Rome in Egypt's Eastern Desert: Volume One 9781479810680

Citation preview

Rome in Egypt’s Eastern Desert Volume 1 Hélène Cuvigny edited with an introduction by Roger S. Bagnall

New York New York University Press Institute for the Study of the Ancient World 2021

© 2021 Institute for the Study of the Ancient World NYU Press ISBN 978-1-479-81-0611 (hardcover, vol. 1) ISBN 978-1-479-81-0673 (ebook, vol. 1) ISBN 978-1-479-81-0680 (ebook other, vol. 1) ISBN 978-1-479-81-0697 (hardcover, vol. 2) ISBN 978-1-479-81-0703 (ebook, vol. 2) ISBN 978-1-479-81-0741 (ebook other, vol. 2) Library of Congress Cataloging-in-Publication Data Names: Cuvigny, Hélène, author. | Bagnall, Roger S., editor, translator. Title: Rome in Egypt's Eastern Desert / Hélène Cuvigny ; edited with an introduction by Roger S. Bagnall. Other titles: ISAW monographs. Description: New York : New York University Press, Institute for the Study of the Ancient World, 2021. | Series: ISAW monographs | Translated from French. | Includes bibliographical references and indexes. | Summary: "Rome in Egypt's Eastern Desert collects Prof. Cuvigny's most important articles on Egypt's Eastern desert during the Roman period. From the excavations of the forts that she has directed have come a wealth of material, including tens of thousands of texts written on potsherds (ostraca). Some of these are administrative texts, but many more are correspondence, both official and private, written by and to the people (mostly but not all men) who lived and worked for periods of time in these remote and harsh environments, supported by an elaborate network of defense, administration and supply that tied the entire region together. The contents have all been published earlier in peer-reviewed venues, but almost entirely in French. All contributions have been translated or checked by the editor and brought up to date with respect to bibliography and in some cases significantly rewritten by the author, in order to take account of the enormous amount of new material discovered in the intervening time and subsequent publications. A full index will make this body of work far more accessible than it now is. This book brings together thirty years of detailed study of this material, bringing to life the geography, administration, military, quarry operations, life in the forts, and the religion and expressive language of the population who lived in them"-- Provided by publisher. Identifiers: LCCN 2021017969 (print) | LCCN 2021017970 (ebook) | ISBN 9781479810642 (v. 1 ; hardcover) | ISBN 9781479810697 (v. 2 ; hardcover) | ISBN 9781479810611 (hardcover) | ISBN 9781479810673 (v. 1 ; ebook) | ISBN 9781479810680 (v. 1 ; ebook other) | ISBN 9781479810703 (v. 2 ; ebook) | ISBN 9781479810741 (v. 2 ; ebook other) Subjects: LCSH: Ostraka--Egypt--Eastern Desert. | Eastern Desert (Egypt)--Antiquities. | Mons Claudianus Site (Egypt) | Rome--History--Empire, 30 B.C.-284 A.D. | Egypt--History--Greco-Roman period, 332 B.C.-640 A.D. Classification: LCC DT73.E27 C89 2021 (print) | LCC DT73.E27 (ebook) | DDC 932/.1022--dc23 LC record available at https://lccn.loc.gov/2021017969 LC ebook record available at https://lccn.loc.gov/2021017970

The Greek fonts are IFAO-Grec Unicode and IFAO-Grec Exposant. Design by Andrew Reinhard Printed in the United States

ISAW Monographs ISAW Monographs publishes authoritative studies of new evidence and research into the texts, archaeology, art history, material culture, and history of the cultures and periods representing the core areas of study at NYU’s Institute for the Study of the Ancient World. The topics and approaches of the volumes in this series reflect the intellectual mission of ISAW as a center for advanced scholarly research and graduate education whose aim is to encourage the study of the economic, religious, political, and cultural connections between ancient civilizations, from the Western Mediterranean across the Near East and Central Asia, to China. Roger S. Bagnall and Giovanni R. Ruffini, Ostraka from Trimithis, Volume 1 (Amheida I) (2012) George Hatke, Aksum and Nubia: Warfare, Commerce, and Political Fictions in Ancient Northeast Africa (2013) Jonathan Ben-Dov and Seth Sanders (eds.), Ancient Jewish Sciences and the History of Knowledge in Second Temple Literature (2014) Anna L. Boozer, A Late Romano-Egyptian House in the Dakhla Oasis: Amheida House B2 (Amheida II) (2015) Roger S. Bagnall, Nicola Aravecchia, Raffaella Cribiore, Paola Davoli, Olaf E. Kaper, and Susanna McFadden, An Oasis City (2016) Roger S. Bagnall, Roberta Casagrande-Kim, Cumhur Tanrıver, Graffiti from the Basilica in the Agora of Smyrna (2016) Rodney Ast and Roger S. Bagnall, Ostraka from Trimithis, Volume 2 (Amheida III) (2016) Nicola Aravecchia, ‘Ain el-Gedida: 2006–2008 Excavations of a Late Antique Site in Egypt’s Western Desert (Amheida IV) (2019) Roger S. Bagnall and Alexander Jones, Mathematics, Metrology, and Model Contracts: A Codex From Late Antique Business Education (P.Math.) (2020) Clementia Caputo, The House of Serenos, Part I: The Pottery (Amheida V) (2020) Jonathan Valk and Irene Soto Marín, Ancient Taxation: The Mechanics of Extraction in Comparative Perspective (2021) Hélène Cuvigny, Rome in Egypt’s Eastern Desert (2021)

