lives 22-27), however, gives an extensive listing of Theophrastus’ writings, which includes no less than eleven works dealing with animals, though nothing is immediately identifiable as dealing with the other topics. 491 Theoph. Hist. Pl. 7.13.8. 492 NH 19.32. 493 Hist. Pl. 3.15.6; NH 27.63. 494 NH 33.114. 495 Hist. Pl. 58, 59; cf. Caley and Richards 1956: 199. There are numerous other errors of the same nature, to the extent that Lloyd (1983: 141) describes Pliny’s transcription of Theophrastus as simply ‘garbled.’ 89 text he was using was quite good, the nature of his research methods, his methods of dictating notes at all times and places, would seem very prone to auditory lapse and simple errors, as Stannard asserts.496 The question of the quality of the texts that Pliny was using may all come back to the somewhat convoluted story of the arrival of Theophrastus’ writings in Rome during the 1st Century BC. According to Strabo and Plutarch, Theophrastus willed his library, including the collection he had inherited from Aristotle, to one Neleus, whose descendants kept the library hidden underground for a long time, where they were quite badly damaged. Eventually the books were sold to one Apellicon, who had the misfortune to cross swords with Sulla during that general’s conquest of Athens, and had his collection confiscated and taken to Rome.497 Here they became the basis of Sulla’s own library in his villa, where firstly the grammarian Tyrannio, and then, more significantly, Andronicus of Rhodes, worked on the collection. Andronicus is generally credited with having produced the first reliable editions of Aristotle and Theophrastus for a Roman audience, as well as creating the accepted canon of works which is still used today.498 It is simply not possible to know whether Pliny was using the Andronicus translation or not. If he was indeed utilising a second-hand version through a compilation or some intermediate source, then arguably the chance of getting a garbled or incorrect version of Theophrastus’ work is greater. Then again, if Pliny’s own frenetic research methods, with the concomitant danger of errors through mishearing or making a mistake in notation, was to blame, the quality of the translation is to some extent a moot question. It seems quite extraordinary at first sight that Homer is the most-cited Greek source in the Natural History. Pliny makes much of his work being a practical source of information, and so, to modern readers at least, the idea of using poetry as a source of information for an encyclopaedia seems quite bizarre. Yet Pliny made much use of poets in writing the Natural History. Virgil is an important source, and Catullus and Ovid, amongst others, are also used. With regard to Homer, what must also be taken into account is the extraordinary respect that Homer engendered in Rome, and the impact that Homeric epic had on Roman culture, particularly in the upper levels of society. As Farrell points out, most Roman writers from Cicero to Fronto make wide use of Homeric quotations in their work, and Homeric study was an integral part of Roman education.499 Even among the lower echelons of society, Homer obviously had an impact, to judge from such evidence as the graffiti from Pompeii. The impact of Homeric epic on Roman life began very early in the history of Roman society, with Homer’s work

496 1965: 421. 497 Strabo 13.1.54 ; Plut. Sulla 26.1-3. 498 Cf. Kearney 1968: 293 ff; Huby 1969: 241 ff; Grayeff 1974: 69 ff; Dix 2004: 58 ff. 499 2004: 265 ff. 90 having been transmitted through Greek settlers in Southern Italy to the Etruscans. One of the very first works of literature in Latin was Livius Andronicus’ version of the Odyssey, dating from the 3rd Century BC. Pliny was obviously drawing on a well-established and long-lasting heritage of Homeric epic in his writing of the Natural History.500 As Farrell notes, the Romans had a great deal of respect for Homer as a repository of great wisdom and an authority in moral and ethical matters, but few utilised him for practical matters.501 Pliny, however, did find such a use for Homer in compiling the Natural History. It seems that Pliny had a very specific reason for using Homer’s Iliad and Odyssey. Simply put, he uses Homer as a source for anything that was considered of great antiquity. Homer fills in a gap for Pliny, in that his writing covers a period earlier than that covered by any of the other sources Pliny used. The overwhelming prevalence of the descriptors ‘ancient’, ‘very early’ and ‘as early as’ confirms that, for Pliny, Homer is the source par excellence for anything very old. Typically, when describing carpets as being in use at a very early time, he cites the Odyssey.502 When stating that warriors fought in linen corselets, he adds ‘as early as the Trojan War.’503 Lotus, saffron and hyacinth, were in use ‘already at the time of the Trojan War.’504 In another variation, he states that Helen was given Egyptian herbs ‘before Egypt even existed.’505 In other words, Homer was describing a period that seemed so far back in time that the world as the Romans knew it did not yet exist. If Pliny needs to describe something of great antiquity, then Homer is the most convenient source. Pliny registers his gratitude for this by praising Homer as the ‘prince of literature’, ‘the -head of all genius’ and the ‘first ancestor of ancient learning.’506 Pliny could in this case get away with using a poet as a major source for two basic reasons. First, because Homer predated all other sources, his work could not be gainsaid by any other source. He was, in effect, an absolutely impeachable source. Second, he was also literally unimpeachable, too, because of the enormous respect with which his work was treated in Rome’s literary circles. In some ways, King Juba II of Mauretania is the most remarkable of Pliny’s major sources. Although the idea of kings producing literature was not so remarkable in the ancient world, as Roller

500 Not surprisingly, given that Homer was so well-known and so ingrained in Roman culture, Pliny makes fewer mistakes with transcribing the Iliad and the Odyssey than with many of his other sources. One misinterpretation that does occur, however, is when Pliny mistakes Homer’s description of the Styx (Il. 2.751) as the ‘river of dread oath’, as a proper name and calls it the . 501 2006: 269. 502 NH 8.191;Hom. Od. 4.298. 503 NH 19.25; Hom. Il. 2.529, 830. 504 NH 21.34. 505 NH 25.12; Od. 4.210-34. 506 NH 2.13, 17.37, 25.12. 91 points out, Juba is interesting because he effectively straddles both the Greek and Roman worlds, to some extent at home in each.507 Raised in Rome in a household of Greek speakers, he spoke and wrote in both Latin and Greek. All of his books, with the possible exception of On Arabia, seem to have been written in Greek.508 He married one of the Ptolemies, and was effectively part of the Hellenistic world, yet he was a confidant of Augustus and travelled on a journey of exploration with Gaius Caesar. He certainly gained the respect of Pliny as a source, to the extent that when Pliny noticed a mistake in Juba’s writing, he put it down to copy error rather than any failing on Juba’s part.509 Pliny very presciently said that Juba would be ‘remembered more for his scholarship than his kingship’, and in this he was accurate, because Juba became a key source in Medieval Europe, but it was through the medium of the Natural History.510 Juba is known to have written at least nine works, though none have survived intact. More than half his works are represented in the Natural History.511 In fact, Pliny made such extensive use of these works that more than half the known fragments of Juba’s work come from Pliny.512 There are a few fragments from Roman , The Wanderings of Hanno, On Assyria, and most notably On Euphorbion, which, despite only being cited twice by Pliny, was arguably the most influential work that Juba wrote, because, as Roller notes, it became a fundamental part of modern botany and started an entire medical industry in the Middle Ages.513 The two most important of Juba’s works, as far Pliny is concerned are Libyka and On Arabia, which account for more than 90% of the fragments within Pliny’s work. These two works, in addition to the geographical knowledge they provided, were a significant source of mirabilia. Some of Pliny’s more fantastic stories, particularly concerning animals and plants, were taken from Juba. They included stories of crested elephants from Ethiopia, which strained even Pliny’s credibility, lions responding to pleas of mercy, and creatures such as Mantichores.514 However, Pliny also took some stories from Juba which ultimately turned out to be true. These include plants which produce wool (cotton) and transparent stone used to make windows (onyx).515 The issue of Pliny’s credulity has exercised many scholars over the years, with the general viewpoint today being relatively sympathetic towards Pliny. As Beagon notes, it would have been

507 2003: 163. 508 2003: 165. 509 NH 6.170. 510 NH 2.170. Cf. 2003: 260 ff. 511 Surprisingly, two of his works which are not represented in the Natural History are On Painting and Theatrical History, though both art and theatre are covered by Pliny. Roller (2003: 173 ff) argues that both were juvenile works of Juba, were cited by very few authors, and hence were probably not very significant. 512 2003: 164. 513 5.16, 25.77-9; cf. 2003: 178 ff. 514 NH 8.35, 48, 107. 515 NH 12.39, 36.163. 92 extremely hard for Pliny, lacking corroborating sources, and with so much of the world being completely unknown, to decide what was true and what was not.516 Steiner adds that, as far as writers in those days were concerned, anything that came from a distant locale was in a sense mirabilia.517 In the end Pliny apparently decided to be sceptical about some things but to accept others. Part of the question comes back again to something that has already been brought up numerous times, which is simply the exhaustion factor of Pliny’s task. The size of the task he set himself and his own frenetic research methods would not have allowed him to verify in any way, if that was at all possible, what was true and what was not. In the end, so much of what can be learned about Pliny’s library and how he used it comes down to the methods he used to write his books. His extremely hectic methods of taking notes and excerpts, on the run as it were, and in places that were not conducive to such work, seem to have coloured his usage of his library. He was apparently, through the sheer size of his task and the amount of information he collected, forced to produce some type of retrieval system for his vast collection of notes and excerpts, and this retrieval system ultimately affected how he used his sources and how he put them together in his work. His frenetic methods also appear to have contributed to numerous errors, mistakes and misinterpretations. The size of the task he set himself dictated how he would access his sources, in that he decided, or was forced, to use compilations and intermediary sources rather than relying completely on originals. This perhaps contributed to the errors.

516 1992: 11 n. 20. It should also be remembered that, with regard to Juba writing about Africa, Pliny himself is likely to have had some familiarity with Africa, as it is believed that he served as a procurator there during his career. Cf Syme 1969: 208 ff. 517 1955: 138. 93

Chapter 5: A Survey of Roman Libraries

This final chapter will present a summary of all that has gone before and a summation of all the existing knowledge about Roman libraries. However, it will not be a haphazard summary, but will be carried out according to the principles by which modern libraries survey and examine their own activities. As closely as possible, it will adhere to the methodologies which are utilised in modern libraries to gain a picture of how well they are serving their community. Obviously, at a remove of some two thousand years, it is not possible to survey Roman libraries in the same manner and for the same reasons that modern libraries undertake surveys. As will be seen, however, many of the principles and methodologies can be used, and it is to be hoped that through this survey a clear overall picture can be gained about how well Roman libraries functioned in serving their communities. A library survey, as defined by Line, is ‘a systematic collection of data concerning libraries, their activities, operations, staff, use and users, at a given time, over a given period.’518 As will be shown, this is perfectly applicable to the survey here. What Line’s definition lacks is a statement of the purpose of a library survey. The main purpose of a modern library survey, as defined by most authorities, is to gather information about the library with a view to ultimately improving its service. For obvious reasons, this cannot be the ultimate purpose of this survey, but the gathering of information for the survey becomes in this case the be-all and end-all. To this end, Gothberg lists two possible reasons for conducting a library survey which are relevant here. First, to provide an in-depth study of a library or group of libraries, and second, to gain a perspective on a library or group of libraries in relation to the history of the region or community.519 These, then, will be the objectives of this study. The planning of the survey is extremely important, as Line notes.520 The whole survey must be planned before a start is made. It must be decided what information is wanted, and what is envisaged as being the end-product of the survey. In the case of the present study, the first step does not really apply. As the study is totally reliant on the small amount of information about Roman libraries preserved in the ancient written sources and the archaeology, any information that can possibly be gleaned will be included in the study. The end product of this study will be a summary of the information available on Roman libraries in the hope of gaining an overall picture of the state of Roman libraries at their peak, their level of sophistication and their importance to the community. The

518 1982: 12. 519 1990: 554. 520 1982: 28. 94 other significant aspect of planning a library survey is choosing the methodology which will be used. The literature lists a number of different methods which have been used to undertake library surveys in the modern world. These include observation, self-investigation, distribution of questionnaires and interviews with concerned parties. For obvious reasons, however, none of these are practical for the purposes of this study. Only one method can be used in this case, viz. the use of documentary evidence. Line tends to be sceptical of the value of such evidence, because of the lack of standards for documentation which was created before the survey was planned.521 However, this should not be an insuperable problem for this survey. If the inherent and known bias of the literary sources, in this case predominantly upper-class Roman writers, is taken into account, as to a certain extent it was in Chapter 2, then valuable information can still be obtained from a survey of Roman libraries. As to exactly what aspects of library operation can be studied by a survey, Line lists seven, these being income and expenditure, stock, buildings and equipment, administration and technical processes, staffing, library use and background.522 This study will examine all these aspects, though for some, obviously, the information available will be scarce. Yet for the others, enough information exists to draw some definite conclusions about the standards of Roman library service and its impact on the community. No distinction will be made between public and private libraries in this survey. There are good grounds for this, both in the scarcity of general information available, and also the fact that Rome’s public libraries developed directly from private libraries and in their general level of sophistication shared many of the same characteristics.523 Obviously, there were some differences, in the level of staffing for example, and these will be noted when necessary, but for all intents and purposes this will be a survey of Roman libraries in general. McDiarmid notes that, since a library cannot be run without money, a survey of the library’s finances must of necessity be a priority.524 For that reason, the first element to be examined in the survey will be the finance and expenditure of the libraries of Rome. Obviously, information relating directly to the financing of Roman libraries is scarce, if not non-existent. However, by examining financing in general, and expenditure on public works in particular, it might be possible to glean some insights into how library finance and expenditure was managed. McDiarmid elaborates on how a survey of library finance should be conducted by noting that such a survey should look to answer four questions. First, what are the sources of library income? Second, what are the library’s expenditures?

521 1982: 48. 522 1982: 16. 523 Dix and Houston 2006: 672. 524 1940: 31. 95

Third, what is the cost of various library activities? Finally, what financial records and accounts does the library keep?525 Although not all these questions can be answered satisfactorily in this case, given the paucity of surviving evidence, there is still merit in examining them. In relation to private libraries, there is little mystery as to the source of finance. Obviously, a Roman aristocrat financed his own library from whatever personal source of wealth was available to him, usually his own fortune. Private libraries were certainly expensive to establish and run, as Cicero laments.526 However, aristocrats found ways to fund their libraries that did not exact a heavy toll on their finances. One method, which featured in Chapter 1, was to loot libraries from conquered cities, as did Sulla and Lucullus. This method, however, was not available to aristocrats who lacked conspicuous military records, including such notable bibliophiles as Cicero, Atticus and Varro. Lacking a foreign conquest, another less salubrious way to establish a library was to obtain one from a fellow aristocrat who had been proscribed. One such instance occurred when Marc Antony raided the collection of Marcus Varro while he was under proscription.527 There was also the possibility of obtaining the collection of another aristocrat who was in debt and needed to sell assets, as Cicero did when the library of Faustus Sulla was auctioned.528 Other methods of obtaining library resources without great expenditure included copying books belonging to friends, receiving gift copies from friends when their books were published, and by bequest, all practices which are known to have been utilised by Roman aristocrats to augment their libraries.529 However, none of these methods would have defrayed the cost of establishing a library, setting up rooms with shelves and cabinets to hold the books, and purchasing or paying staff with the skills to arrange and tend the collection. The brunt of this expense must have been borne by the bibliophile’s own financial resources. With the advent of the Imperial era and the appearance of public libraries, the funding of libraries becomes much less clear-cut. Certainly, it seems likely that the building of some public libraries was financed by foreign loot. It is known that Asinius Pollio funded his library in the Atrium Libertatis through manubiae from his Illyrian conquest.530 Augustus himself records that he funded the Temple of Palatine Apollo with its attached library through the spoils of war.531 Later in the 1st Century, it seems likely that Vespasian used loot from the Judaean war to build the Forum of Peace

525 1940: 8. 526 Cic. Fam. 13.77.3. 527 Gell. NA 3.10.17. 528 Att. 9.11; Plut. Cic. 27.3. 529 Dix and Houston 2006: 672. 530 Isid. Etym. 6.52. 531 RG 21.2. 96 with its library.532 Trajan used the proceeds of his Dacian campaign to build his forum containing the Bibliotheca Ulpia.533 This trend is in keeping with the practice of Republican aristocrats who used manubiae for the benefit of the public, and, not coincidentally, won public renown for themselves in the process.534 In this instance at least, library practice continued unchanged from the Republic through the early Empire. The day-to-day funding of the Imperial libraries is more problematic. There is no clear-cut evidence indicating the source of funding for the running of the libraries. No writer states directly that the libraries were funded from a particular source. The question is complicated even further by a long- standing debate about the finances of the Empire, the question of the aerarium and the fiscus. The aerarium was the state treasury, situated in the temple of , and, until the advent of Augustus, the depository for all state revenues. However, with the beginning of the , a large portion of the revenue began to accrue to the Emperor personally, and this eventually became the fiscus, the Emperor’s personal finances.535 It is generally accepted that the aerarium was eventually supplanted by the fiscus, that the finances of the Emperor and the State became indistinguishable in all but name, but that this did not largely occur until after the 1st Century.536 Further complicating the picture is the fact that some scholars see a further subdivision in the Emperor’s finances, between the fiscus proper, that part of the Emperor’s revenue that he was obliged to spend on public projects, and his patrimonium, that part that was his alone to spend as he wished.537 Yet regardless of these uncertainties, it can be determined with some confidence whence the funding for the Imperial libraries came. The first clues are provided by Augustus himself. As the first Emperor to build a public library, there is little doubt that he set the precedent which later Emperors followed. It is known that Augustus was scrupulous about not tapping the aerarium for funds for his building projects. Indeed he boasts in the Res Gestae of actually pumping money into the state treasury.538 What is notable about the first of Augustus’ libraries, the Library of Palatine Apollo, is its proximity to the Imperial palace. This proximity provides the first clue about how the library was probably funded. Its situation indicates that the library was clearly part of the palace complex, leading to the conclusion that the library would

532 Joseph. BJ 7.158. 533 Gell. NA 13.25.1. 534 There has been an ongoing debate about whether the use of manubiae for public benefit was compulsory or not. Cf Shatzman 1972: 177 ff; Churchill 1999: 85 ff. 535 It is generally accepted that, although the idea of the fiscus originated with Augustus, no use of the term in connection with the Emperor’s personal finances occurred before the reign of Claudius. See Sen. Ben. 7.6.3; cf. Sutherland 1945: 151 ff; Jones 1950: 22 ff; Millar 1963: 29 ff. 536 1963: 41. 537 1944: 51. 538 RG 17.1; cf. Sutherland 1945: 157. 97 have been funded as part of the Imperial household. This is confirmed by an examination of extant inscriptions recording the slaves and freedmen who worked in the library down to the time of Claudius and probably later. A number of them bear the designation ‘Caesaris’, indicating that they were members of the Familia Caesaris, the Emperor’s own staff of slaves and freedmen.539 This strongly indicates that the library was funded as part of the Imperial household, hence the funding would have come directly from the Emperor’s resources. This practice was continued by later Emperors for the libraries built subsequently, even when the libraries were freed from immediate proximity to the Imperial palace. Inscriptions for library staff down to at least the Hadrianic period indicate the use of members of the Familia Caesaris.540 This gives the strongest indication that funding for the libraries came from the Emperor’s finances and not from the state treasury. The Emperor drew his funds primarily from revenue from the Imperial provinces, but also from the practices of bona vacanta, bona caduca and bona damnatorum, as well as from fines, legacies and inheritances.541 Ultimately these were the sources from which the Imperial libraries were funded. McDiarmid lists as an essential part of a survey of library finances an examination of a library’s expenditures and also the cost of library activities.542 There is no doubt that establishing and running a library is expensive today and it seems certain that it was the same for the Romans. Several writers lament the cost of their libraries, though it seems that owners of Roman private libraries had a number of methods available to them to minimise the cost of establishing and running their own libraries.543 Perhaps surprisingly, it was not the cost of books that constituted the main expense, but the cost of staffing the library. The same circumstance pertains today, as the American Library Association notes.544 The cost of trained library staff could be formidable. Seneca records that a skilled copyist, essential in ancient libraries where the copying of books was the regular means of transferring books, could cost up to 100,000 sesterces.545 In fact, any slave who could read or write was bound to be more expensive than one who could not. It is not surprising, then, that aristocrats setting up their own libraries tried to circumvent the cost of hiring trained staff as much as possible. Cicero makes a great deal of use of the staff of his friend Atticus for work in his own library, at one

539 CIL 6.5188, 5189, 5884. 540 CIL 6.4233, 8679, 8743, 8744, 8907; 14.196; cf. Houston 2002: 141 ff. 541 Bona caduca and bona vacanta related to the property of individuals who died intestate or tried to leave their property to their wife instead of to a male heir. Bona damnatorum concerned the property of condemned persons. Cf. Millar 1963: 34 ff. 542 1940: 31. 543 Marshall 1976: 254-255. 544 The A.L.A. Standards (1933: 514) require for public libraries that 55% of the budget should be allocated for staffing, and only 25% for acquisitions. Although the advent of information technology has made some impact in recent times, that basic split would remain very similar today. 545 Ep. 27.6. 98 stage asking for ‘two of your library slaves for Tyrannio to employ to glue pages together and assist in general.’546 It could be argued that the acquisition of library staff in this period did not entail the ongoing expense of staff today because there was no need to pay wages. However, this is not entirely accurate. Beyond the initial purchase price of trained slaves, they had to be fed, housed and given medical care, if for no other reason than to protect the owner’s investment. There was also the issue of peculium, which, according to Florentinus, was either money or property granted by the owner for the slave’s use, or money or goods the slave was able to acquire through his own efforts.547 This in effect operated as the slave’s wage, and was in the master’s interest because, as Duff notes, it was generally believed that slaves worked better if they were earning money potentially towards purchasing their freedom.548 However, this need not necessarily have been a financial impost for the owner. As Joshell notes, a slave with marketable skills could earn income for his master through being hired out.549 It is not hard to conceive of someone such as Atticus hiring out his valuable library slaves to less well-off aristocrats of Cicero’s ilk. This in turn would have considerably cut the expenses of establishing and running a library. There is little doubt that the cost of acquiring books was much less for Roman library owners than for those who run modern libraries. This is mainly due to the fact that the publishing trade was in its infancy, and the concept of mass-producing books barely existed. Roman readers in the Late Republic and Early Empire simply regarded purchasing books from a retailer as one of a number of options for acquiring reading materials, and certainly, until much later, not the favoured option. The book trade did not help matters by having a reputation for producing unreliable copies, as Cicero warns his brother.550 The simple fact is that for a long period it was more attractive, as well as more practical, to acquire books by other means. Predominant among these was borrowing or being gifted copies through the extensive network of friendships which existed among the Roman aristocracy. Cicero records the borrowing and copying of books, especially from Atticus.551 With no copyright law of any sort in existence, there was no impediment to the free copying of books, the only expense being that of a literate slave, who, as was shown above, could be borrowed or hired, and the cost of a roll or rolls of papyrus. In addition, authors could obtain books, or even entire libraries, as gifts from

546 Att. 4.4a. 547 Dig. 15.1.39; cf. Sen., Ep. 80.4; Apul. Met. 10.14; Juv. 3.189. 548 1928: 16. 549 2010: 128. 550 Q. Fr. 3.5.6; cf. 3.4.5; Att. 1.7, 13.32. 551 Att. 2.20, 8.11, 8.12, 9.9; cf. Starr 1987: 217. 99 friends.552 Whole libraries could also be acquired for free through bequests.553 One expedient which does not seem to have been indulged to any great extent is the purchase of second-hand books. Although there are references to the purchase of used books in the sources, Starr argues that these refer to antiquarian rather than recent used books.554 The case is not entirely proved, since Gellius elsewhere mentions buying a large number of old and dilapidated books at Brindisi, the extremely cheap price seeming to indicate that these were not valuable antiquarian works.555 Nevertheless, it does seem that the buying of second-hand books was not common among the Roman aristocracy. Is it at all possible to make some sort of estimate as to how much a private library may have cost its owner? It is possible to come up with some sort of figure, if some moderately large assumptions are made. The basic equation comes down to the size of the library multiplied by the cost of the books. For both of those factors, there is some evidence available. For the size of the average Roman library, there is one fixed base-point, namely the library in the Villa of the Papyri in Herculaneum, which contained some 1800 scrolls. As has been noted before, it is possible that the Herculaneum library was not typical, because it was largely dedicated to the works of one author. However, there is some evidence that it was not unusual for libraries to contain a large number of works by a single author.556 There are references to private libraries that were considerably smaller than the Herculaneum library. Martial’s library, for instance, ran to only 120 scrolls.557 On the other hand, the library of Lucullus was large enough for both Cicero and Cato to sit in and be surrounded by scrolls.558 Therefore, it is possible to make an assumption that the Herculaneum library may have sat somewhere in the middle in terms of size. For that reason, and for the purposes of this study, it will be assumed that an average private library in the Late Republic or early Empire contained about 2000 scrolls, coincidentally the same number of volumes which Pliny said he consulted in writing the Natural History.559 The next factor to consider is the price of the books. There are few references to the price of books in the sources and there remains some debate as to whether books were cheap enough to be available to any but the wealthy.560 As might be expected, it appears book prices varied widely according to the size and the opulence of the book. Martial’s Xenia, a very short book, could be

552 Mart. 9.84, 14.183; Plin. Ep. 8.15; Att. 1.20.7, 2.1.12; cf. Starr 1987: 217; Marshall 1976: 255. 553 Ulp. Dig. 32.52. 554 1990: 149; cf. Gell. 2.3.5, 5.4.1; Luc. Ind. 102. 555 Gell. NA 9.4.1-5. 556 Suetonius (Vita Auli 1) records that the library of Persius Flaccus contained 700 volumes by Chrysippus. 557 Mart. 14.190. 558 Cic. Fin. 3.2.7. 559 Plin. NH pref. 17. 560 Philips 1985: 36. 100 bought for only 4 sesterces.561 On the other hand, a decorated copy of Martial’s poems cost 5 denarii.562 As it happens, however, Martial himself supplies a price which may be close to the average, a cost of 6 to 10 sesterces, consisting of the combined cost of transcription and of a roll of papyrus.563 If the difference is split, an average price of 8 sesterces is arrived at for the purposes of this study. Therefore, a cost for our hypothetical library can be arrived at by multiplying the number of scrolls by the average cost of a book, producing a total cost of 16,000 sesterces. Of course, this rests on a large number of assumptions, and does not take into account that in most cases, for the reasons outlined above, a library owner would not have purchased his entire library outright. Yet it does at least give some indication of the potential monetary value of a private library.564 Rome’s public libraries were no less subject to the costs of running libraries than the private examples, and on a much larger scale. One reference indicates that in the 4th Century AD, Rome boasted no less than 28 public libraries, a figure which compares favourably to the number of libraries offered by modern cities with substantial populations.565 In order to run these libraries, the Emperors employed a substantial number of skilled staff, with a distinct hierarchy, controlled by the procurator bibliothecarum, appointed by the Emperor to oversee all the libraries. Below him, in each library was either the bibliothecarius or a bibliotheca, who was either in charge of each individual library or each separate Greek or Latin section within the library.566 Below them were a variety of staff, skilled and unskilled, responsible for the upkeep of the collection, including copying and repairing books, shelving and re-shelving, and assisting library users.567 As already indicated, skilled and literate staff were expensive and the cost was doubled for public libraries because of the division of the libraries into Greek and Latin sections, so that each library required two sets of staff.568 Seneca emphasized the extreme expense of skilled copyists, and the fact that libraries would have had to employ numbers of them.569 Added to this cost was the extra layer of bureaucracy necessary to run a city-wide library system. It is known that Augustus was the first to pay a salary to his procurators, and this included the

561 Mart. 13.3.1. 562 Mart. 1.117. 563 Mart. 1.66.1. 564 The price of books may also have been a factor in the composition of the clientele of libraries, as noted in Ch. 2 565 Mirabilia Romae 23. Brisbane, a city of around 2 million people as of 2011, has 32 public libraries. 566 The term bibliothecarius, believed to denote the librarian in charge, was used by Marcus Aurelius in his correspondence with Fronto in relation to the Domus Tiberiana library. However, this term does not appear in inscriptions, where the term a bibliotheca is used. 567 Modern public libraries work on a staff per capita of population ratio, e.g. one full-time staff member per 3000 population served. Cf. Queensland Public Library Standards and Guidelines: Staffing Standard 2008. 568 Inscriptions indicate overwhelmingly that staff were identified by which linguistic section of the library they worked in. Cf. CIL 6.4431, 4433, 4435, 5884, 8679. 569 For example, when the Library of the Porticus Octavia burned down in 80 AD, Domitian was forced to send copyists to Alexandria to replace the lost works. See Suet Dom. 20. 101 procurator bibliothecarum.570 An inscription for L. Baebius Aurelius Juncinus, a procurator from the Hadrianic era, indicated that his salary was 60,000 sesterces per annum, which puts him in the lowest sexagenarius rate on the salary scale for procurators.571 This might indicate that libraries were not given a particularly high status or significance, but, as Mattingly warns, pointing out cases where provincial procurators received the lowest pay rate, the rate of pay does not necessarily indicate the importance of the position.572 Dix and Houston argue that Rome’s public libraries acquired their books in much the same way as the aristocrats had obtained books for their libraries.573 The main methods would have been acquisition by purchase, either individual titles or by block, through copying by trained slaves, and by gift or bequest. There is little evidence for the use of any of these methods. It is difficult to believe that the library staff went to the trouble of purchasing individual titles from a retailer, as this would have been a wasteful and time-consuming practice and it is not a practice used by modern libraries, except in very rare circumstances. However, the lack of a printing industry seems to indicate that it would have been difficult to block purchase titles, unless an existing library assembled by someone else was purchased. It has already been shown that this method was employed by private library builders, so it is certainly possible that the staff of the Emperor’s libraries also indulged in this practice when necessary. There is also little evidence of large-scale copying, though, as has been shown, the libraries did employ copyists. It is possible that such a practice was largely used when a damaged library had to be restocked. This was clearly the case when, as already mentioned, Domitian sent copyists to Alexandria after the Porticus Octivia library burned down in 80. However, it is possible that the public libraries did not have to spend a great deal on acquisitions. As Dix and Houston argue, rather than having to acquire books, they may literally have had books in large quantities thrust upon them.574 As was shown in Chapter 2, it appears to have been common practice for writers to present copies of their works to the public libraries, both in respect of the Emperor being the ultimate patron, and out of a desire to gain greater readership for their works. This would have supplied numbers of free titles for the libraries. It also seems that another way in which the Imperial libraries acquired books was through bequests to the Emperor. It was evidently common practice for citizens,