Preface My first experience in papyrology was during the seminar of Jean Scherer in 1976/77. I was delighted with these sessions of reading under his guidance and found there a way to continue with Greek, while at the same time distancing myself from the school-like, literary approach that I had never liked. Documentary papyrology appealed to my taste for the concrete, the revealing detail and confirmed me in my aversion to rhetoric. Equally I was attracted to the history of normal, insignificant people, which coincided with a tradition in my family. And yet, when the time came to choose the subject for a doctoral thesis, which I would have liked to be an edition, Jean Scherer told me that the collection of the Sorbonne no longer contained any publishable new papyri, and that I would have to find myself another subject.1 I was about to drop Greek and papyrology altogether when a secretary at the EPHESS advised me to meet Pierre Vidal-Naquet. This encounter was a turning-point, and Vidal-Naquet rescued me for papyrology, but he did not have any unpublished papyri to give me. So, when, in September 1984, thanks again to Vidal-Naquet,2 I arrived at the IFAO as “pensionnaire,” I was starving for texts. In Cairo I had the good fortune to find Guy Wagner, who at once associated me to the publication of the ostraca from Douch, and who, in his very special way, made me discover the Egyptian countryside and its possibilities, something that has been of lasting importance for me. It was also he who, in 1985, took me to Mons Claudianus in the company of his “pals” as he called them, Adam Bülow-Jacobsen, Wilfried Van Rengen, and Walter Cockle.3 I can still see him standing on what became the South Sebakh, pointing a finger to the ground, while he said “This is where one ought to dig.” In 1986 we returned to Mons Claudianus in the company of Jean Bingen, and with his support it was easy to convince Paule Posener, then director of the IFAO, of the interest of opening an excavation on Mons Claudianus. This was how it all began. Later on, the successors of Mme Posener never let us down. Though Egyptologists, they deemed it worthwhile to fund searches of which one clearly stated objective was to look for Greek and Latin ostraca: Nicolas Grimal, who allowed us to undertake our program in the Desert of Berenike, which in turn grew in scale with the support of Bernard Mathieu, then of Laure Pantalacci and Béatrix Midant-Reyne. This program benefited from financing both from the Ministry of Foreign Affairs and from the IFAO. I also owe a debt of gratitude to the Supreme Council of Antiq1. During the congress of papyrology in New York, I naïvely told this to Naphtali Lewis and it made him laugh. 2. P. Vidal-Naquet, Mémoires, II (Paris 1998) 222. 3. These “pals” were the participants in the CIPSH photographic mission to the Egyptian Museum, and they had already several times made excursions with Wagner into the Eastern Desert.

v

vi

Rome in Egypt’s Eastern Desert

uities in Egypt for the good will of its directors, and to the long series of devoted inspectors who have accompanied us in the desert. They had much work smoothing out all the administrative difficulties that constantly cropped up. Those responsible at the store house in Dendera, later in Quft, where we spent many months improving readings, making joins, and re-photographing our finds, have always made us most welcome, made sure that we wasted as little time as possible, and moaned with us when work was blocked by power cuts. The two teams with which I have had the good fortune to work were marked by personalities, both learned and engaging, and also able to find, to everyone’s satisfaction, a compromise between the opposing interests of the archaeologists and the papyrologists. At Mons Claudianus it was Jean Bingen. From 1994 and for two decades, Jean-Pierre Brun, as a side-line to his career, excavated the dumps of several praesidia with unchanging patience and good humor, even going so far as to share the joys and disappointments of the papyrologists with amused indulgence. He was extremely popular with the workers and made a remarkable excavator of Baghdadi Mohammed, who was later to become our rais. I cannot mention all the friends and colleagues who have participated in the adventure. At Mons Claudianus, it was fascinating for me to be part of a team that united so many university professors and confirmed scholars. At table the conversation was always lively, and erudition was mixed with amusing anecdotes. Life in the field puts you face-to-face with ambiguous situations, makes you share simple pleasures and material difficulties, relaxes personalities, and simplifies human relations. This spirit continued in the Desert of Berenike, where the two pillars of the excavations, Michel Reddé and Jean-Pierre Brun, were both of humorous disposition. These two bons vivants would certainly have enjoyed the experience much less without the precious help of Ashraf al-Azob who joined us in the first year of Didymoi as a creative cook, a reliable driver, and a solver of all sorts of problems. Year after year the number of ostraca has grown, and they have provided us with fertile research material. Chapter 15 gives a history of our excavations, and Chapter 41 is a synopsis of our finds. Every papyrologist knows the interest of working on documents with a precise provenance. When there are enough of them and they are linked in series, they allow you to make better founded hypotheses. These ostraca have sometimes given me the mischievous pleasure of throwing spanners into learned works, of upsetting received wisdom, and of changing the meaning of words. With one archaeological site following after another, each of them providing new light, I have over the years been able to create a general view, which is found in Chapters 15, 18, and 41. When we began work, the desert sites were mostly as the Romans had left them. A few tourists and Beduins had here and there made holes in the dumps—Jean Bingen used to call them “rabbit holes.” Since 2015 the destruction of the archaeological remains has accelerated dramatically. Now the Egyptian mining companies, and not least the clandestine gold-diggers, attack the sites with bulldozers, cut through walls, dig under cisterns, and flatten the dumps. Just imagine what we might have learned if we had found the ostraca of Abbad, the site where the caravans of Ptolemaic elephant-hunters were formed before departure! In 2013, I passed the direction of the programme to Bérangère Redon and Thomas Faucher, who have given it a new impetus. In order to face conditions that are in all respects more difficult than in my time, youth is needed, an energy that I no longer have, nerves of steel, and an adherence to the new ideal of research as management. I hope that circumstances will not upset the efforts of my two talented successors. Bérangère has set afoot a large project concerning the whole of the Eastern Desert.4 In this context she has asked a young geographer, Louis Manière, following our indications, to create the map that is found at the beginning of this book. 4. “Desert Networks,” financed by the European Research Council.

Preface

vii

I owe a great debt to Adam Bülow-Jacobsen. His practical sense, patience, and sense of diplomacy have been invaluable. During all these years he has borne, along with me, the stress of a desert excavation where logistics are often on a knife’s edge. On top of the administrative complications there were always cars and generator breaking down, which disrupted not only the work, but even the provision of water. I have always been afraid of being without a working car when someone needed to be brought to hospital, but fortunately it has never happened, though emergency transports have been necessary more than once. When our driver, Gamil, had heart trouble in the middle of the night at Xeron, Adam was able to bring him to the hospital in Edfu. Nearly all the photographs in this volume, and certainly all those of ostraca are his. In 2005 he had the bright idea of using a camera that had been adapted for infrared. The results surpassed our expectations, and completely changed our work method on site. Working directly on the ostraca in the blinding light of 500W lamps and with magnifying glasses has been replaced by comfortable work on the screen on high contrast and enlargeable images. During several weeks of every season Adam rephotographed in infrared our finds from previous years. Besides all this, he has been important as a second eye and discussion-partner in the editions. And finally, I need to express my gratitude to Roger Bagnall. Not only did he have the idea for this book, but he also generously offered to translate half of the texts. The rest, with few exceptions, have been translated by Adam. My papers on the desert have become chapters and have been updated. Some have been extensively reworked (Ch. 6 and 7) or made longer (Ch. 8 and 26), and some repetitions have been taken away. Roger as an editor is as agreeable as he is efficient and conscientious. He has left me complete freedom to rework my old papers. He has read the manuscript several times and has found repetitions, contradictions, and slips. His attention to form does not make him lose the overall view. He pushed me at the last moment to improve a reading in an ostracon where I thought I had reached my limits, but he rightly felt that progress was still possible. Finally this book owes a lot to David Ratzan, editor of publications for ISAW. I thank him for all the time and attention he patiently devoted to the manuscript and the proofs. The ostraca that are the subject of these pages are linked to good memories. The superb landscapes of the desert. Night walks in full moon among the rocks at Mons Claudianus, where one saw everything as clearly as in daylight, but in icy shades of black and white. A choir of wild dogs at night, so melodious that in my half-sleep I thought I was hearing opera. The reading sessions in the tent in the late afternoon and at aperitif time, when the acrid smell of newly excavated ostraca mixed with the aroma of peated whisky.

Contents Preface List of Illustrations List of Tables Original Publications of the Chapters Map of the Eastern Desert

v xii xvii xix xxii

Volume 1 Introduction and Survey Introduction (Roger S. Bagnall) 1. A survey of place-names in the Egyptian Eastern Desert during the principate according to the ostraca and the inscriptions

1 7

Part I. Senior Administrators 2. Ulpius Himeros, imperial procurator (I.Pan 53) 3. Sulpicius Serenus, procurator Augusti, and the titulature of the prefects of Berenike 4. Claudius Lucilianus, prefect of an ala and of Berenike 5. Vibius Alexandros, prefect and epistrategos of the Heptanomia 6. Procurator Montis

81 91 113 117 127

Part II. The Quarries 7. Greek ostraca from Mons Claudianus, revisited 8. An inscription of an ἐργοδότης in a quarry at Mons Claudianus 9. The amount of the wages paid to the quarry-workers at Mons Claudianus 10. Two ostraca from Mons Claudianus: O.Bahria 20 and 21 11. The organization chart of the personnel of an imperial quarry according to an ostracon of Mons Claudianus 12. A dedication to Zeus Helios Great Sarapis in honor of a desector on an ostracon from Mons Claudianus

ix

135 147 165 175 181 213

x Part III. The Road System and the Mail 13. An unpublished ostracon from the Eastern Desert and the provenance of O.Amst. 9 14. Kaine, a new town: an experiment in familial reunification in the second century AD 15. The road system of the Eastern Desert of Egypt under the Early Empire in the light of the excavated ostraca 16. Collection of cases of irregularities in the transmission of official mail 17. The postal register of Turbo, curator of the praesidium of Xeron Pelagos 18. Men and gods in a network

223 227 233 261 267 299

Volume 2 Part IV. Military Rations 19. A receipt for military rations in exchange for payment of publica 20. The monthly ration of a cavalryman and his horse according to an ostracon from the praesidium of Dios 21. An unrecognized type of military administrative document: the order for payment of frumentum praeteritum (O.Claud. inv. 7235 and Ch.L.A. XVIII 662)

325 337 355

Part V. Business and Prostitution 22. Conductor praesidii 23. Quintana, a woman transformed into a tax 24. Rotating women: remarks on prostitution in the Roman garrisons of the Desert of Berenike 25. “Me too” in the praesidia, or when reality meets theatrical fiction

367 375 379 389

Part VI. Desert Dwellers 26. Kinaidokolpitai in a Greek ostracon from the Eastern Desert 27. Papyrological evidence on “Barbarians” in the Egyptian Eastern Desert 28. Public post, military intelligence, and dry cisterns: the letters of Diourdanos to Archibios, curator Claudiani

397 415 439

Part VII. Religion and Magic 29. Twilight of a god: the decline of the cult of Pan in the Eastern Desert 30. A soldier of the cohors I Lusitanorum at Didymoi: once again on the inscription I.Kanaïs 59bis 31. The shrine in the praesidium of Dios (Eastern Desert of Egypt): graffiti and oracles in context 32. The prefect of Egypt demobilizes some overage men and imposes a preventive “seal” (tattoo?) 33. “The wheat for the Jews” (O.KaLa. inv. 228) 34. The oldest representation of Moses, drawn by a Jew around AD 100

465 473 479 527 539 545

xi Part VIII. Language 35. Πλήρωμα in the identification of soldiers in the navy 36. Remarks on the use of ἴδιος in the epistolary prescript 37. Πέμπειν/ἀγοράζειν τῆς τιμῆς in Greek letters from Egypt 38. The names of cabbage in the Greek ostraca from the Eastern Desert: κράμβη, κραμβίον, καυλίον 39. Χίλωμα = Haversack 40. “When Heroïs has given birth…” ἐάν = ὅταν in temporal clauses referring to the future

557 561 571 577 585 591

Conclusion 41. Are ostraca soluble in history?

597

Abbreviations Bibliography

617 621

Indexes 1. Sources A. Greek and Latin authors B. Greek and Latin ostraca, papyri, and tablets C. Greek and Latin inscriptions 2. Persons 3. Places and ethnic names 4. Greek and Latin words A. Greek a. Lexical b. Personal and geographical names B. Latin a. Lexical b. Personal and geographical names 5. Subjects