570 Dio Cass. 52.25.2, 53.15.5. 571 CIL 10.7580; cf. Pflaum 1960: 678 ff. From the second century there were four rates of pay for procurators, from highest to lowest, trecenarii, ducenarii, centenarii and sexagenarii, earning respectively 300,000, 200,000, 100,000 and 60,000 sesterces per annum. Cf. Hirschfeld 1963: 432 ff. 572 1910: 100; ILS 6930. 573 2006: 672. 574 2006: 673. 102 particularly wealthy ones, to bequeath at least a part of their estates. In the case of the early Julio- Claudian emperors at least, this practice was governed by self-imposed rules of ethics. Augustus, for example, refused to accept bequests from anyone unknown to him, while Claudius refused to accept bequests when the individual had living relatives.575 Nevertheless, it is recorded that Augustus expected to be remembered in the wills of his friends and clients, and was put out if this did not occur.576 While there are no surviving wills that mention the bequest of books to the Emperor, it is possible to identify some individuals who are highly likely to have done this. It is known that Virgil, for example, left a quarter of his estate to Augustus.577 It seems likely that if Virgil had a library, and as a writer it is almost certain that he would have had one, he would have left at least part of it to the Emperor, and hence to the Imperial libraries. Another possible candidate was Gaius Maecenas, Augustus’ close friend, advisor and literary patron extraordinaire. Maecenas left the entirety of his estate to Augustus.578 It seems certain that this patron of Virgil and numerous other writers would have had a large library, and almost as certain that he would have left this library to Augustus. Then it would probably have ended up in the Imperial libraries. In addition, there were other methods by which the Emperors may have acquired book collections gratis. There were the practices of bona vacantia, where the ownership of property was deemed to have lapsed, and bona caduca, where a property bequest was deemed legally invalid.579 In these cases it appears the property automatically reverted to the Emperor’s fiscus.580 Finally, there was outright seizure or confiscation of properties from persons condemned of certain crimes, where the proceeds also reverted to the Emperor’s fiscus.581 This is a direct follow-on from the proscriptions of the Republican era, but given a patina of legal sanction by the Emperors. It seems likely through all these methods that the Imperial libraries were able to obtain books virtually free of charge, and hence the cost of stocking the libraries was probably far less than that experienced by modern libraries. Stock-taking is an essential function in modern libraries, carried out with varying frequency depending on the size of the library and the turnover of its resources. McDiarmid lists a number of ways in which a library’s stock can be measured, but, for practical purposes, only the simplest is possible in this study, namely a numerical count of the number of volumes on the shelves.582 The

575 Dio Cass. 60.6; cf. Millar 1977: 154 ff. 576 Suet. Aug. 66.4. 577 Donat. Virgil 37. 578 Dio Cass. 55.7.5. 579 Cf. Millar 1977: 158 ff. 580 Tac. Ann. 2.48; Pliny, Pan. 42.1. 581 Suet. Aug. 41.1; Ov. Tr. 5.2.57; cf. Millar 1977: 163 ff. 582 1940: 100. 103 advantages of this method are apparent. It is simple to perform, easily understood and makes comparison with other libraries easy. Its disadvantages include the fact that it is quantitative rather than qualitative and that it fails to distinguish between titles and volumes, a problem readily apparent when studying ancient libraries, where books were often spread over more than one scroll.583 Nevertheless, it is possible to carry out this study for Roman libraries, based on the remains of the library buildings and an analysis of their shelving. Of Rome’s public libraries, only two have discernible extant remains. The most complete is the West Library of the Bibliotheca Ulpia in Trajan’s Forum. As Casson notes, enough of it remains to enable an almost complete reconstruction of the library.584 The floor, substantial pieces of marble, and, most importantly, the lower part of the walls survive. These reveal the structure of the all- important niches into which the armaria which held the books were placed. The size of these niches can give an indication of how many books the library probably held. At least one scholar has already attempted to calculate this number. Based on his analysis of the average rolled diameter of scrolls, Packer has come up with a figure for the total number of scrolls held in the Bibliotheca Ulpia. He argues that, with an average rolled scroll width of 5 cm, the niches in the lower part of the West Library, with measurements of 3.23 x 1.61 m, and with ten shelves per niche, would have held 32 scrolls on each shelf, or 320 in each niche.585 This gives a total of 4480 scrolls in the lower level. The upper level, also with 14 niches, but smaller, would have held 4032 scrolls. The niches in the western wall of the library add another 1900 scrolls, according to Packer, thus giving the West Library a grand total of some 10,400 scrolls. Adding an identical total for the East library, the total for the whole of the Bibliotheca Ulpia falls just short of 21,000 scrolls. It should be noted, however, that Packer’s analysis has not been accepted without question. For example, Houston and Dix have noted a number of variables in Packer’s calculations, the alteration of which would substantially have affected the final total.586 Packer himself notes that the shelves may have held double the number of scrolls he mentions, thus doubling the library’s total to 42,000 scrolls. The only other library in Rome for which enough extant remains survive to attempt a reconstruction is the Library of Palatine Apollo. Its location was disputed for a long time, but

583 1940: 102. 584 2001: 84. The East library also remains, but is more fragmented than its counterpart, and less accessible. There is, however, no reason to assume that it was not identical to its twin. 585 1997: 454. 586 2006: 708. They argue, for instance, that the niches could have held 15 shelves rather than 10, that there may have been free-standing bookcases as well, and that the average diameter of the rolls may have been less than the 5 cm which Packer cites. The last is certainly a valid concern. Packer himself notes evidence from other sources for rolls between 3 and 6 centimetres in diameter. Cf. Kenyon 1951: 66 ff; Callmer 1944: 153 n. 1. As a purely average figure, however, Packer’s estimate appears to be acceptable. 104 eventually a set of remains was discovered adjacent to the Domus Flavia and close to the Temple itself, which have now been generally accepted as the ruins of the Palatine Library.587 It consists of two identical adjacent halls, approximately 10 metres by 15 metres. Each hall has 18 niches together with one large niche in the rear wall, presumably for a statue. The niches measure 3.8 by 1.8 by 0.6 metres. From this figure it is possible to work out, using Packer’s methodology, a possible stock count for the Palatine Library. The niches being slightly larger than those in the Bibliotheca Ulpia, it seems an estimate of 12 shelves per niche is appropriate. Using Packer’s formula, a number of 36 scrolls per shelf, and 432 scrolls per niche is arrived at. Extrapolating from this to one of the halls, a figure of 7776 scrolls per hall is arrived at. Doubling this gives a figure of 15,552 for the whole library. Not enough remains to determine whether the Palatine Library had an upper level. If it did, then the figure for the library would be in excess of 31,000 scrolls. It must be remembered, of course, that this does not represent a total book count, but only a count of scrolls. This is similar to the disadvantage mentioned by McDiarmid, in that a book count only gives a total of volumes, not titles. A count of the titles in a Roman library would be much less than the count of scrolls. As Kearney notes, a single scroll could only contain part of a sizeable work.588 Therefore, it can be assumed that the actual count of titles for the Bibliotheca Ulpia and the Palatine Library would have been very much less than the volume count. If an average of three scrolls per title is assumed, then the title count for the Bibliotheca Ulpia would be in the region of 7000 books (or 14,000). For the Palatine Library, the figure would be in the region of 5000 titles (or 10,000). These are not unreasonable figures, and stack up well against modern suburban public libraries.589 Although there are only two public libraries in Rome with enough extant remains to make an estimate of their holdings, it is worthwhile noting that there are libraries outside the city which also have remains sufficient to estimate a book count. Examination of these libraries may provide a more balanced picture of what the average book count of a Roman public library might have been. The libraries in question are the Library of Hadrian in Athens, the Library of Celsus in Ephesus, and the Library in Timgad, North Africa. The Library of Hadrian was a donation by that Emperor to Athens in about the year 132. The library consists of a large colonnade with the book collection area in a large central room at the rear. It is possible that the library had as many as three levels, with up to sixty-six

587 Cf. Callmer 1944: 157 ff; Makowiecka 1978: 31; Strocka 1981: 307 ff. 588 1951: 64 ff. Kearney cites the example of the Iliad, where perhaps two or three books only could be contained on a single scroll. 589 The Australian Public Libraries Statistical Report 2008-2009 (2010: 5) gives an average of 28,029 library materials per library. However, since this includes materials other than books, and the bulk of public libraries in Australia are small suburban establishments, with only a small proportion of larger flagship city branches, the average for a suburban library would probably be close to 10,000 books. 105 niches.590 For this reason, quite large estimates of the size of the collection have been made. Gotze, for example, suggests a collection of 22,000 scrolls.591 Stracka makes a slightly smaller estimate of 20,000.592 However, Packer argues that the niches only had eight shelves in each, and that, allowing for the frames and wooden cases of the armaria, they would have held only 120 scrolls each, giving a total for the library of 7920 scrolls.593 The library itself is quite substantial, measuring 23 meters by 15 meters, an area which compares favourably with the Bibliotheca Ulpia and the Palatine Library. However, crucially, it was not paired, possibly because, being located in Athens, there was little demand for Latin books, so a bi-lingual establishment was not deemed necessary. Therefore, there are some grounds for arguing for a lower figure, though perhaps not as low as Parker indicates. The Library of Celsus in Ephesus is the best-preserved Roman library remaining today. Enough of it exists for the structure to have been reconstructed up to the ceiling, and it has the requisite niches on all three of its levels. There are thirty niches in all, with dimensions of 2.8 metres by 1 metre by 0.5 metres. Estimates of the size of the library’s collection vary widely. Johnson makes an estimate of 9500 scrolls, equating to 730 books.594 Casson, however, makes an estimate of only 3000 scrolls.595 If the difference is split, the collection would seem to be in the region of 5000 scrolls. Again, as in the case of Hadrian’s Library, there was no apparent pairing for this library. The other remaining Roman library is the one at Timgad in North Africa. It might be the latest Roman library whose remains have survived, since it was probably erected in the 3rd Century, and may show how library structures changed over time, as it differs from other existing provincial libraries in being apsidal and roofed with a dome. Like the other known provincial libraries, it was not paired, though, unlike the libraries of Hadrian and Celsus, which would have been oriented towards Greek literature, this library would have been largely Latin in character, as Casson notes.596 It was also quite small by Roman standards. With only eight niches, its collection may have been no more than 2500 scrolls, and quite possibly many fewer. An estimate of up to 23,000 scrolls has been made, but this presupposes a second level, for which no evidence has been found.597 The study of the size of libraries leads into the next area to be surveyed, namely that of buildings and equipment. Line lists five aspects to be considered. They are size, lighting, shelving,

590 Sisson 1929: 59; Callmer 1944: 174; Strocka 1981: 318; Casson 2001: 113. 591 1937: 242. 592 1981: 318. 593 1997: 454. 594 1980: 652. 595 2001: 116. 596 2001: 118. 597 Cagnat 1906: 106; Pfeiffer 1931: 163. 106 work space and reader space.598 As the question of the size of library buildings has been largely addressed in the previous section, it is worthwhile to move on to an examination of the lighting of Rome’s libraries, a topic which has only been lightly touched on in research on ancient libraries, and, surprisingly, in the study of modern libraries as well. Lighting is one of the most fundamental yet little discussed aspects of library design. By their very nature, libraries are entirely dependent on good lighting, but little regard has been paid to the lighting of ancient libraries. It is a simple fact that ancient libraries had only two options for lighting, daylight or fire, and, for this reason perhaps, there has been little discussion of their lighting. These means should not be regarded as inferior options because they were the only ones available, natural lighting in particular. With regard to natural lighting, modern libraries are coming to use it as a first option rather than as a poor second to artificial lighting, as Waters and Wonters note.599 A number of modern libraries have been built with an emphasis on natural lighting in their design.600 Natural lighting is highly regarded in modern library construction not only because it is cost-efficient and environmentally friendly, but because it has been shown that it contributes to a high degree of satisfaction from library users.601 Natural light is seen as pleasing and a significant contributor to the attraction of a library. Given that ancient libraries would have relied so heavily on natural lighting, it should not therefore be seen automatically as simply a necessary option forced onto them through lack of alternatives, but quite possibly as a positive aspect of their nature. As far as private libraries go, the only physical evidence regarding lighting comes from the Villa of the Papyri in Herculaneum. As has been previously noted, the library here was merely a storeroom, and reading was probably undertaken in the nearby colonnade, which would have been ideal for daylight reading. This scenario is supported by literary evidence regarding the library of Lucullus, which records reading and discussion taking place in the colonnades surrounding the library.602 On the other hand, Cicero indicates that he found Cato in the act of reading inside Lucullus’ library, so it appears that the library itself was well enough lit to permit satisfactory reading.603 There is ample evidence of Romans reading, or being read to, in circumstances where artificial lighting

598 1982: 16. 599 1987: 336. 600 Rastorfer 1985: 158. Examples include the San Juan Capistrano Regional Library in California, the Conrad Sulzer Regional Library in Chicago and the Mt. Airy Public Library in North Carolina. Cf. Waters and Winters 1987: 337 ff. 601 1985: 154. 602 Plut. Luc. 42.1. 603 Cic. Fin. 3.7. Cicero (Top. 1.1) also mentions talking with Trebatius in his own library, so presumably that library was also more than merely a storeroom. 107 would have been necessary, so it is probably safe to say that lighting was not an issue in small private libraries, and that lamps would have been commonly used to supplement natural light.604 With regard to public libraries, the only structure in Rome with enough extant remains to draw some conclusions about how it was lit is the Bibliotheca Ulpia. Although nothing of the roof survives, there has been much conjecture as to how it was constructed, which incidentally impacts on how many and of what size were its windows. There are two main schools of thought. The majority argue for a tiled flat roof supported by beams and rafters.605 This would dictate, as Packer indicates, either a row of small windows above the wall niches, or large windows in lieu of some of the niches.606 Packer himself argues instead for a barrel vaulted roof, which would permit a full complement of wall niches, with large semi-circular windows on all four sides.607 Packer maintains that such windows would have provided pleasing indirect light without fear of direct sunlight damaging the books.608 He challenges the main evidence for a flat roof, which is provided by a 16th Century drawing by Du Perac, purporting to show the remains of the East Library. Meneghini cites this drawing as evidence of a flat roof supported by beams, claiming that it shows sockets and brackets that could have supported the roof beams, and that the walls were too thin to have supported vaults.609 This argument is supported by Bianchi, who claims further that the Du Perac drawing shows a staircase three levels high, indicating that the library itself was at least three storeys high, a height which would favour a flat roof.610 Packer counters by arguing that the scale of the Du Perac drawing is not necessarily accurate and so what is shown may not be the original remains of the Library.611 He also points out that the remains of the West Library indicate that the walls, with a thickness of between 2 and 3 metres, would easily have supported a vault, and that they were in keeping with the Roman builders’ maxim that a wall must be 1/10 the width of the span of the vault in order to support it.612 Modern library practice would tend to support Packer’s thesis. Waters and Winters state that the best use of indirect lighting in libraries comes when the openings are placed as high as possible.613 The Romans certainly understood the importance of proper location and orientation for suitable lighting of libraries, so it can be assumed