639 639 639 640 641 643 647 647 647 649 650 650 651 652

xii

List of Illustrations Introduction Map of the Eastern Desert. © Louis Manière (ERC project “Desert Networks,” CNRS) Chapter 1 Figure 1. The Eastern Desert in Roman times. © J.-P. Brun Figure 2. Pliny’s marmor Tibereum: column in granodiorite from Tiberiane in Zenon’s chapel at St. Praxedes. © A. Bülow-Jacobsen Figure 3. A porphyry plate from Umm Tuwat polished on one side, found in the midden at Umm Balad. © Rights reserved by H. Cuvigny Figure 4. A latomia at Umm Tuwat. © A. Bülow-Jacobsen Figure 5. The track towards Umm Tuwat. © A. Bülow-Jacobsen Figure 6. Porphyry column from Umm Tuwat in Zenon’s chapel at St. Praxedes. © A. Bülow-Jacobsen Figure 7. Map of the Tabula imperii romani, Sheet Coptos, published by Meredith 1958. © Rights reserved Figure 8. O.KaLa. inv. 269. © A. Bülow-Jacobsen Figure 9. O.Claud. IV 816.1 (detail). © A. Bülow-Jacobsen Figure 10. O.Claud. IV 866.4 (detail). © A. Bülow-Jacobsen Figure 11. O.Claud. IV 841.11 (detail). © A. Bülow-Jacobsen Figure 12. O.Claud. IV 710.3 (detail). © A. Bülow-Jacobsen Figure 13. O.Claud. IV 779.2 (detail). © A. Bülow-Jacobsen Figure 14. O.Claud. IV 783.1 (detail). © A. Bülow-Jacobsen Figure 15. O.Claud. IV 841.30–33. © A. Bülow-Jacobsen Figure 16. Column base inscribed on the underside, from the quarry of Myrismos. © A. Bülow-Jacobsen Figure 17. Nikotychai quarry, remained untapped. © A. Bülow-Jacobsen Figure 18. The signpost of Nikotychai quarry, with the name of the foreman. © A. Bülow-Jacobsen Figure 19. O.Claud. IV 747.6: [Νικ]οτυχ( ) ἀκισκ(λάριοι) β. © A. Bülow-Jacobsen Figure 20. Ο.Claud. IV 739.1. © A. Bülow-Jacobsen Figure 21. O.Claud. IV 742.1. © A. Bülow-Jacobsen Figure 22. O.Claud. IV 841.51. © A. Bülow-Jacobsen Figure 23. Dipinto under the base of the giant column. © A. Bülow-Jacobsen Figure 24. O.Claud. IV 658.1 (detail). © A. Bülow-Jacobsen Figure 25. Mons Claudianus. Column shafts on the krepis at the bottom of the Pillar-Wadi. © A. BülowJacobsen Figure 26. The rock of Krokodilo seen from the northeast. © H. Cuvigny Figure 27. A rock graffito on a nearby cliff, probably inspired by the shape of the hill. © H. Cuvigny Figure 28. The praesidium of Phalakron. © M. Reddé Figure 29. O.Xer. inv. 995, fr. c: Kabalsi? mentioned at the end of line 14. © A. Bülow-Jacobsen Figure 30. Ostracon found at Abu Zawal. © R. Klemm Figure 31. Tracks to Porphyrites, according to Maxfield and Peacock 2001: 194. © S. Goddard Figure 32. Dayr al-Atrash. © A. Bülow-Jacobsen Figure 33. Black hills behind Dayr al-Atrash. © A. Bülow-Jacobsen

xiii Figure 34. The lintel of the room of cisterns at Mons Claudianus. © A. Bülow-Jacobsen Figure 35. O.Claud. inv. 7955. © A. Bülow-Jacobsen Figure 36. O.Claud. IV 632.1–2. © A. Bülow-Jacobsen Figure 37. Maximianon (Al-Zarqaʾ), Myos Hormos, Biʾr Karim (from Meredith 1958). © Rights reserved Figure 38. O.Krok. I 120. © A. Bülow-Jacobsen Figure 39. O.Max. inv. 1099.7. © A. Bülow-Jacobsen Chapter 2 Figure 40. Sites in Wadi Samna (drawing by Tregenza, after BIFAO 96 [1996] 91). Figure 41. I.Pan 53 (AE 2001, 2036). © A. Lecler (IFAO) Figure 42. Fragments found in 1994 at Samna. © J.-Fr. Gout (IFAO) Chapter 3 Figure 43. O.Dios inv. 90. © A. Bülow-Jacobsen Chapter 4 Figure 44. I.Did. 9. © A. Bülow-Jacobsen Chapter 5 Figure 45. SB XXVIII 16941. © A. Bülow-Jacobsen Chapter 6 Figure 46. O.Did. 40. © A. Bülow-Jacobsen Chapter 9 Figure 47. SB XXIV 16173. © A. Bülow-Jacobsen Chapter 10 Figure 48. SB XXIV 16061. © A. Bülow-Jacobsen Chapter 11 Figure 49. O.Claud. inv. 1538+2921. © A. Bülow-Jacobsen Chapter 12 Figure 50. O.Claud. inv. 7363. © A. Bülow-Jacobsen Chapter 13 Figure 51. SB XXII 15541. © A. Bülow-Jacobsen Chapter 16 Figure 52. P.Worp 51. © A. Bülow-Jacobsen

xiv Chapter 17 Figure 53. 1, col. i. © A. Bülow-Jacobsen Figure 54. 1, col. iii. © A. Bülow-Jacobsen Figure 55. 1, col. iii, bottom. © A. Bülow-Jacobsen Figure 56. 2. © A. Bülow-Jacobsen Figure 57. 3. © A. Bülow-Jacobsen Figure 58. 4. © A. Bülow-Jacobsen Figure 59. 5. © A. Bülow-Jacobsen Figure 60. 6. © A. Bülow-Jacobsen Figure 61. 7. © A. Bülow-Jacobsen Figure 62. 8. © A. Bülow-Jacobsen Chapter 18 Figure 63. O.Xer. inv. 995, fr. C, end of line 14. © H. Cuvigny Figure 64. O.Xer. inv. 809. © A. Bülow-Jacobsen Figure 65. O.Xer. inv. 810. © A. Bülow-Jacobsen Figure 66. O.Xer. inv. 829. © A. Bülow-Jacobsen Figure 67. O.Krok. I 1. © A. Bülow-Jacobsen Figure 68. O.Dios inv. 807. © A. Bülow-Jacobsen Figure 69. O.Dios inv. 145. © A. Bülow-Jacobsen Figure 70. O.Dios inv. 1246. © A. Bülow-Jacobsen Figure 71. O.Blem. 57. © A. Bülow-Jacobsen Chapter 19 Figure 72. O.Dios inv. 480. © A. Bülow-Jacobsen Figure 73. O.Claud. inv. 4524+5700. © A. Bülow-Jacobsen Figure 74. O.Dios inv. 972. © A. Bülow-Jacobsen Chapter 20 Figure 75. O.Dios inv. 626. © A. Bülow-Jacobsen Figure 76. O.Xer. inv. 464. © A. Bülow-Jacobsen Figure 77. O.Krok. I 70. © A. Bülow-Jacobsen Chapter 21 Figure 78. O.Claud. inv. 7235. © A. Bülow-Jacobsen Figure 79. Ch.L.A. XVIII 662. © A. Bülow-Jacobsen Chapter 22 Figure 80. P.Bagnall 11. © A. Bülow-Jacobsen Figure 81. P.Bagnall 12. © A. Bülow-Jacobsen Chapter 25 Figure 82. O.Dios inv. 439. © A. Bülow-Jacobsen

xv Chapter 26 Figure 83. SB XXIV 16187. © A. Bülow-Jacobsen Figure 84. Map of the Red Sea. © H. David Chapter 27 Figure 85. O.Krok. I 49. © A. Bülow-Jacobsen Figure 86. O.Blem. 83. © A. Bülow-Jacobsen Chapter 28 Figure 87. 1. © A. Bülow-Jacobsen Figure 88. 2. © A. Bülow-Jacobsen Figure 89. 3. © A. Bülow-Jacobsen Figure 90. 4. © A. Bülow-Jacobsen Figure 91. After TAVO B V 21. © Dr. Ludwig Reichert Verlag Wiesbaden; reprinted by permission Figure 92. The intelligence cycle. Figure 93. 5. © A. Bülow-Jacobsen Figure 94. 6. © A. Bülow-Jacobsen Figure 95. 7. © A. Bülow-Jacobsen Figure 96. 8. © A. Bülow-Jacobsen Figure 97. 9. © A. Bülow-Jacobsen Chapter 29 Figure 98. O.Max. inv. 1214. © A. Bülow-Jacobsen Figure 99. The route of Myos Hormos. Chapter 30 Figure 100. I.Did. 7. © A. Bülow-Jacobsen Chapter 31 Figure 101. Plan of Dios (drawing by J.-P. Brun, M. Reddé, E. Botte). Figure 102. 1. © A. Bülow-Jacobsen Figure 103. The aedes in its surroundings. © J.-P. Brun Figure 104. Schematic plan of the aedes (drawing by J.-P. Brun). Figure 105. A–B section of the aedes looking south (drawing by J.-P. Brun). Figure 106. The aedes, phase 5. © J.-P. Brun Figure 107. Receptacle at the southern end of the podium. © J.-P. Brun Figure 108. The podium seen from above. © J.-P. Brun Figure 109. The brick floor, phase 7. © J.-P. Brun Figure 110. The podium at the time of the discovery. © J.-P. Brun Figure 111. 2. © A. Bülow-Jacobsen Figure 112. 3. © A. Bülow-Jacobsen Figure 113. 4. © A. Bülow-Jacobsen Figure 114. 5–7. © A. Bülow-Jacobsen

xvi Figure 115. 8. © A. Bülow-Jacobsen Figure 116. 9–10. © A. Bülow-Jacobsen Figure 117. 11. © A. Bülow-Jacobsen Figure 118. 12. © A. Bülow-Jacobsen Figure 119. 13. © A. Bülow-Jacobsen Figure 120. 14. © A. Bülow-Jacobsen Figure 121. 15. © A. Bülow-Jacobsen Figure 122. 16. © A. Bülow-Jacobsen Figure 123. 17. © A. Bülow-Jacobsen Figure 124. 18. © A. Bülow-Jacobsen Figure 125. 19. © A. Bülow-Jacobsen Figure 126. 20. © A. Bülow-Jacobsen Figure 127. 21. © A. Bülow-Jacobsen Figure 128. 22. © A. Bülow-Jacobsen Figure 129. 23. © A. Bülow-Jacobsen Chapter 32 Figure 130. O.Dios inv. 568. © A. Bülow-Jacobsen Chapter 33 Figure 131. O.KaLa. inv. 228. © A. Bülow-Jacobsen Chapter 34 Figure 132. The praesidium of Umm Balad. © A. Bülow-Jacobsen Figure 133. O.KaLa. inv. 179. © A. Bülow-Jacobsen Figure 134. Etching by Marc Chagall, part of a cycle of illustrations to the Bible. © Centre Pompidou, MNAM-CCI. Dist. RMN-Grand-Palais/Philippe Migeat. © Adagp, Paris 2021 Figure 135. The naturalist Michel Aymerich manipulating a cobra. © Clothilde Rojat Chapter 36 Figure 136. O.Krok. I 81. © A. Bülow-Jacobsen Chapter 38 Figure 137. A variety of kale: Brassica oleracea italica. Figure 138. A καυλίον. © A. Bülow-Jacobsen Figure 139. P.IFAO II 6, r°. Chapter 41 Figure 140. Types of texts on ostraca at Xeron Pelagos. Figure 141. O.Xer. inv. 348. © A. Bülow-Jacobsen Figure 142. O.Claud. inv. 6366+7149. © A. Bülow-Jacobsen Figure 143. The most active people in the sutlers’ network according to the ostraca from Krokodilo (after O.Krok. II, p. 12).

xvii

List of Tables Chapter 1 Table 1.1. Distinction between appellative and proprial. Table 1.2. Names of metalla in the Eastern Desert: semantic classification. Table 1.3. Δομιτιανή and Καινὴ Λατομία. Distribution of occurrences according to stratigraphy. Table 1.4. Taxonomy of the quarry-names at Mons Claudianus. Table 1.5. The toponyms mentioned in the ostraca from Umm Balad. Table 1.6. Names of the praesidia. Table 1.7. Use of the article before place-names. Chapter 3 Table 3.1. The prefects of Berenike: prosopography. Table 3.2. Geological formations of the Late Proterozoic in the vicinity of Dios. Table 3.3. Hardness (Mohs) of the rocks present in the vicinity of Dios. Chapter 5 Table 5.1. Ostraca dated to the reign of Commodus found at Mons Claudianus. Chapter 9 Table 9.1. Dacian mining contracts that mention wages. Chapter 11 Table 11.1. Number of blacksmiths and quarrymen in each work area. Table 11.2. Staffing of four quarries on different days. Chapter 15 Table 15.1. Proportion of private letters in earlier and later middens. Chapter 17 Table 17.1. Complete list of the fragments of Turbo’s postal register. Table 17.2. Movement of messengers in Turbo’s register. Chapter 18 Table 18.1. Proskynemata in ostracological corpora from praesidia in the Berenike Desert. Table 18.2. Appendix. The names of the praesidia on the roads to Myos Hormos and to Berenike. Chapter 20 Table 20.1. Conversion of Egyptian weight units in kg. Table 20.2. Equine rations in O.Dios inv. 626 and O.Krok. 18. Table 20.3. Equine rations in O.Dios inv. 626 and P.Oxy. 4087–4088.

xviii Chapter 24 Table 24.1. Itemized bill for the hire of the prostitute Serapias. Table 24.2. Hire and conditions of payment of the quintana in the correspondence of the procurers. Chapter 29 Table 29.1. Incidence of proskynemata in O.Krok. and O.Max. Chapter 31 Table 31.1. Dios oracles. Table 31.2. The gods of the praesidia according to the written evidence and the archaeology. Chapter 36 Table 36.1. Papyrological attestations of ἴδιος in apposition to the name of the addressee. Table 36.2. Ἴδιος as a possessive adjective relating to a trade, a title, a noun. Chapter 38 Table 38.1. Distribution of the occurrences of κράμβη and καυλίον in the ostraca of various periods. Chapter 41 Table 41.1. Number of ostraca inventoried by site.