604 Cic. Fam. 9.26.1, 14.21.4; Plin. Ep. 3.1.4. 605 Amici 1982: 82; Meneghini 2001: 250; Dix and Houston 2006: 698. 606 2001: 58. 607 1997: 454. Blake (1973: 15), also argues for a vaulted roof, but does away with upper niches altogether, instead favouring an upper walkway that would have allowed library users to savour the view through the windows. 608 1997: 125. 609 1993: 18 ff. 610 2001: 267. 611 2003: 130. 612 2003: 131; cf. DeLaine 1997: 56 ff. 613 1987: 343. 108 that in general they chose the best possible solutions. Allied with the archaeological evidence, it seems more likely for these reasons that the libraries were roofed with barrel vaults. Did the public libraries in Rome make use of artificial lighting? Makowiecka is of the opinion that they did not, since the use of open flame in an environment consisting of thousands of very flammable papyrus rolls was simply too hazardous.614 However, as has already been shown, the Romans were not afraid to use lamps in their personal libraries, and it is difficult to believe that the public libraries would have needed a total ban on all artificial lighting. After all, up until the late nineteenth century, all libraries needed to make use of flame for lighting. Undoubtedly, accidents did happen, and conflagration resulted. However, with proper care, there is no reason why the public libraries in Rome could not and did not make use of artificial lighting, on cloudy days, for example, and during the short, dark days of winter. The use of artificial lighting might in fact have depended on the opening hours of the libraries. It seems certain that they were only open during daylight hours, but for how long each day? The only firm evidence surviving from the ancient world comes from an inscription from the Library of Pantainos in Athens, probably placed over the door, which indicates that the library ‘will be open from the first hour until the sixth.’615 This indicates that the library was open from roughly 6 am until noon. Whether hard and fast opening hours such as this would have been feasible in Rome is uncertain, given that in midwinter a 6 am start would have been at least an hour before sunrise. It seems more likely that opening hours would have been more flexibly based on the season and the rising and setting of the sun. It also seems likely that the lighting in the major libraries was sufficient for them to stay open as long as daylight permitted.616 On those days when overcast conditions prevailed, it certainly seems feasible that the libraries made some use of artificial lighting, as modern libraries do, even those that make extensive use of natural lighting. One aspect of the libraries’ operations that may have determined whether or not artificial lighting was needed is the question of whether the books could be borrowed or not. This is of course a fundamental and accepted practice in modern libraries, but it needs to be remembered that it is a very recent practice. Borrowing from libraries was almost unheard of before the nineteenth century, and it is really only the advent of modern mass production of cheap books that has made it feasible. Evidence of borrowing from Roman libraries is almost non-existent. Two quotations are often cited as evidence that books could be borrowed, but each is problematic. The first is from the letters of Marcus

614 1978: 60. 615 Oldfather 1937: 437. 616 The libraries of the Bibliotheca Ulpia, for example, almost certainly had windows on all sides, allowing an equal amount of lighting in the morning and afternoon. Cf. Packer 1997: 125. 109

Aurelius to Fronto, wherein he boasts that he has been able to borrow two books from the Library of Palatine Apollo and take them with him to the country.617 However, as Hall argues, this cannot be taken as prima facie evidence that books could be freely borrowed.618 He makes the point that a member of the Imperial family was unlikely to be refused such a request in the Emperor’s own library. The second reference, which comes from Vopiscus’ biography of Aurelian, mentions that the City , Tiberianus, offered to lend books from the Bibliotheca Ulpia to Vopiscus.619 Again, this does not provide firm evidence of a general policy of borrowing, since it seems likely that, as with the Imperial family, a high-ranking official such as the City Prefect would hardly be refused if he made such a demand of library staff. However tenuous is the evidence in favour of borrowing, the evidence that attests to borrowing being banned is even more so. There is in fact only one piece of evidence that supports this case. The same inscription from the Pantainos library in Athens that gives the library hours also declares that ‘no book shall be taken out, for we have sworn an oath.’620 This reads as a definitive statement, but, as Oldfather notes, it appears to be a statement peculiar to the circumstances of that library and therefore cannot be taken as prima facie evidence of a general ban on borrowing.621 The best evidence that Roman libraries in general did not allow borrowing comes from later centuries. It is known that the borrowing of books only became a general phenomenon at a very late stage, in the nineteenth century, after the development of public libraries and the advent of cheap mass-produced books. There is no evidence to suggest that Roman libraries would have been any different from the libraries of medieval monasteries or universities, where books were also hand-written and too valuable to lend. If the libraries did not allow borrowing, as seems likely, then the most obvious practice would have been to allow the books to be copied. This was an established Roman practice. It is known from the late Republic that Roman aristocrats were inveterate copiers and lenders of their books to friends, and that they went to the trouble of buying or hiring expensive copyists to accomplish this. As the public libraries largely followed the practices of the private owners, it seems possible that the same thing happened. Either patrons desiring a book brought their own copyist or possibly used the library’s own staff to make copies. Inscriptions reflect the existence of copyists in the public libraries, as does the reference by Suetonius to Domitian sending copyists to Alexandria to obtain replacements

617 Fronto, Ep. 4.5. 618 1922: 31. 619 Vop. Aur. 1. 620 Oldfather 1937: 437. 621 1937: 437. 110 after the Porticus Octavia library was damaged by fire.622 Alternatively, if the patron did not want to copy the books, he could simply read them within the library. There is literary evidence to support this practice, especially reading and discussing the books with friends in situ.623 In this sense the public libraries would have differed little from their predecessors in the Greek world, the gymnasia, which were established as centres for the discussion of intellectual matters. In either of these cases, however, copying or reading in situ, the libraries would have needed to be well-lit. Therefore, it seems likely that on dark days some use of artificial lighting must have been made. This would have been especially necessary if books were copied on the spot, a practice which would demand very good lighting. It can be assumed therefore, that, as in the private libraries, the threat of fire was not enough to deter the use of lamps, with all practical safeguards being utilised. While daylight is an excellent source of lighting for libraries, as is increasingly being recognised in the modern world, it has its drawbacks. It thus seems that the Roman public libraries would have needed to make use of both natural and artificial lighting to fulfil their purpose to the maximum possible efficiency. Another fundamentally important aspect of library operations is shelving. Shelving as a means of storing library resources has existed in one form or another since the very first libraries were constructed in Mesopotamia.624 It remains very important in modern libraries, regardless of the advent of modern non-print media. Modern libraries have a number of different shelving options to call upon, but as far as the Romans were concerned, there were basically two.625 It is these two different shelving options which will be explored in this survey. The two options available to the Romans were simply standard wall shelving, and cabinets or bookcases (armaria). Conventional wisdom has generally held that the first option was more suitable, and more widely used, for the storing of scrolls. Armaria, on the other hand, are supposed to have come into prominence when the codex took over.626 Yet the evidence does not entirely support this conclusion. The only physical remains of library shelving which have survived until the modern era are those of the Villa of the Papyri in Herculaneum. The library itself dates no later than 79 AD, of course, and probably from a good deal earlier, most likely the late 1st Century BC. This was well within the era of book rolls, and the codex form had barely made an impression. An examination of

622 CIL 6.6314, 9520, 9525; Suet. Dom. 20 623 Gell. NA 11.17; 13.20; cf. Houston 2004: 8 ff. 624 The royal palace at Ebla contains what is probably the earliest known library, dating to 3000 BC, which consists of clay tablets arranged on wooden shelves around the room. Cf. Weihs 2009: 12. 625 Boll 1980: 16 ff. 626 Petroski 1999: 26ff. 111 eyewitness accounts of the library in the Villa, however, seems to confirm that the shelving was some type of cabinet or press. Paderni, in his first-hand account of the library, uses the word ‘press’ to describe the type of shelving used.627 A decade later, another first-hand account by Winckelmann describes ‘book-cases’ as being present in the library.628 Both of these accounts would seem to refer to some free-standing case or cabinet, which sounds closer to the conventional description of armaria. This would appear to be backed up by the Oxford English Dictionary, which describes a press as a ‘large (usually shelved) cupboard, especially placed in a niche in the wall, for holding linen, clothes, books, etc.’, and provides an example from 1693 of a mention of ‘5 duble preses for books.’629 This clearly indicates that, at the time, a press was understood as a cupboard or cabinet, so that Paderni’s description of what he saw in the Herculaneum library must be taken to mean exactly that. There is in fact little evidence to suggest that the Romans made use of wall-mounted or free- standing shelving to house books. The evidence that does exist is assembled by Clark to make a case that pegmata (shelving fixed to the wall) or plutei (free-standing shelves) were used to hold books, but the literary evidence is limited.630 Cicero makes mention of workmen belonging to Atticus installing pegmata in Cicero’s library.631 There is even less evidence for plutei, which is not a term used in connection with a library, but with shelves used to hold busts.632 The Oxford Latin Dictionary makes no connection between plutei and the storage of books, defining the concept as originally referring to military equipment, but as being subsequently transferred to furniture as ‘the upright board forming the back, or far side of a couch.’633 If this is referring to some type of side or backboard, it is physically possible that it may have contained some sort of shelving capable of holding busts, for example, as is indicated by Juvenal, or perhaps even a few books. Definite evidence, however, is lacking. The physical evidence for wall-shelving is even more limited. Basically it comes down to a room in the House of Menander in Pompeii, which is identified by Richardson as having been a library at one stage because of the presence of holes in the walls indicating that a set of shelves had been fixed there.634 This indeed matches the description of pegmata, but there is no evidence that books were ever stored there. In fact, the archaeological finds from the room, consisting of small glass bottles and other

627 1753: 823. 628 1831: 197. 629 OED online. 630 1909: 35. 631 Att. 4.8. 632 Juv. Sat. 2.4. 633 1968: 1394. 634 1977: 398; cf. Maiuri 1933: 88 ff; Richardson 1988: 160. 112 household items, seem to indicate that the room was merely a storeroom.635 It can be concluded from the limited literary evidence that pegmata were sometimes used in the private libraries of bibliophiles such as Cicero, but equally evidence exists for the use of bookcases and cabinets. The simple reason why the Romans appear to have preferred cabinets and bookcases to wall shelving was the adoption of the wall niche. As previously outlined, the niche is one of the unique and defining features of Roman libraries. Niches are found in all surviving remains of Roman public libraries and in the majority of libraries in large villas, especially those occupied by the emperors. In fact the presence of niches is such a defining aspect of Roman libraries that there has been a tendency to identify some buildings or rooms as libraries because they possess niches, even though niches were also used for other purposes, most notably for statues.636 The use of niches in libraries because of their very nature favours the use of cabinets and bookcases over wall shelving, and it seems almost certain that the presence of niches in a library meant that such furnishings were used. In at least one case the physical remains of such cabinets were found in the niches of a public library.637 The argument that armaria only came into general use with the adoption of the codex tends not to hold up when the literary evidence is examined. There are a number of references to armaria dating from the late Republic up to the early 2nd Century, before the codex had supplanted the scroll.638 It has been suggested that the armarium was not entirely suitable for holding scrolls, but it appears that Roman bibliophiles made adaptations to the armarium to make it more suitable. Clark lists a number of words that are used in conjunction with bookcases. These words are nidus, forulus and loculamentum.639 There are some variations, but basically these words refer to pigeon-holes and cells of bee-hives, and appear to have been largely adapted from agricultural use to describe adaptations to bookcases which made them suitable for holding scrolls.640 It is easy to imagine that a pigeon-hole arrangement dividing up the shelves of a cabinet or bookcase would make it more amenable to holding scrolls. It seems likely that early versions of the armarium were adapted to holding scrolls and only became more like modern bookshelves when the codex form came into general use. The question of shelving in a library leads inevitably to discussion of the use of space. In most libraries, there are two distinct types of space, public space and work space. The first is largely self-

635 Allison 1997: 308. 636 Casson 2001: 90. 637 When the remains of the Bibliotheca Ulpia were excavated in the 19th Century, metal parts of such cabinets, including hinges, were discovered to have survived in some of the niches. See Nibby 1839: 188 ff. 638 Vitr. 7 pr. 7; Sen. Dial. 9.9.6; Pliny, NH 35.6; Pliny, Ep. 2.17.8; Juv. 7.11. 639 1909: 32ff; Mart. 1.117, 7.17; Juv. Sat. 3.219; Suet. Aug. 31.1; Sen. Tranq. 9.7. 640 Columella (Rust. 8.8, 9.12), for instance, uses loculamentum to describe both a compartment in a dove-cote and a bee- hive. 113 explanatory and refers to the space in the library to which the users have access. At first it would seem that this space is mainly devoted to the collection and consists of the area which is given over to library users for reading. However, modern libraries increasingly are utilising their public space for many things beyond simply the use of the collection for reading purposes. The public space of a typical public library these days houses facilities for computer use, displays, rooms for clubs and groups to assemble for community meetings and discussion, and play areas for children. The public has responded to these facilities by increasingly identifying libraries not simply as places to read, but as places in which to socialise and take part in joint activities, as has been noted in recent surveys.641 However, lest it be thought that such activities are entirely modern, there is considerable evidence that the Romans also made use of the space in their libraries for purposes other than simply reading. In fact, it appears that modern have merely revived practices in which the Romans routinely engaged in their libraries. It has generally been an article of faith among scholars that Rome’s (supposedly) low literacy rate meant that the public libraries were little used except by the elite. That is probably accurate, if it is assumed that the only purpose of the libraries was to supply reading material. However, considerable evidence exists that Rome’s public libraries provided more to the community than books, and that in fact they were used for a variety of purposes. It also appears likely that such usage began with the very first of Rome’s public libraries. As Dix notes, while Rome’s libraries may not have hosted a multitude of readers, they may well have been full of listeners instead.642 There is substantial literary evidence which indicates that public recitation of authors’ works was carried out in the libraries. This is certainly an obvious means of getting around the literacy barrier and the cost barrier which were noted in Ch. 2 and may well have thrown the libraries, and the authors’ works, open to a wider audience. The man credited with introducing public recitations is Asinius Pollio, the creator of Rome’s first public library in the Atrium Libertatis.643 It seems almost certain that his innovation with regard to public recitation coincided with his establishment of the library in the Atrium Libertatis, and that therefore his library became a venue for these recitations. There were precedents in the Greek world for using libraries as a place for public recitation, for example in the Library of Alexandria.644 In Rome, where large venues for recitations of

641 Goulding 2009: 37 ff; Aabo, Audunson and Varheim 2010: 16 ff; Dahlkild 2011: 11 ff; May and Black 2011: 5 ff. 642 1994: 287. 643 Sen. Contr. 4 pref. 2. It is now believed that, rather than introducing recitations, which had probably been brought to Rome by Greek writers in the 3rd Century {Suet. Gramm. 2), Pollio’s achievement was instead to present public recitation on a more formal basis. Cf Dalzell 1955: 26 ff. 644 Vitr. De Arch. 7 pref. 4-7. 114 works were lacking, it seems logical that the open space of the library, already possessing an association with books, would have seemed ideal.645 Public recitation continued to be practised in public libraries constructed subsequently. Horace records a competition between poets in the Library of Palatine Apollo, while Pliny records that Claudius overheard a recitation by Nonianus while walking on the Palatine. The location is unspecified, but it could have been any of the three libraries known to have been located on the Palatine.646 Juvenal refers to a particularly unsavoury patron who has abandoned the Palatine library for recitation and offers his protégés squalid venues instead.647 It seems clear that the libraries offered a very suitable location for recitation. Dix asserts that the libraries must have provided rooms especially for recitation, but this is not supported by the archaeological evidence.648 It seems more likely that the recitation was carried out in the library itself. Both of the existing library remains, those of the Library of Palatine Apollo and of the Bibliotheca Ulpia, appear to have been large enough library halls to accommodate public recitation, and it seems likely that, when a recitation was scheduled, the furniture in the library was simply moved to accommodate seated or standing patrons. This solution was of course easy because the shelving in Roman public libraries was located along the walls, leaving a central space clear for readers, and, when needed, for an audience. It could be argued that public recitation was in fact the most ‘public’ function offered by Rome’s libraries. Since literacy was not an issue, it is possible that the recitations attracted a more diverse audience than that usually found reading. The Emperor Augustus found another use for his libraries when suffering from ill-health late in his reign. Rather than travel to Senate-meetings in the Forum, Augustus chose to hold Senate- meetings on the Palatine adjacent to his palace.649 The venue he chose is almost certain to have been the library.650 The library, with its large open space conveniently divided into Greek and Latin sections, useful for the purpose of divisions, would have been the perfect location.651 This practice