xix

Original Publications of the Chapters 1. “La toponymie du désert Oriental égyptien sous le Haut-Empire d’après les ostraca et les inscriptions,” in J.-P. Brun, T. Faucher, B. Redon, and S. Sidebotham (eds.), Le désert Oriental d’Égypte durant la période gréco-romaine: bilans archéologiques (Paris 2018) 35 figg. http://books.openedition.org/ cdf/5154. 2. “Ulpius Himerus, procurateur impérial (I.Pan 53),”BIFAO 96 (1996) 91–101, 1 figure, 1 map, 1 pl. 3. “Sulpicius Serenus, procurator Augusti, et la titulature des préfets de Bérénice,” Chiron 37 (2007) 11–33, 1 pl. (with Adam Bülow-Jacobsen). 4. “Claudius Lucilianus, préfet d’aile et de Bérénice,” in Traianos Gagos and Roger S. Bagnall (eds.), Essays and Texts in Honor of J. David Thomas = American Studies in Papyrology 42 (Oakville, Connecticut, 2001) 171–77 and pl. 23. 5. “Vibius Alexander, praefectus et épistratège de l’Heptanomie,” CdE 77 (2002) 238–48, 1 figure. 6. “Procurator Montis,” in Henri Melaerts (ed.), Papyri in honorem Johannis Bingen octogenarii editae (P.Bingen) = Studia varia Bruxellensia ad orbem Graeco-Latinum pertinentia V (Leuven, 2000) 415–18 and pl. 63. 7. “Nouveaux ostraca grecs du Mons Claudianus,”CdE 61 (1986) 271–86, 5 figg. 8. “Inscription inédite d’un ἐργοδότης dans une carrière du Mons Claudianus,” in Itinéraires d’Égypte. Mélanges offerts au père Maurice Martin s.j. = Bibliothèque d’étude 107 (Cairo 1992) 73–88, 1 figure. 9. “The amount of wages paid to the quarry-workers at Mons Claudianus,” JRS 86 (1996) 139–45, 1 table and pl. II. 10. “Deux ostraca du Mons Claudianus: O.Bahria 20 et 21,” CdE 72 (1997) 112–18, 1 figure. 11. “L’organigramme du personnel d’une carrière impériale d’après un ostracon du Mons Claudianus,” Chiron 35 (2005) 309–53, table, 1 plan, 1 pl. 12. “Une dédicace à Zeus Hèlios Grand Sarapis honorant un desector sur un ostracon du Mons Claudianus,” in Christel Freu, Sylvain Janniard, and Arthur Ripoll (eds.), Libera curiositas. Mélanges d’histoire romaine et d’antiquité tardive offerts à Jean-Michel Carrié (Turnhout 2016) 17–21, 1 figure. 13. “Un ostracon inédit du désert Oriental et la provenance de O.Amst. 9,” in Proceedings of the 20th International Congress of Papyrologists (Copenhagen 1994) 229–30 and pl. 11. 14. “Kainè, ville nouvelle: une expérience de regroupement familial au IIe s. è.chr.,” in Olaf E. Kaper (ed.), Life on the Fringe. Living in the Southern Egyptian Deserts during the Roman and EarlyByzantine Periods. Proceedings of a Colloquium Held on the Occasion of the 25th Anniversary of the Netherlands Institute for Archaeology and Arabic Studies in Cairo 9–12 December 1996 = CNWS Publications 71 (Leiden 1998) 87–94. 15. “Le système routier du désert Oriental égyptien sous le Haut-Empire à la lumière des ostraca trouvés en fouille,” in Jérôme France and Jocelyne Nelis-Clément (eds.), La statio. Archéologie d’un lieu de pouvoir dans l’empire romain (Bordeaux 2014) 247–78.  16. “Recueil d’irrégularités dans la transmission du courrier officiel,” in Francisca A. J. Hoogendijk and Brian P. Muhs (eds.), Sixty-Five Papyrological Texts Presented to Klaas A. Worp on the Occasion of his 65th Birthday = Papyrologica Lugduno-Batava XXXIII (Leiden 2008) 317–23, 1 figure.

xx 17. “Le livre de poste de Turbo, curateur du praesidium de Xèron Pelagos,” in Anne Kolb (ed.), Viae Romanae/Roman Roads. New Evidence, New Perspectives (Berlin and Boston 2019) 67–106, 11 figg. 18. “Hommes et dieux en réseau : bilan papyrologique du programme ‘désert Oriental’,” CRAI 2013: 405–42, 12 figg. 19. “Un reçu de rations militaires contre paiement des publica,” in Katja Lembke, Martina Minas-Nerpel, and Stefan Pfeiffer (eds.), Tradition and Transformation: Egypt under Roman Rule. Proceedings of the International Conference, Hildesheim, Roemer- and Pelizaeus-Museum, 3–6 July 2008 = Culture and History of the Ancient Near East 41 (Leiden 2010) 37–51, 3 figg. 20. “La ration mensuelle d’un cavalier et de son cheval d’après un ostracon du praesidium de Dios (désert Oriental d’Égypte),” in Michel Reddé (ed.), De l’or pour les braves! Soldes, armées et circulation monétaire dans le monde romain = Scripta antiqua 69 (Bordeaux 2014) 71–90, 3 tables, 2 figg. 21. “Un type méconnu de document administratif militaire: la demande de versement de frumentum praeteritum (O. Claud. inv. 7235 et Ch.L.A. XVIII 662),” in Tomasz Derda, Adam Łajtar and Jakub Urbanik (eds.), in cooperation with Andrzej Mirończuk and Tomasz Ochała, Proceedings of the 27th International Congress of Papyrology, Warsaw, 29 July – 3 August 2013. Volume II. Subliterary Papyri, Documentary Papyri, Scribal Practices, Linguistic Matters = The Journal of Juristic Papyrology, Supplements Volume XXVIII (Warsaw 2016) 931–41, 2 figg. 22. “Conductor praesidii,” in Rodney Ast, Hélène Cuvigny, Todd M. Hickey, and Julia Lougovaya (eds.), Papyrological Texts in Honor of Roger S. Bagnall = American Studies in Papyrology 53 (Durham, North Carolina 2013) 67–74, 2 figg. 23. “Quintana, la femme métamorphosée en taxe,” in H. Cuvigny (ed.), La route de Myos Hormos, 2nd ed. (Cairo 2006) 698–93. 24. “Femmes tournantes: remarques sur la prostitution dans les garnisons romaines du désert de Bérénice,” ZPE 172 (2010) 159–66, 2 tables. 25. “Serapias balance ses porcs, ou quand la réalité rejoint la fiction théâtrale,” in A. Ricciardetto, N. Carlig, G. Nocchi Macedo, and M. de Haro Sanchez (eds.), Le médecin et le livre. Hommages à Marie-Hélène Marganne (Lecce forthcoming) 163–70. 26. “Des Kinaidokolpites dans un ostracon grec du désert Oriental (Égypte),” Topoi 6 (1996) 697–720, 2 figg. (with Christian Robin). 27. “Papyrological Evidence on ‘Barbarians’ in the Egyptian Eastern Desert,” in Jitse H. F. Dijkstra and Greg Fisher (eds.), Inside and Out. Interactions between Rome and the Peoples on the Arabian and Egyptian Frontiers in Late Antiquity = Late Antique History and Religion 8 (Leuven 2014) 165–98, 2 figg. 28. “Poste publique, renseignement militaire et citernes à sec: les lettres de Diourdanos à Archibios, curator Claudiani,” Chiron 49 (2019) 271–97, 11 figg. 29. “Le crépuscule d’un dieu : le déclin du culte de Pan dans le désert Oriental,” BIFAO 97 (1997) 139–47, 2 figg. 30. “Un soldat de la cohors I Lusitanorum à Didymoi: du nouveau sur l’inscription I.Kanaïs 59 bis,” BIFAO 101 (2001) 153–57, 1 figure. 31. “The shrine in the praesidium of Dios (Eastern Desert of Egypt): graffiti and oracles in context,” Chiron 40 (2010) 245–99, 2 tables, 29 figg.

xxi 32. “Le préfet d’Égypte démobilise des hommes âgés et impose un “sceau”(tatouage?) prophylactique,” Chiron 44 (2014) 325–339, 2 figg. 33. “‘Le blé pour les Juifs’ (O.Ka.La. inv. 228),” in Gaëlle Tallet and Christiane Zivie-Coche (eds.), Le myrte et la rose. Mélanges offerts à Françoise Dunand par ses élèves, collègues et amis, Tome 1 = CENiM 9 (Montpellier 2014) 9–14, 1 pl. 34. “La plus ancienne représentation de Moïse, dessinée par un juif vers 100 è. chr.,” CRAI 2014/1: 339– 51, 4 figg. 35.“Plèrôma dans l’identification des soldats de marine,” ZPE 110 (1996) 169–73. 36. “Remarques sur l’emploi de ἴδιος dans le praescriptum épistolaire,” BIFAO 102 (2002) 143–53, 1 figure. 37. “Πέμπειν/ἀγοράζειν τῆς τιμῆς dans l’épistolographie grecque d’Égypte,” CdE 80 (2005) 270–76. 38. “Les noms du chou dans les ostraca grecs du désert Oriental d’Égypte: κράμβη, κραμβίον, καυλίον,” BIFAO 107 (2007) 89–96, 1 table, 3 figg. 39. “Χίλωμα = musette,” ZPE 166 (2008) 195–98. 40. “‘Quand Hèroïs aura accouché...’ ἐάν = ὅταν dans l’expression de l’éventuel,” BIFAO 112 (2012) 97– 99. 41. “Les ostraca sont-ils solubles dans l’histoire?” Chiron 48 (2018) 193–217, 4 figg.

xxii

Rome in Egypt’s Eastern Desert

The Eastern Desert © Louis Manière (ERC project “Desert Networks,” CNRS)

INTRODUCTION AND SURVEY

Introduction Roger S. Bagnall Although the study of Graeco-Roman Egypt as a whole has made enormous progress, in both archaeology and documentary studies, over the last half century, it is no exaggeration to say that it is the desert zones for which our knowledge has been most revolutionized. It may be difficult for someone beginning studies today to realize how little we knew of the Eastern Desert in 1970. A number of early travelers visited various parts of the region, and Steven Sidebotham has paid tribute to the quality of J. Gardner Wilkinson’s sketch plan of Berenike from 1826.1 These early travelers and further visitors in the first half of the twentieth century also copied a number of inscriptions, and in the course of work for the Tabula Imperii Romani sheet “Coptos” David Meredith carried out a certain amount of survey as well as epigraphical work in the late 1940s and early 1950s, some in collaboration with Leon Tregenza. This resulted in a considerable number of articles. Their work was in the very recent past when André Bernand began in 1957 the long process of producing a corpus of inscriptions for the Eastern Desert, which he focused on the figure of the god Pan.2 But it was only in the 1970s that the volumes of this corpus appeared, providing a more systematic treatment of the epigraphy and, to some extent, of the travelers (I.KoKo., 1972; I.Kanaïs, 1972; I.Pan, 1977). Although Bernand traveled in the Eastern Desert and made some discoveries of new material, he did not carry out any systematic program of travel, let alone survey, in this inhospitable region; for example, his description of Porphyrites and Mons Claudianus depends entirely on others.3 That ostraca also might play a significant role in understanding the area began to become clear from the article of Octave Guéraud on ostraca from what is now known to have been the site of Persou, published in 1942, as well as a handful published by David Meredith in 1956.4 Also of importance, but found at Koptos rather than in the desert proper, was the substantial archive of Nicanor and his family, proprietors of a transportation enterprise based in this gateway city during the first century of Roman 1. Sidebotham 2011: 16. 2. Fortunately, his apparent initial idea of producing a corpus simply of all of the inscriptions mentioning Pan gave way to a broader vision during the nearly two decades devoted to the project (cf. I.KoKo., pp. xiii–xiv). 3. He says explicitly that he did not visit Mons Claudianus: I.Pan, p. 78; his information on Porphyrites is also all second-hand. 4. “The Myos Hormos road: inscriptions and ostraca,” CdE 31 (1956) 356–62; texts reprinted as SB VI 9165–1966.