645 There were, of course, other venues. Dalzell (1955: 28) notes that the Forum and baths would probably have been used, and Horace (Ep. 1.19) mentions recitations taking place in the theatre, though it seems likely that the association between libraries and books would have made libraries a particularly attractive venue. 646 Hor. Ep. 2.2.92-105; Plin. Ep. 1.13.3. 647 Juv. Sat. 7.36-53. 648 1994: 287. 649 Suet. Aug. 29.3. 650 Augustus also used the library for other public functions. In 13 AD he is recorded as hearing an embassy form Alexandria there. See P. Oxy 2435; cf. Millar 1977: 19. 651 Thompson (1981: 338) notes that the ruins currently identified as the Library of Palatine Apollo would have been unsuitable, as they are divided into separate halls. This structure, however, is almost certainly a Domitianic reconstruction, the original library having been destroyed in the fire of 64. It seems likely that the original library had one hall divided into its Greek and Latin sections. Cf. Richmond 1914: 201. 115 was carried on by subsequent emperors. At least two Senate-meetings were held here by Tiberius. The first of these, in 16 AD, definitely identifies the venue as the library, because Hortalus, who was attending the meeting to seek financial aid, is recorded by Tacitus as glancing at the orators’ portraits decorating the walls.652 Pliny records that such portraits were placed in the public libraries, and the Horatian scholia specifically record these imagines being placed in the bibliotheca Apollinis.653 Subsequent meetings of the Senate were held here in the reigns of Claudius and Nero.654 While no evidence is given that these meetings were indeed held in the library, Thompson argues that there is simply no evidence for any other suitable venue on the Palatine, and that the precedent established by Augustus of meeting in the library is likely to have been followed by subsequent emperors.655 It is perhaps significant that no Senate-meetings on the Palatine are recorded after Nero’s reign, when the original library was destroyed. When its replacement was erected under Domitian, it would have been, as noted above, unsuitable for such meetings, and thus the practice ended. The use of public libraries for government purposes is increasingly the case today. Local councils are using public libraries as avenues to distribute information to the community. A good example of this is the Gateshead community in the north of England, where the local council has evolved a programme of using the public libraries as ‘one-stop shops’ for council information and advice.656 While ideas like this are somewhat different to the Roman idea of holding Senate-meetings in the library, it is clear that the Emperors from early on envisaged their libraries as offering more to the community than just the availability of books to read. This is demonstrated by another function fulfilled by the libraries in Rome, that of offering a place to socialize. Surveys of users of modern public libraries have indicated that, increasingly, users see libraries as a place to meet and to socialize.657 Library users have identified public libraries as a safe place to meet, to engage in group activities and simply to socialize with friends. This is in keeping with the concept of the ‘third place’, developed by Ray Oldenbourg. This elucidates the concept of an increasing interest in modern society of finding a place to meet away from home or work.658 Although Oldenburg does not mention libraries in his list of ‘third places’, modern researchers on library use are increasingly of the opinion that public libraries are coming to be seen as such.659 However, it appears

652 Tac. Ann 2.37. 653 Plin. NH 35.9; [Hor.] ad Serm 1.4.21. 654 Joseph. AJ 19.266-268; Dio Cass. 50.3.2; Tac. Ann. 13.5. 655 1981: 337. 656 Goulding 2009: 44. 657 May and Black 2010: 8; Aabo, Audunson and Varheim 2010: 25. 658 1999: 16. 659 Goulding 2008: 42. 116 that such a concept was already known in Rome. It is clear that Rome largely lacked the ‘third spaces’ common in modern society. There were no cafes or hair salons, bars were unlikely to be patronised by the literate classes, and bookstores were rudimentary. It appears that the main ‘third place’ was the baths, though those with literary interests appear to have added the public libraries as another venue to meet.660 Evidence exists of Roman writers and others with literary interests meeting in the library for discussion. Gellius relates how he and his friends were seated in the domus Tiberiana library when a volume of Marcus Cato Nepos was brought to them. This precipitated a discussion which was joined by a young man unknown to Gellius.661 It is clear that the library was a place where discussions on literary matters took place on an ad hoc basis. This is not out of keeping with the earlier Greek models, where discussion was carried out in the exedrae joining the library. The difference with the Roman libraries was that the discussion was being carried out in the library itself. The next area to be examined in this survey is that of technical services. Technical services basically consist of two aspects of library operations, acquisitions and cataloguing. Since the acquisition of books by Rome’s libraries has largely been dealt with earlier, this section will concentrate on the operation of cataloguing in the libraries of Rome. Information on Roman cataloguing practices is almost entirely lacking, but an examination of the evidence that does exist strongly suggests that the Romans followed models set up by the Greeks, particularly those established at the Library of Alexandria.662 These practices are reasonably well-known, so it becomes possible to examine them and extrapolate from them to Roman libraries. Harrod’s Librarian’s Glossary defines cataloguing as ‘the process of making entries for a catalogue, [which] additionally….may cover all the processes involved in preparing books for the shelves.’663 Essentially the process therefore involves two different aspects. The first is entering the books in the catalogue, the second is placing them on the shelves in such a manner that they can be easily found by consulting the catalogue. This second step has its own name, classification. Modern classification systems, such as Dewey and Library of Congress Classification, are sophisticated systems whereby books are arranged on the shelves by subject with a high degree of precision, usually by means of an alphanumeric code where the exact subject area can be defined to a very exacting degree. However, the basic principle of modern classification is that items of the same and similar

660 Oldenbourg (1999: 17) extols the Forum as the most prominent ‘third place’ in Roman society. Certainly it was the major meeting-place, but given that the major business of the Forum was political, and that the business of Roman men of the upper classes was largely politics, it appears to qualify more as Oldenbourg’s ‘second place’, in other words as a workplace, rather than as a comfortable, casual meeting place. 661 Gell. NA 13.20.1. 662 Hanson and Daly 2003: 433. 663 1971: 128. 117 subject areas be shelved together. To a large extent, with some variation, this also appears to have been the principle followed by the Greeks and by extension the Romans. The practice of cataloguing certainly predates both the Greeks and Romans. The first library catalogue known comes from the Assyrian in , dating from c. 700 BC. This library, discovered in the 1840s, consisted of a roomful of clay tablets, arranged according to subject matter, indicating that this basic principle is probably as old as the concept of libraries themselves. Among the tablets was one apparently containing a list of the subjects and a list of the tablets pertaining to each subject.664 The tablets themselves appear to have a number of features which, Pfeiffer argues, were adopted by the cataloguers at the Library of Alexandria, including titles at the end of the tablet and stichometric counts of the numbers of lines.665 It is certainly possible that the knowledge of these practices was brought to Alexandria following Alexander’s conquest of Mesopotamia, but it is also possible that they were developed quite independently by the Alexandrian cataloguers in response to their own needs. There is little doubt that the pre-eminent cataloguer of the Alexandria library was Callimachus. Callimachus worked at the Library of Alexandria during the 3rd Century BC. Although known to posterity primarily as a poet, he was also described as a grammarian. Some evidence suggests that he may have held the post of Director of the Library, but this is disputed.666 Whether he did or not, undoubtedly his greatest service to the Library was his creation of a catalogue to the library. Although he created a number of catalogues, this catalogue in particular is the one that came to be seen as a model for future catalogues that was still in use centuries later. This was his Pinakes, apparently titled ‘Tables of all those who were eminent in any kind of literature, and of their writings, in 120 books.’667 This provides an excellent description of the Pinakes, as it was in fact not just a simple catalogue of what was contained within the Alexandria library, but an inventory of existing works of Greek literature, with biographical details of the authors, as Too notes.668 It was in fact a reference work, not specifically attached to the library, but something that could be used elsewhere, even in other libraries. Basing his Pinakes on work already done to organise and list the collection by two of his predecessors, Alexander the Aetolian and Lychophron the Chalcidian, Callimachus created a sophisticated classification system based on subject and author.669 The complete list of divisions has

664 Severance 116. 665 1968: 126. 666 Blum 1991: 127 ff. 667 Suidas 3.19; cf. Parsons 1952: 208; Witty 1958: 133: Blum 1991: 150: Too 2010: 55 668 2010: 55 669 Parsons 1952: 207. 118 not survived, but it appears likely that there were 8 (or possibly 10) subject classes.670 Within these classes were further subdivisions and then the classes and subdivisions were arranged alphabetically by author. The entry for each author included biographical data, the titles of his works, the opening words of each work and the line-count for each work. As can be seen, it was a detailed index providing a lot of information about the author and his works. It did not, however, have any link to the actual copies on the shelves, which is why some scholars draw the line at calling it a catalogue, instead describing it as a bibliography, an independent reference work which, while based on the holdings of the library, was not tied to it.671 This sense of a catalogue as an independent reference work that could be used in any library was carried on by the Romans. There are a number of references in literature of the 1st Century BC through to the 2nd Century AD of similar works, called by the Romans ‘indices.’ Cicero refers to an Index Tragicorum, presumably a catalogue of dramatic tragedies, while Seneca mentions an Index Philosophorum.672 Similar, if unnamed,673 catalogues are mentioned by Quintilian and Pliny. However, the most significant reference to a catalogue of this kind comes from a recently discovered work of Galen, the Peri Alupias (On the Avoidance of Grief). In this work, Galen laments the loss of both his personal library and of the public libraries destroyed by the fire of 192 AD. Galen describes using the catalogues, but finds faults in them, either books ascribed to the wrong authors or books found in the libraries which are not in the catalogues.674 As Nicholls notes, this does not appear to be a specific catalogue created for this library, but an independent reference work that lists authors of a particular genre and their works.675 Hence, Galen gives a good indication of how such a work should be used, not to locate where a particular copy is on the shelf as in a modern catalogue, but as a guide to what books a library should hold.676 Circumstantial evidence would seem to indicate that such catalogues, if they were not copies of Callimachus’ Pinakes, were probably based on the structure set up by Callimachus. First, the tremendous importance and clout of the Alexandrian library, the premier library of the ancient world, and the enormous respect accorded to the Pinakes, would surely have seen it used or copied elsewhere. It is known that the Pinakes were still in use into the 3rd and 4th Century

670 The eight basic divisions are 1. Epic poetry; 2. Drama; 3. Law; 4. Philosophy; 5. History; 6. Oratory; 7. ; 8. Miscellaneous. Of these only Law, Oratory and Miscellanea are definitely known from fragments of the Pinakes. The two categories added by some scholars are Medicine and Science. Cf. Parsons 1952: 210 ff. 671 Blum (1991: 160) describes it as a catalogue of works, rather than a catalogue of copies. 672 Cic. Hort. fr, 9; Sen., Ep. 39. 673 Quint. Inst. Or. 10.1.57; Plin. Ep. 3.5. 674 16-17. 675 2011: 135. 676 Although some commentary identifies it as an Alexandrian catalogue, Nicholls (2011: 135) argues that it is in fact the catalogue of Aristotle’s works compiled by Tyrannio and Andronicus of Rhodes. 119

AD. Clearly it was a highly-regarded work.677 Second, there is the respect in which Callimachus himself was held by the Romans. Cicero puts him in the pantheon of the most illustrious Greeks, and other authors are similarly laudatory.678 However, as useful as such independent catalogues would be, they must have been supplemented by a local catalogue of some description, a shelf list that showed where works were located on the shelves. Most scholars agree that a Greek (and by extension a Roman) library needed in fact two catalogues, the independent reference work already discussed and a shelf list. Evidence of such shelf lists is limited, but there appears to be some archaeological evidence which may show how these shelf-list catalogues functioned. It appears that they may have taken the form of tablets or inscriptions attached to the wall of the library that gave directions on where particular genres of books could be found. One example has been found on the island of Rhodes. It dates to the late second or early first century BC.679 It records works by 5 different authors in the left-hand column and then, in the right-hand column, the number of books which make up each title. The authors are arranged alphabetically and it appears to deal with similar subjects. It has been characterised as a list of donations, but Nicholls argues that it was in fact a library catalogue.680 It is not difficult to see that it could have been affixed to the wall near a particular collection of books to show what titles were contained within the collection. There is a similar example of what might have been a Roman counterpart to this inscription. Three fragments of an inscription were located in Rome which appear to list writers and their plays in the order in which they were placed in a dramatic festival in Athens.681 Although somewhat different in emphasis from the Rhodes inscription, it does appear to follow the same principle of listing the authors and then their works. Although not found in connection with any known library, it is not difficult to conceive of it having been attached to the wall of a library and providing an indication of what works the library held, as both Pickard-Cambridge and Pfeiffer argue.682 However, both it and the Rhodes inscription clearly indicate the flaw with which this type of catalogue is inherently encumbered. Lacking a proper classification system, it cannot point exactly to where a book may be found on the shelves. As Nicholls argues, these catalogues could all lead a reader to approximately the

677 Blum 1991: 125. 678 Cic. De or. 3.132; Gell. NA 17.21. 679 Platthy 1968: 148 ff. 680 1991: 186. 681 IG 1097, 1098. 682 1968: 121 ff; 1968: 132. 120 right spot, but not the exact spot on the shelves where a book was located.683 Knowledge of exactly how the Romans organised their books upon the shelves is lacking. It has been generally assumed that some sort of subject order was used, though evidence once again from the Peri Alupias suggests that other forms of order were also used. Galen makes reference to particular collections of books which are organised by the names of what are believed to be the collectors or compilers of these collections.684 It appears that these books were either donated or given as a bequest, or otherwise acquired as a block, and kept together within the library, some of them for between 200-250 years by the time Galen saw them. It can be seen therefore that many of the practices which were undertaken in Roman libraries parallel those of modern libraries. There was no equivalent to the Roman public library system in the West until the 19th Century, and although many of the practices developed by modern libraries evolved independently, they followed roughly similar courses to what the Romans had developed almost 2000 years earlier. Some practices which the Romans undertook through necessity, such as the use of natural lighting, are only now being appreciated as valuable undertakings in modern libraries, not because of necessity, but because they add to the library experience. It is fitting to end this study here, with a look at the parallels and differences between the Roman libraries and modern libraries. In a sense, therefore, the journey which we began in the late stages of the Roman Republic will end in here in the early years of the third millennium.

683 2011: 137. 684 13. The collections are the Callinia, Atticiana, Pedoucinia and Aristarcheia. Jones (2009: 393) argues that the Atticiana was indeed the collection of Cicero’s friend Titus Pomponius Atticus. As for the other purported collections, the Aristarcheia has been linked to the Alexandrian grammarian of the 2nd Century BC, whereas the others are a matter of some debate. Cf. Jones 2009: 391 ff; Nicholls 2011: 130 ff. 121

CONCLUSION

From the late first century BC to the 5th Century AD, the city of Rome had the most advanced library system seen in the world until the 19th Century. A system of public libraries financed by the Imperial household and run by Imperial bureaucrats provided Rome’s writers and intellectuals with a substantial resource of books that was only equalled in the ancient world by the great library of Alexandria. This library system was built on a foundation created by the aristocrats of the late Republic with their own significant and sophisticated private libraries in their large country villas. This thesis has endeavoured to show how these libraries came into being, how they developed and became more sophisticated over time, and how the Romans built on the basic library model provided by the Greeks and adapted that model for their own needs. The first chapter examined the beginnings of Roman libraries, suggesting that libraries were in existence in Rome before the commonly accepted date for the inception of libraries of 168 BC, including the possibility that proto-libraries of religious texts were put together by Rome’s priesthoods, and that early in Rome’s literary history, playwrights had access to libraries of Greek originals. It was then shown that the private libraries of Rome’s aristocrats, first the looted libraries of the conquerors of the East, then those of the great bibliophiles of the Republic’s last years, steadily developed in size and sophistication. These eventually led to the first public libraries, first the library posited by Caesar, and then the libraries brought into being by Pollio and Augustus. The second chapter outlined arguments by scholars that the public libraries were of little significance, and provided counter- arguments to demonstrate that they were considered valuable by Rome’s writers, and also provided evidence that the usage of the libraries might have been wider in social terms than previously thought. The third chapter examined the development of private libraries through the 1st and early 2nd Centuries, demonstrating their continuing development, from the very basic Greek models found in Herculaneum and Pompeii, through the private libraries of the Imperial households, showing how they continued to develop in sophistication, following the innovations provided in the public libraries, with some breaking away from the conventional rectilinear shape towards more innovative designs. The fourth chapter examined the library of one particular writer of the 1st Century, Pliny the Elder, and how he made use of his collection and the written resources available to him to write the Natural History, using a relatively sophisticated information retrieval system. The last chapter was a survey of Rome’s libraries utilising modern library survey methodology, and demonstrating, by comparison with modern library practice, how these ancient libraries coped with the issues common to libraries of all eras.

122

This last chapter indicates the potential for further research in this area. There is scope for a more detailed comparison between ancient and modern library practice. Such areas as library use patterns, the relationship between a library, its local area and its clientele, efficiency of library practice and effective use of resources are all current foci of study in modern libraries. Although the ancient evidence is naturally limited, there are grounds for examination and comparison with the modern picture, including the use of statistical models used in modern library surveys. There is a basis for library practice which holds true for libraries from all eras, and there is certainly value in examining this from its earliest appearances in human history.

123

ANCIENT SOURCES

Anthologia Palatina (trans. W.R. Paton). 1916-18. Anthologia Palatina, London: Heinemann. Aelius Spartianus (trans. D. Magie). 1921-32. Hadrian, Cambridge, Mass.: Harvard University Press. Ammianus (trans. J.C. Rolfe). 1950-52. Rerum Gestarum, London: Marcellinus Heinemann. Appian (trans. H. White). 1912-13. Bella Civilia, London: Heinemann. (trans. H. White). 1912-13. Mithridatic Wars, London: Heinemann. (trans. J.A. Hanson). 1989. , Cambridge, Mass.: Harvard University Press. Arnobius (trans. G.E. McCracken). 1949. Adversus Nationes, Westminster: Newman Press. Athenaeus (trans. C.B. Gulick). 1927-41. Deipnosophistae, London: Heinemann. Augustus (trans. A.E. Cooley). 2009. Res Gestae Divi Augusti, Cambridge: Cambridge University Press. Augustine (trans. D.B. Zema…[et al.]). 1950-54. De Civitate Dei, Washington: Catholic University Press of America. (trans. S.J.B. Barnish). 1992. Variae, Liverpool: Liverpool University Press. Cassius Dio (trans. E. Cary). 1914-1927. Historia Romana, London: Heinemann. Cato (trans. W.D. Hooper). 1934. De Agricultura, London: Heinemann. Catullus (trans. G.P. Goold). 1988. Carmina, London: Heinemann (2nd rev. ed.). Cicero (trans. H. Rackham). 1933. Academica, London: Heinemann. (trans. G.L. Hendrickson). 1962. Brutus, London: Heinemann. (trans. W.A. Falconer). 1983. Cato Maior de Senectute, London: Heinemann. (trans. W.A. Falconer). 1923. De Divinatione, London: Heinemann. (trans. N.H. Watts). 1923. De Domo Sua, London: Heinemann. (trans. H. Rackham). 1931. De Finibus Bonorum et Malorum, Cambridge, Mass.: Harvard University Press. (trans. J.H. Freese). 1930. De Lege Agraria, London: Heinemann. (trans. E.W. Sutton and H. Rackham). 1948. De Oratore, Cambridge, Mass.: Harvard University Press.

124

(ed. H.E. Butler and M. Cary). 1924. De Provinciis Consularibus, Oxford: Clarendon Press. (trans. C.W. Keyes). 1928. De Republica, London: Heinemann. (trans D.R. Shackleton Bailey). 1912-18. Epistulae ad Atticum, London: Heinemann. (trans. W.G. Williams). 1927-29. Epistulae ad Familiares, London: Heinemann. (trans. D.R. Shackleton Bailey). 2002. Epistulae ad Quintum Fratrem, Cambridge, Mass.: Harvard University Press. (trans. N.H. Watts). 1953. In L. Calpurnium Pisonem, London: Heinemann. (trans. H. Rackham). 1948. Paradoxa Stoicorum, London: Heinemann. (trans. W.C.A. Ker). 1926. Philippics, London: Heinemann. (trans. N.H. Watts). 1923. Pro Archia Poeta, London: Heinemann. (trans. R. Gardner). 1958. Pro Sestio, London: Heinemann. (trans. H.M. Hubbell). 1949. Topica, London: Heinemann. (trans. J.E. King). 1945. Tusculanae Disputationes, London: Heinemann. Columella (trans. H.B. Ash). 1941-55. De Re Rustica, Cambridge, Mass.: Harvard University Press. Diogenes Laertius (trans. R.D. Hicks). 1925. Vitae Philosophorum, London: Heinemann. Dionysius of (trans. E. Cary). 1937-50. Antiquitates Romanorum, London: Halicarnassus Heinemann. Donatus, Aelius (trans. D.S. Wilson-Okamura). 1996. Virgil, www.virgil.org/vitae/a-donatus.htm (accessed 16 February 2012). Florus (trans. E.S. Forster). 1929. Epitome de T. Livio Bellorum Omnium Annorum DCC Libri Duo, London: Heinemann. (trans. M.B. McElwain). 1925. Stratagemata, London: Heinemann. Fronto (trans. C.R. Haines). 1929. Epistulae, London: Heinemann. Galen (trans. C.K. Rothschild and T.W. Thompson). 2011. Peri Alupias, in Early Christianity 2: 110-129. Gellius, Aulus (trans. J.C. Rolfe). 1946. Noctes Atticae, London: Heinemann. Gregory (trans. J.R.C. Martin). 2004. Epistulae, Toronto: Pontifical Institute of Medieval Studies. Homer (trans. A.T. Murray). 1924-25. Iliad, London: Heinemann.

125

(trans. A.T. Murray). 1995. Odyssey, Cambridge, Mass.: Harvard University Press. Horace (trans. E. De Faur). 1906. Carmen Saeculare, Sydney: Brooks Press. (trans. H.R. Fairclough). 1929. Epistulae, Cambridge, Mass.: Harvard University Press. Isidorus (trans. S.A. Barney). 2006. , Cambridge: Cambridge University Press. (trans. M.D. Donalson). 1996. Chronicon, Lewiston: Mellen University Press. (trans. H. St.J. Thackeray). 1926-65. Antiquitates Judaicae, London: Heinemann. Josephus (trans. H. St.J. Thackeray). 1926-1965. Bellum Judaicum, London: Heinemann. Juvenal (trans. S.M. Braund). 2004. Satura, Cambridge, Mass.: Harvard University Press. (trans. M.F. McDonald). 1964. Divinarum Institutionum Libri VII, Washington: Catholic University Press of America. Livy (trans. by B.O. Foster …[et. al.]). 1919-. , London: Heinemann. (trans. by A.M. Harmon). 1913-67. Adversus Indoctum, London: Heinemann. Macrobius (trans. P.V. Davies.). 1969. , New York: Columbia University Press. Martial (trans. W.C.A. Ker). 1968. Epigrammata, London: Heinemann. Nepos, Cornelius (trans. N. Horsfall). 1989. Atticus, Oxford: Clarendon Press. (trans. N. Horsfall). 1989. Cato, Oxford: Clarendon Press. (ed. E.S. Shuckburgh). 1894. Cimon, Cambridge: Cambridge University Press. (trans. R.J. Deferrari). 1964. Historiae Adversus Paganos, Washington: Catholic University Press of America. Ovid (trans. G. Showerman). 1921. Amores, London: Heinemann. (trans. A.L. Wheeler). 1924. Epistulae ex Ponto, London: Heinemann. (trans. J.G. Frazer). 1989. Fasti, Cambridge, Mass.: Harvard University Press (2nd rev. ed.). (trans. A.L. Wheeler). 1924. Tristia, London: Heinemann. Paulus, Julius (trans. S.P. Scott). 1973. Sententiae, New York: AMS Press. Plautus (trans. P. Nixon). 1965. Casina, London: Heinemann.

126

Pliny (trans. H. Rackham, W.H.S Jones and D.E. Eicholz). 1963. Historia Naturalis, London: Heinemann. Pliny (trans. B. Radice), 1969. Epistulae, London: Heinemann. (trans. B. Radice). 1969. Panegyricus, London: Heinemann. Plutarch (trans. B. Perrin). 1914-26. Aemilius Paullus, London: Heinemann. (trans. B. Perrin). 1914-26. Cato Maior, London: Heinemann. (trans. B. Perrin). 1914-26. Cicero, London: Heinemann. (trans. B. Perrin). 1914-26. Crassus, London: Heinemann. (trans. B. Perrin). 1914-26. Lucullus, London: Heinemann. (trans. B. Perrin). 1914-26. Marcellus, London: Heinemann. (trans. B. Perrin). 1014-26. Pompey, London: Heinemann. (trans. B. Perrin). 1914-26. Sulla, London: Heinemann. Polybius (trans. W.R. Paton). 1927. Historiae, Cambridge, Mass.: Harvard University Press. Propertius (trans. H.E. Butler). 1912. Carmina, London: Heinemann. Quintilian (trans. H.E. Butler). 1921-22. Institutio Oratoria, London: Heinemann. Seneca (trans. M. Winterbottom). 1974. Controversiae, Cambridge, Mass.: Harvard University Press. Seneca (trans. R.M. Gummere). 1917-25. Ad Lucilium Epistulae Morales, London: Heinemann. (trans. A.N. Athanassakis). 1973. Apocolocyntosis Divi Claudii, Lawrence: Coronado Press. (trans. J.W. Basore). 1928-35. De Tranquillitate Animi, London: Heinemann. Sidonius (trans. W.B. Anderson). 1936-65. Carmina, London: Heinemann. (trans. W.B. Anderson). 1936-65. Epistulae, London: Heinemann. (trans. D.R. Shackleton Bailey). 2003. Thebaid, Cambridge, Mass.: Harvard University Press. Strabo (trans. H.L. Jones). 1944-49. Geographica, London: Heinemann. Suetonius (trans. J.C. Rolfe). 1998. Augustus, Cambridge, Mass.: Harvard University Press. (trans. J.C. Rolfe). 1998. Caligula, Cambridge, Mass.: Harvard University Press. (trans. J.C. Rolfe). 1998. Claudius, Cambridge, Mass.: Harvard University Press.

127

(trans. J.C. Rolfe). 1998. De Illustribus Grammaticis, Cambridge, Mass.: Harvard University Press. (trans. J.C. Rolfe). 1998. Divus Julius, Cambridge, Mass: Harvard University Press. (trans. J.C. Rolfe). 1998. Domitian, Cambridge, Mass.: Harvard University Press. (trans. J.C. Rolfe). 1998. Tiberius, Cambridge, Mass.: Harvard University Press. (trans. J.C. Rolfe). 1998. Aulus Persius Flaccus, Cambridge, Mass.: Harvard University Press. Tacitus (trans. J. Jackson). 1925-37. , London: Heinemann. Terence (trans. J. Barsby). 2001. Hecyra, Cambridge, Mass.: Harvard University Press. (trans. by R. Arbesmann…[et. al]). 1950. Apologeticus, Washington: Catholic University Press of America. Theophrastus (trans. A. Hort). 1916. Historia Plantarum, London: Heinemann. Ulpian (trans. S.P. Scott). 1973. Digest, New York: AMS Press. (trans. D.R. Shackleton Bailey). 2000. Factorum ac Dictorum Memorabilium Libri IX, Cambridge, Mass.: Harvard University Press. Varro (trans. R.G. Kent). 1938. De Lingua Latina, London: Heinemann. (trans. W.D. Hooper). 1934. Rerum Rusticarum Libri III, London: Heinemann. (trans. F.W. Shipley). 1924. Historiae Romanae, London: Heinemann. Virgil (trans. H.R. Fairclough). 1934. Aeneid, London: Heinemann. Vitruvius (trans. I.D. Rowland). 1999. , Cambridge: Cambridge University Press. Vopiscus (trans. D. Magie). 1921-32. Aurelian, Cambridge, Mass.: Harvard University Press. (trans. D. Magie). 1921-32. Probus, Cambridge, Mass.: Harvard University Press.

128

BIBLIOGRAPHY

Aabo, S., Audunson, R. and Varheim, A. 2010. ‘How do public libraries function as meeting places?’, Library and Information Science Research 32: 16-26. Adembri, B. 2000. Hadrian’s Villa, Rome: Electa. Allison, P.M. 2006. The Insula of the Menander at Pompeii: Volume 3 – The Finds: A Contextual Study, Oxford: Clarendon Press. Amici, C.M. 1982. Foro di Traiano: Basilica Ulpia e Biblioteche, Rome: L’Erma di Bretschneider. Anderson, J.C. 1984. The Historical Topography of the , Collection Latomus 182, Brussels: Latomus. Andreae, B. 1977. The Art of Rome, New York: Abrams. Ashby, T. 1906, ‘The Classical topography of the Roman Campagna: Part 2’, Papers of the British School at Rome 3. _____ 1907, ‘The Classical topography of the Roman Campagna: Part 3 – Section 1’, Papers of the British School at Rome 4: 121-124. _____ 1910 ‘The Classical topography of the Roman Campagna: Part 3 – Section 3’, Papers of the British School at Rome 5: 215-425. Astin, A.E. 1967. Scipio Aemilianus, Oxford: Clarendon Press. _____ 1978. Cato the Censor, Oxford: Clarendon Press. Ball, L.F. 2003. The Domus Aurea and the Roman Architectural Revolution, Cambridge: Cambridge University Press. Balsdon, J.P.D.V. 1960, ‘, dignitas, otium’, Classical Quarterly 10: 43-50. Barnes, R. 2000. ‘Cloister bookworms in the chicken-coop of the Muses: the ancient library of Alexandria’, in R. MacLeod ed., The Library of Alexandria: Centre of Learning in the Ancient World, London: I.B. Tauris, 61-78. Barton, I.M. 1996. Palaces, Exeter: University of Exeter Press. Barwick, K. 1933, ‘Das Kultlied des Livius Andronicus’, Philologus 88: 203-221. Beagon, M. 1992. Roman Nature: The Thought of Pliny the Elder, Oxford: Clarendon Press. Beare, W. 1964: The Roman Stage: A Short History of Latin Drama in the Time of the Republic, London: Methuen. Best, E.E. 1965. ‘Attitudes toward literacy reflected in Petronius’, Classical Journal 61: 72-76. Bishop, J.H. 1956. ‘Palatine Apollo’, Classical Quarterly 6: 187-192. 129

Blake, M.E. 1947. Ancient Roman Construction in Italy from the Prehistoric Period to Augustus, Washington: Carnegie Institution. _____ 1959. Roman Construction in Italy from Tiberius through the Flavians, Washington: Carnegie Institution. _____ 1973. Roman Construction in Italy from Nerva through the Antonines, Philadephia: American Philosophical Society. Bloch, H. 1940. ‘L. Calpurnius Piso in Samothrace and Herculaneum’, American Journal of Archaeology 44: 485-493. Blum, R. 1991. Kallimachos: The Alexandrian Library and the Origins of Bibliography, Madison, WI: University of Wisconsin Press. Boethius, A. 1960. The Golden House of Nero: Some Aspects of Roman Architecture, Ann Arbor: University of Michigan Press. Boll, J.J. 1980. To Grow or not to Grow?, New York: Bowker. Bowersock, G.W. 1990. ‘The dawn of reading’, The New Republic, April 2: 37-39. Boyd, C.E. 1915. Public Libraries and Literary Culture in , Chicago: University of Chicago Press. Bruce, L. 1983. ‘The Procurator Bibliothecarum at Rome’, Journal of 18: 143-162. _____ 1986. ‘Palace and villa libraries from Augustus to Hadrian’, Journal of Library History 21: 510- 522. Burnett, A. 1999. ‘Buildings and monuments on Roman coins’, in G.M. Paul, ed., Roman Coins and Public Life under the Empire: E. Togo Salmon Papers, Ann Arbor: University of Michigan Press, 137-164. Burrell, B. 2009. ‘Reading, hearing and looking at Ephesos’, in W.A. Johnson and H.N. Parker, eds., Ancient Literacies: The Culture of Reading in Greece and Rome, Oxford: Oxford University Press, 69-95. Byrne, A.H. 1920. Titus Pomponius Atticus: Chapters of a Biography, PhD Thesis, Bryn Mawr (UMI). Cagnat, R. 1906. L s biblio s ni i al s dans l ir ro ain, Paris: Imprimerie nationale. Caley, E.R and Richards, J.F.C. 1956. Theophrastus on Stones, Columbus: Ohio State University Press. Callmer, C. 1944. ‘Antike bibliotheken’, Opuscula Archaeologica 3: 145-193. Calvino, I. 1986. ‘Man, the sky and the elephant’, in I. Calvino ed., The Uses of Literature: Essays, San Diego: Harcourt Brace Jovanovich.

130

Cameron, C. 1772. The Baths of the Romans, London: G. Scott. Canfora, L. 1989. The Vanished Library, Berkeley: University of California Press Carcopino, J. 1951. Cicero: The Secrets of his Correspondence, London: Routledge. Carettoni, G. 1969. ‘The House of Augustus – 1’, Illustrated London News 255: 24-25. Casson, L. 2001. Libraries in the Ancient World, New Haven: Yale University Press. Chroust, A. 1973, Aristotle: New Light on his Life and Some of his Lost Works – Vol. 1: Some Novel Interpretations of the Man and his Life, London: Routledge. Churchill, J.B. 1999, ‘Ex qua quod vellent facerent: Roman magistrates’ authority over praeda and manubiae’, Transactions and Proceedings of the American Philological Association, 129: 85- 116. Clark, J.W. 1909. The Care of Books: An Essay on the Development of Libraries and their Fittings, from the Earliest Times to the End of the Eighteenth Century, Cambridge: Cambridge University Press. Clift, E.H. 1945, Latin Pseudepigraphia: A Study in Literary Attributions, Baltimore: Johns Hopkins University Press. Coarelli, F. 2000. The Column of Trajan, Rome: Editore Colombo. _____ 2007. Rome and Environs: An Archaeological Guide, Berkeley: University of California Press. Conte, G.B. 1994. G nr s and R ad rs: L r i s, Lov l gy, Pliny’s n y lo a dia, Baltimore: Johns Hopkins University Press. Coulson, W.D.E. 1976. ‘The reliability of Pliny’s chapters on Greek and Roman sculpture’, Classical World 69: 361-372. Dahlkild, N. 2011. ‘The emergence and challenge of the modern library building: ideal types, model libraries, and guidelines, from the Enlightenment to the Experience Economy’, Library Trends 60: 11-42. Dalzell, A. 1955. ‘C. Asinius Pollio and the early history of public recitation at Rome’, Hermathena 86: 20-28. Thompson, D.L. 1981. ‘The meetings of the on the Palatine’, American Journal of Archaeology 85: 335-339. D’Arms, J.H. 1970, Romans on the Bay of Naples: A Social and Cultural Study of the Villas and their Owners from 150 BC to AD 400, Cambridge: Harvard University Press. De Lacey, P. 1993. ‘Review of: M. Capasso, Manuale di Papirologia ercolanese’, American Journal of Philology 114: 178-180.

131

DeLaine, J. 1997. The Baths of Caracalla: A Study in the Design, Construction and Economics of Large-Scale Building Projects in Imperial Rome, Portsmouth, RI: Journal of Roman Archaeology. Della Corte, M. 1940. ‘Virgilio nell’ epigrafia pompeiana’, Epigraphica 2: 171-178. Dix, T.K, 1986, Private and Public Libraries at Rome in the 1st Century BC: A Preliminary Study in the History of Roman Libraries, PhD thesis, Ann Arbor: University of Michigan (UMI). _____ 1994. ‘Public libraries in ancient Rome: ideology and reality’, Libraries and Culture 29: 282- 296. _____ 1996. ‘Pliny’s library at Comum’, Libraries and Culture 31: 85-102. _____ 2000. ‘The library of Lucullus’, A na : s di di l ra ra a s oria d ll’an i i a 88: 441- 464. _____ 2004, “Aristotle’s “Peripatetic” library’, in J. Raven, ed., Lost Libraries: The Destruction of Great Book Collections since Antiquity, London: Palgrave Macmillan. Dix, T.K. and Houston, G.W. 2006. ‘Public libraries in the city of Rome: from the Augustan age to the time of Diocletian’, Melanges de l'ecole Francaise de Rome – Antiquite 118: 671-717. Doody, A. 2001. ‘Finding facts in Pliny’s encyclopedia: the summarium of the Natural History’, Ramus 30: 1-22. _____ 2010. Pliny’s n y lo dia: T R ion of Na ral His ory, Cambridge: Cambridge University Press Duckworth, G.E. 1952, The Nature of Roman Comedy: A Study in Popular Entertainment, Princeton: Princeton University Press. Duff, J.W. 1928. Freedmen in the Early Roman Empire, Oxford: Clarendon Press. Dunlap, L.W. 1972, Readings in Library History, New York: Bowker. Dupont, F. 2009. ‘The corrupted boy and the crowned poet: or, the material reality and the symbolic status of the literary book at Rome’, in W.A. Johnson and H.N. Parker, eds., Ancient Literacies: The Culture of Reading in Greece and Rome, Oxford: Oxford University Press, 143-163. During, I. 1956. ‘Ariston or Hermippus? A note on the catalogue of Aristotle’s writings’, Classica et Medievalia, 17: 11-21. Evans, H. 1983. P bli a Car ina: Ovid’s Books fro xil , Lincoln: University of Nebraska Press. Farrell, J. 2004. ‘Roman Homer’, in R. Fowler, ed., The Cambridge Companion to Homer, Cambridge: Cambridge University Press, 254-271.

132

Forsythe, G. 1991, ‘A philological note on the Scipionic circle’, American Journal of Philology 112: 363-364. Frank, T. 1928. Catullus and Horace: Two Poets in their Environment, Oxford: Blackwell. _____ 1929. Roman Imperialism, New York: Macmillan. Franklin, J.L. 1996. ‘Vergil at Pompeii: a teacher’s aide’, Classical Journal 92: 175-184. Fraser, P.M. 1972. Ptolemaic Alexandria, Oxford: Clarendon Press. French, R. 1994. Ancient Natural History: Histories of Nature, London: Routledge. Goodyear, F.R.D. 1982. ‘Technical writing’, in E.J. Kenney, ed., Cambridge History of Classical Literature: Vol. 2 – Latin Literature, Cambridge: Cambridge University Press. Goold, G.P. 1970. ‘Servius and the Helen episode’, Harvard Studies in Classical Philology 74: 101- 168. Gothberg, H.M. 1990. ‘The library survey: a research methodology rediscovered’, College and Research Libraries 6: 553-559. Gotze, B. 1937. Antike Bibliotheken, Berlin: De Gruyter. Goulding, A. 2009. ‘Engaging with community engagement: public libraries and citizen involvement’, New Library World 110: 37-51. Granger, F. 1933. ‘Julius Africanus and the Library of the Pantheon’, Journal of Theological Studies 34: 157-161. Grayeff, F. 1974. Aristotle and his School: An Inquiry into the History of the Peripatetos, London: Duckworth. Guzman, P. 1904. La Villa Imperiale de Tibur, Paris: Albert Fontemoing. Hall, F.W. 1922. ‘Review’, Classical Review 36: 31-32. Hankinson, R.J. 1988. ‘Review of: R. French and F. Greenaway, eds., Science in the Early Roman Empire – Pliny the Elder: His Sources and Influence’, Isis 79: 340-341. Hanson, E.R. and Daily, J.E. 2003. ‘Catalogs and cataloging’, in M.A. Drake, ed., Encyclopedia of Library and Information Science, 2nd edn., New York: Marcel Dekker. Harris, M.H. 1995, in the Western World, 4th edn., Lanham: The Scarecrow Press Harris, W.V. 1989. Ancient Literacy, Cambridge: Harvard University Press. Healy, J.F. 1999. Pliny the Elder on Science and Technology, Oxford: Oxford University Press. Hill, P.V. 1989. The Monuments of Ancient Rome as Coin Types, London: Seaby.

133

Hillscher, A. 1892. ‘Hominum Litteratorum Graecorum ante Tiberii Mortem in Urbe Commemoratorum Historia Critica’, Jahrbucher für Classiche Philologie, Supplement-band 18: 354-444. Hirschfeld, O. 1963. Die Kaiserlichen verwaltungsbeamten bis auf Diocletian, Berlin: Weidmann. Holford-Strevens, L. 1988. , London: Duckworth. Horsfall, N. 1972. ‘Varro and Caesar: three chronological problems’, Bulletin of the Institute of Classical Studies 19: 120-125. _____ 1976, ‘The Collegium Poetarum’, Bulletin of the Institute of Classical Studies 23: 79-95. _____ 1987. ‘Review: The Text of Virgil in Antiquity’, Classical Review 37: 177-180. _____ 1993. ‘Empty shelves on the Palatine’, Greece and Rome 40: 58-67. _____ 1995. ‘Rome without spectacles’, Greece and Rome 42: 49-56. Houston, G.W. 2002. ‘The slave and freedman personnel of public libraries in ancient Rome’, Transactions and Proceedings of the American Philological Association 132: 139-176. _____ 2004. ‘How did you get hold of a book in a Roman library? Three second-century scenarios’, Classical Bulletin 80: 5-13. Howe, N.P. 1985. ‘In defense of the encyclopedic mode: on Pliny’s preface to the Natural History’, Latomus 44: 561-576. Huelsen, C. 1928. The Forum and the Palatine, New York: A. Bruderhausen. Johnson, D.R. 1980. ‘The Library of Celsus: an Ephesian phoenix’, Wilson Library Bulletin 54: 651- 653. Johnson, E.D. 1970. History of Libraries in the Western World, Metuchen: Scarecrow. Johnson, W.A. 2000. ‘Towards a sociology of reading in ’, Latomus 47: 593-627. _____ 2006. ‘The story of the Villa de Papyri’, Journal of Roman Archaeology 19: 483-496. Jolowicz, H.F, and Nicholas, B. 1972, Historical Introduction to the Study of , Cambridge: Cambridge University Press. Jones, A.H.M. 1950. ‘The aerarium and the fiscus’, Journal of Roman Studies 40: 22-29. Jones, C. 2009. ‘Books and libraries in a newly-discovered treatise of Galen’, Journal of Roman Archaeology 22: 390-397. Jory, E.J. 1970, ‘Associations of actors in Rome’, Hermes 98: 224-253. Joshel, S.R. 2010. Slavery in the Roman World, Cambridge: Cambridge University Press. Kaehler, H. 1950. Hadrian und seine Villa bei Tivoli, Berlin: Gebr. Mann. Kahler, H. 1965. The Art of Rome and her Empire, New York: Crown Press.

134

Keaney, J.J. 1963. ‘Two notes on the tradition of Aristotle’s writings’, American Journal of Philology 84: 52-63. _____ 1968. ‘The early tradition of Theophrastus’ Historia Plantarum’, Hermes 96: 293-298. Kenney, E.J. 1982. ‘Books and readers in the Roman world’, in E.J. Kenny, ed., Cambridge History of Classical Literature: Vol. 2 – Latin Literature, Cambridge: Cambridge University Press. Kenyon, F.G. 1951. Books and Readers in and Rome, 2nd edn., Oxford: Clarendon Press. Kleberg, T. 1973. ‘Book auctions in ancient Rome?’, Libri 23: 1-5. Lanciani, R.A. 1889. Ancient Rome in the Light of Recent Discoveries, London: Macmillan. _____ 1897. The Ruins and Excavations of Ancient Rome: A Companion Book for Students and Travellers, London: Macmillan. Last, H. 1944. ‘The fiscus: a note’, Journal of Roman Studies 34: 51-59. Lerner, F. 2009. The Story of Libraries: From the Invention of Writing until the Computer Age, 2nd edn., New York: Continuum Lindsay, H. 1997. “Strabo on Apellicon’s Library”, Rheinisches Museum 140: 290-298 Line, B. 1982. Library Surveys: An Introduction to the Use, Planning, Procedure and Presentation of Surveys, London: Clive Bingley. Lloyd, G.E.R. 1983. Science, Folkore and Ideology: Studies in the Life Sciences in Ancient Greece, Cambridge: Cambridge University Press. Locher, A. 1986. ‘The structure of Pliny the Elder’s Natural History’, in R. French and F. Greenaway, eds., Science in the Early Roman Empire: Pliny the Elder, his Sources and Influence, London: Croom Helm. Maiuri, A. 1932. La casa de Menandro e il svo Tesoro di argentaria, Rome: Libreria della Stato. _____ 1958. Capri: Its History and its Monuments, Rome: Libreria della Stato. _____ 1974. Herculaneum and the Villa of the Papyri, Novara: Istituto Geografico de Agustini. Makowiecka, E. 1978. The Origin and Evolution of the Architectural Form of the Roman Library, Warsaw: Wydawnictwa Uniwersytetu Warszawskiego. Marrou, H.I. 1932. ‘La vie intellectual au forum de Trajan et au forum d’Auguste’, Melanges de l’É ol Fran ais d Ro – Antiquite 49: 93-110. Marshall, A.J. 1976. ‘Library resources and creative writing at Rome’, Phoenix 30: 252-264. Marshall, D.N. 1983. History of Libraries: Ancient and Medieval, New Delhi: Oxford & IBH Publishing

135

Marshall, E.A. 1986. A Biography of Titus Pomponius Atticus, PhD Thesis, Cambridge, Mass: Harvard University (UMI). Mattingly, H. 1919. The Imperial Civil Service of Rome, Cambridge: Cambridge University Press. Mattingly, H.B. 1957. ‘The date of Livius Andronicus’, Classical Quarterly 7: 159-163. May, F. and Black, F. 2010. ‘The life of the space: evidence from Nova Scotia public libraries’, Evidence-Based Library and Information Practice 5: 5-34. McCracken, G 1942. ‘The villa and tomb of Lucullus at Tusculum’, American Journal of Archaeology 46: 325-340. McDiarmid, E.W. 1940. The Library Survey: Problems and Methods, Chicago: American Library Association. McDonald, W.L. 1982. The Architecture of the Roman Empire: 1 – An Introductory Study, New Haven: Yale University Press. McKay, A.G. 1975. Houses, Villas and Palaces in the Roman World, Ithaca, NY: Cornell University Press. Mendell, C.W. 1951. Tacitus: The Man and his Work, New Haven: Yale University Press. Meneghini, R. 1993. ‘Nuovi dati sulle biblioteche e il Templum divi Traiani nel Foro di Traiano’, Bollettino di Archeologia 19-21: 13-21. _____ 2001. ‘Il Foro di Traiano: riconstruzione archittonica e analisi strutturale’, Mitteilungen des Deutschen Archaiologische Institut Romische Abteilung 108: 245-273. Merivale, C. 1890. History of the Romans under the Empire, London: Longmans, Green and Co. Millar, F. 1963. ‘The fiscus in the first two centuries’, Journal of Roman Studies 53: 29-42. _____ 1910. The Emperor in the Roman World, London: Duckworth. Milnor, K. 2009. ‘Literary literacy in Roman Pompeii: the case of Virgil’s Aeneid’, in W.A. Johnson and H.N. Parker, eds., Ancient Literacies: The Culture of Reading in Greece and Rome, Oxford: Oxford University Press, 288-319. Murphy, T. 2004. Pliny ld r’s Na ral His ory: T ir in n y lo a dia, Oxford: Oxford University Press. Nash, E. 1961-62. Pictorial Dictionary of Ancient Rome, London: Zwemmer. Nettleship, H. 1898. ‘The ancient commentators on Virgil’, in J. Conington, ed., The Works of Virgil – Vol. 1, London: George Bell and Sons, liv-c. Nibby, A. 1839. Ro a n ll’anno Mill o o-cento-trentotto, Rome: Tipografia delle Belle Arti.

136

Nicholls, M.C. 2011. ‘Galen and libraries in the Peri Alupias’, Journal of Roman Studies 101: 123- 142. Oakley, S.P. 1997, A Commentary on Livy, Vol. 1, Books VI-X, Oxford: Oxford University Press. Ogilvie, R.M. 1980. Roman Literature and Society, Brighton: Harvester Press. Oldenbourg, R. 1999. The Great Good Place; Cafés, Shops, Bookstores, Bars, Hair Salons and Other Hangouts at the Heart of a Community, Cambridge, MA: Da Capo Press. Oldfather, W.A. 1937. ‘Earliest recorded library regulation’, Library Quarterly 7: 437. Olinder, B. 1974. Porticus Octavia in Circo Flaminia: Topographical Studies of the Campus Region of Rome, Stockholm: Svenska Institutet I Rom. Oliver, J.H. 1979. ‘Flavius Pantaenus, Priest of the Philosophical Muses’, Harvard Theological Review 72: 157-160. Orlin, E.M. 1997. Temples, Religion and Politics in the Roman Republic, Leiden: E.J. Brill. Packer, J.E. 1997. The Forum of Trajan in Rome: A Study of the Monuments, Berkeley: University of California Press. _____ 2001. The Forum of Trajan in Rome: A Study of the Monuments in Brief, Berkeley: University of California Press. Paderni, C. 1753. ‘Extract of a letter from Signor Camillo Paderni, to Dr Mead, concerning the Antiquities dug up from ancient Herculaneum, dated from Naples, Nov. 18, 1752’, Philosophical Transactions 48: 71-73. Parsons, E.A. 1952. The Alexandrian Library: Glory of the Hellenic World, London: Cleaver-Hume Press. Pavord, A. 2005. ‘Pliny the plagiarist’, in A. Pavord, ed., The Naming of Names: The Search for Order in the World of Plants, London: Bloomsbury. Petroski, H. 1999. The Book on the Bookshelf, New York: Vintage Books. Pfeiffer, H.F. 1931. ‘The Roman library at Timgad’, Memoirs of the American Academy in Rome 9: 157-165. Pfeiffer, R. 1968. History of Classical Scholarship: From the Beginninsg to the End of the Hellenistic Age, Oxford: Clarendon Press. Phillips, J.J. 1985. ‘Book prices and Roman literacy’, Classical World 79: 36-38. _____ 1986, ‘Atticus and the publication of Cicero’s works’, Classical World 79: 227-237. Pickard-Cambridge, A. 1968. The Dramatic Festivals of Athens, Oxford: Clarendon Press.

137

Pierce, E.D. 1922. Gaius Asinius Pollio: A Man of Letters, PhD Thesis, New York: Columbia University. Platner, S.B. 1965. A Topographical Dictionary of Ancient Rome, completed and revised by T. Ashby, Rome: ‘L’Erma’ di Bretschneider. Platthy, J. 1968. Sources on the Earliest Greek Libraries, with the Testimonia, Amsterdam: Adolf M. Hakkert. Pollitt, J.J. 1974. The Ancient View of Greek Art: Criticism, History and Terminology, New Haven: Yale University Press. Posner, E. 1972. Archives in the Ancient World, Cambridge: Harvard University Press Purcell, N. 1993. ‘Atrium Libertatis’, Papers of the British School at Rome 61: 125-155. Rawson, E. 1985. Intellectual Life in the Roman Republic, London: Duckworth. Reichmann, F. 1938. ‘The book trade at the time of the Roman Empire’, Library Quarterly 8: 40-77. Reid, N. 1976. ‘Some neglected evidence on the early career of Tacitus’, Classical Quarterly 26: 309- 314. Reid, J.S. 1885, M. Tulli Ciceronis Academica: The Text Revised and Explained, London; Macmillan. Reynolds, L.D. and Wilson, N.G. 1991. Scribes and Scholars: A Guide to the Transmission of Greek and Latin Literature, 3rd edn., Oxford: Clarendon Press. Richardson, L. 1976. ‘The evolution of the Porticus Octaviae’, American Journal of Archaeology 80: 57-64. _____ 1977. ‘The libraries of Pompeii’, Archaeology 30: 394-402. _____ 1988. Pompeii: An Architectural History, Baltimore: Johns Hopkins University Press. _____ 1992. A New Topographical Dictionary of Ancient Rome, Baltimore: Johns Hopkins University Press. Riche, P. 1976. Education and Culture in the Barbarian West: From the Sixth through the Eighth Century, Columbia: University of South Carolina Press. Richmond, O.L. 1914. ‘The Augustan Palatium’, Journal of Roman Studies 4: 193-226. Riddle, J.M. 1988. ‘Review: R. French and F Greenaway, eds., Science in the Early Roman Empire: Pliny the Elder, his Sources and Influence’, American Historical Review 93: 398. Roller, D.W. T World of J ba II and Kl o a ra S l n : Royal S olars i on Ro ’s Afri an Frontier, New York: Routledge. Rose, H.J. 1954. A Handbook of Latin Literature, London: Methuen.

138

Rushforth, G.M. 1902. ‘The Church of S. Maria Antiqua’, Papers of the British School at Rome 1: 1- 119. Savage, S. 1945, ‘Remotum a notitia vulgarum’, Pro dings and Transa ions of A ri an Philological Association 76: 157-165. Schanz, M. 1935. Geschichte der römischen Literatur: bis zum Gesetzgebungswerk des Kaisers Justinian, Munich: C.H. Beck. Scheid, J. 2006, ‘Oral and written tradition in the formation of sacred law in Rome’, in C. Ando and J. Rupke, eds., Religion and Law in Classical and Christian Rome, Stuttgart: Franz Steiner. Schultze, C. 1986. ‘Dionysus of Halicarnassus and his audience’, in I.S Moxon, J.D. Smart and A.J. Woodman, eds., Past Perspectives: Studies in Greek and Roman Historical Writing, Cambridge: Cambridge University Press. Schulz, F. 1946. History of Roman Legal Science, Oxford: Clarendon Press. Schwartz, J. 1951. ‘Pompeius Macer et la jeunesse d’Ovide’, Revue de philologie de litterature et d’ is oir an i nn s 25: 182-194. Segala, E. and Sciortino, I. 1999. Domus Aurea, Rome: Electa. Sellers, E. (commentary and historical introduction) and Jex-Blake, K. (trans.) 1896. T ld r Pliny’s Chapters on the History of Art, London; Macmillan. Sesso, F.B. 1968. ‘History of the public library’, in A. Kent and H. Lancour, eds., Encyclopedia of Library and Information Science 24: 267-290. Severance, H.O. 1910. ‘Three of the earliest book catalogs’, Public Libraries 10: 116-117. Shatzman, I 1972, ‘The Roman general’s authority over booty’, Historia 21: 177-205. Sider, S. 1990, ‘Herculaneum’s library in 79 AD: the Villa of the Papyri’, Libraries and Culture 25: 534-542. Sihler, E.G. 1905, ‘The Collegium Poetarum at Rome’, American Journal of Philology 26: 1-21. Sisson, M.A. 1929. ‘The Stoa of Hadrian at Athens’, Papers of the British School at Rome 11: 250- 272. Sommer, R. 1926. ‘Titus Pomponius Atticus und die verbreitung von Werken’, Hermes 61: 389-422. Stambaugh, J.E 1978, “The functions of Roman temples”, in Aufstieg und Niedergang der Römischen Welt, 2.16.1, Berlin: Walter de Gruyter, 554-608. Stannard, J. 1965. ‘Pliny and Roman botany’, Isis 56: 420-425.

139

Starr, R.J. 1987. ‘The circulation of literary texts in the Roman world’, Classical Quarterly 37: 213- 223. _____ 1990. ‘The used book trade in the Roman world’, Phoenix 44: 148-157. _____ 1991. ‘Reading aloud: lectores and Roman reading’, Classical Journal 86: 337-343. Stein, P. 1966, Regulae iuris: from juristic rules to legal maxims, Edinburgh: Edinburgh University Press. Steiner, G. 1955. ‘The scepticism of Pliny the Elder’ Classical Weekly 48: 137-143. Stevenson, S.W. 1964. Dictionary of Roman Coins, London: Seaby. Strocka, V.M. 1981. ‘Römische bibliotheken’, Gymnasium 88: 298-329. Sutherland, C.H.V. 1945. ‘Aerarium and fiscus during the early Empire’, American Journal of Philology 66: 151-170. Swain, S.C.R. 1990, ‘Hellenic heroes and Roman heroes of Plutarch’, Journal of Hellenic Studies 110: 126-145. Syme, R. 1969. 1969. ‘Pliny the procurator’, Harvard Studies in Classical 73: 201-236. _____ 1978. History in Ovid, Oxford: Clarendon Press. _____ 1991. ‘Hadrian as intellectual’, in A.R. Birley, ed., Roman Papers VI, Oxford: Clarendon Press, 103-114. Thompson, J.W. 1940, Ancient Libraries, Hamden: Archon Press. Too, Y.L. 2010, The Idea of the Library in the Ancient World, Oxford: Oxford University Press Towend, G.B. 1961. ‘The Hippo inscription and the career of Suetonius’, Historia 10: 96-109. Tucci, P.L. 2008. ‘Galen’s storeroom, Rome’s libraries, and the fire of A.D. 192’. Journal of Roman Archaeology 21: 133-149. Van Deman, E.B. 1924. ‘The House of Caligula’, American Journal of Archaeology 28: 368-398. Van de Woestyne, P. 1929. ‘Un ami d’Ovide: C. Iulius Hyginus’, Le Musée Belge 33: 33-45. Wallace, R. 2005. An Introduction to Wall Inscriptions from Pompeii and Herculaneum, Wauconda, Ill.: Bolchazy-Carducci Publishers. Wallace-Hadrill, A. 1983. Suetonius: The Scholar and his Caesars, London: Duckworth. Ward-Perkins, J.B. 1956. ‘Nero’s Golden House’, Antiquity 30: 209-219. _____ 1981, Roman Imperial Architecture, New York: Penguin Books. Waters, B.A. and Winters, W.C. 1987. ‘On the verge of a revolution: current trends in library lighting’, Library Trends 36: 327-349.

140

Weichardt, C.F.W. 1900. Tib ri s’ Villa and o r Ro an B ildings on the Isle of Capri, Leipzig: K.F. Koehler. Weihs, J. 2009. ‘A brief history of shelving’, Technicalities 29: 11-14. Wendel, C. 1949. ‘Die bauliche Entwicklung der Antike Bibliothek’, Zentralblatt für Bibliothekswesen 63: 407-418. Westrup, C.W. 1929, ‘On the antiquarian-historiographical activities of the Roman pontifical college’, Det Kongelige Danske Viderskabernes Selskab Historiske-Filologiske 16: 1-49. White, K.D. 1973. ‘Roman agricultural writers 1: Varro and his predecessors’, in Aufstieg und Niedergang der Römischen Welt, Berlin: Walter de Gruyter 1.4: 439-497. White, P. 2009. ‘Bookshops in the literary ’, in W.A. Johnson and H.N. Parker, eds., Ancient Literacies: The Culture of Reading in Greece and Rome, Oxford: Oxford University Press, 268-287. Winckelmann, G.G. 1831. Opere di G.G. Winckelmann: Tomo VII, Prato: Giachetti. Wiseman, T.P. 1987, ‘Conspicui postes tectaque digna deo: the public image of aristocratic and imperial houses in the Late Republic and early Empire’, Collection d l’É ol fran ais d Rome 98: 393-414. _____ 1998. Roman Drama and Roman History, Exeter: University of Exeter Press. _____ 2007. ‘Practice and theory in Roman historiography’, History 66: 375-393. Witty, F.J. 1958. ‘The Pinakes of Callimachus’, Library Quarterly 28: 132-136. Woolf, G. 2009. ‘Literacy or literacies in Ancient Rome?’, in W.A. Johnson and H. Parker, eds., Ancient Literacies: The Culture of Reading in Greece and Rome, Oxford: Oxford University Press. Zecchini, G. 1982. ‘Asinio Pollione: dall’attavita politico all riflessione storiografica’, Aufstieg und Niedergang der Römischen Welt 2.30.2, Berlin: Walter de Gruyter, 1265-1296. Zetzel, J.E.G. 1972. ‘Cicero and the Scipionic Circle’, Harvard Studies in Classical Philology 76: 173- 179. _____ 1981. Latin Textual Criticism in Antiquity, New York: Arno Press.

141