1

2

Rome in Egypt’s Eastern Desert

rule, amounting to almost 100 texts (reedited in O.Petrie Mus.; Trismegistos Archives 154). But at the beginning of the 1970s little could have prepared anyone for the sheer volume and interest of the written documentation that would begin to emerge from the desert. It is no doubt for this reason that when the antiquities market began to produce groups of ostraca in the early 1970s none of those of us involved in publishing them could see any compelling reason to doubt the attributions of provenance supplied by dealers, variously connecting them with Contrapollonopolis Magna and Laton polis.5 If we now know that O.Amst. 9 instead comes from Maximianon (see Chapter 13), it has been impossible to establish a secure provenance for the rest of the texts among any of the other desert forts excavated to date, or probably their near neighbors; and Willy Clarysse’s attempt to place them in the neighborhood of Thebes does not seem persuasive.6 They may not, of course, all come from the same source, but a desert origin seems certain, quite likely a praesidium that has not been excavated—indeed, very possibly one of which the dump was destroyed before scientific excavations in the area began. It is from two years after the appearance of the Florida and Amsterdam ostraca, and a year after the publication of the final volume of André Bernand’s epigraphical corpus, that one may date the beginning of properly scientific investigation of the region, with the start of the University of Chicago excavations at Qusayr al-Qadim in 1978 and 1980.7 The identification of this site with an ancient place-name was still uncertain at the time; it was not until after new evidence was discovered that it could be shown definitively that this was the ancient Myos Hormos.8 But the excavations, even though not continued after their second season, did show the potential for excavation in an Eastern Desert site to bring wellpreserved evidence for the ancient and medieval activity in the region. The years 1985–1987 brought three decisive, if very different, impulses to the enormous flourishing of studies of the Eastern Desert that the last third of a century has seen. The first of these was the publication in 1985 of the extraordinary (even if fragmentary) document usually called the “Muziris papyrus,” with on one side a contract connected with a trading voyage between India and Egypt, on the other an inventory of the cargo of the ship that made this trip.9 This rich source has given rise to a long series of articles and books on the trade between the Mediterranean and India, mostly recently Federico de Romanis’s just-published monograph on the papyrus.10 The second was the publication the following year of Steven Sidebotham’s monograph on Roman economic policy in the Red Sea region, a revised version of his 1981 dissertation.11 He followed this book up with fieldwork aimed at pursuing this interest, notably in excavations at Abu Sha’ar in 1990–1993.12 And then, eight years after the publication of the book, Sidebotham was able to begin survey and excavation at the key port site of Berenike, thus inaugurating a fruitful project, with many ramifications, that still continues. The third initiative, arguably the most consequential, and certainly at the root of the present volume, was the beginning in 1987 of the excavations at Mons Claudianus, by an international team under 5. A list of the various groups is provided by Georges Nachtergael in O.Hombert II, pp. 9–10. The Hombert group was, he says, purchased in Luxor in 1953 but acquired by Hombert only in 1981. Nachtergael also (p. 10, n. 3) lists several small groups of unprovenanced ostraca published in scattered articles. O.Florida and O.Amsterdam both appeared in 1976. 6. Clarysse 1984. Nachtergael (O.Hombert II, pp. 11–13) was inclined as late as 2003 to retain the attribution to Contrapollonopolis Magna. 7. Whitcomb and Johnson 1979, 1982. 8. Peacock 1993 and Bülow-Jacobsen, Cuvigny, and Fournet 1994 were the decisive contributions. Cf. Peacock and Blue 2006a: 4–5 for the debate. 9. Harrauer and Sijpesteijn 1985. 10. F. de Romanis, The Indo-Roman Pepper Trade and the Muziris Papyrus (Oxford 2020). 11. Sidebotham 1986. 12. The modest yield of texts from this work was published in R. S. Bagnall and J. A. Sheridan, “Greek and Latin Documents from ‘Abu Sha’ar, 1990–1991,” JARCE 31 (1994) 157–66 and in Bagnall and Sheridan 1994.

Introduction and Survey

3

the leadership of Jean Bingen; this group included Adam Bülow-Jacobsen and Hélène Cuvigny, and from their experience at Mons Claudianus in 1987–1993 resulted the remarkable series of subsequent Eastern Desert excavations led by Cuvigny that are, along with the documents from Mons Claudianus itself, the foundations of this volume. Those who knew that remarkable man Jean Bingen will readily recognize his spirit alive throughout its pages, in the breadth of interests and knowledge, the endless curiosity, and the generous collegiality that mark the enterprise throughout. The series of excavations led by Cuvigny were not the only offspring of Bingen’s initiative at Mons Claudianus. David Peacock with Valerie Maxfield and Lucy Blue worked first at Porphyrites (1994–1998) and then at Myos Hormos (1999–2003), with a range of important finds. The archaeological results were published very promptly and figure in the present volume in various areas. The documents—mostly ostraca, as at all of the Eastern Desert sites—still await publication, although some have seen preliminary publication or have been communicated privately. Even from this brief description of the work on the Eastern Desert in the last thirty years it is evident that the Romans brought to it a distinctive approach. Despite the Ptolemies’ early interest in African elephants (said to be the reason for the founding of Berenike) and the supposed opening up of Indian Ocean trade later in the Ptolemaic period, the Eastern Desert before the Romans was predominantly a zone of resource extraction—gold, stone, and gems. Under Roman rule, the desert kept that vocation, but with a dramatic expansion of activity in the first two centuries as imperial building projects took advantage of the distinctive stone of the quarries in the northern part of the desert. And to that extraction they added a large-scale trade through the Red Sea corridor to East Africa, South Arabia, and above all India, carried out through the ports of Myos Hormos and then, for the most part, Berenike, which was better suited to the massive cargo ships that came into use, as Federico de Romanis has argued. This book has little to say about that trade, in part because it reflects a series of excavations (listed in Chapter 41) of sites in the desert, not of the ports, which have been the work of other teams. But this is in itself of some interest. One might have expected extensive traces of the trade passing through the way-stations of the desert, particularly the Desert of Berenike, to be present in the documents found in those little forts and their dumps. But they are not to be found. It is almost as if the traffic on the desert roads was hermetically sealed from its environment. In some sense, this was actually the case, because the trade goods were transported under seal; customs examination at either end of the route (Berenike and Koptos) required them not to be tampered with in between. The desert stations had to provide security for overnight stays and at least some part of the water required by men and animals, but they played no role in the trade itself. Cuvigny’s work (particularly in Chapter 1) has greatly clarified the Romans’ conception of the Eastern Desert, showing that it needs to be thought of as two great and distinct zones, that north of the road from Koptos to Myos Hormos, dominated by the quarrying of Mons Claudianus, Porphyrites, and some smaller sites; and that to the south, called the Desert of Berenike in administrative parlance and, while not without mineral extraction,13 mainly devoted in this period to the road network that supported caravan transport between the Red Sea ports and the Nile valley. The dividing line and axis of the northern zone, the road between Koptos and Myos Hormos, was dealt with in detail in the massive collective study La route de Myos Hormos, published in 2003 and in a second edition 2006 (Cuvigny [ed.] 2006). Topics dealt with there come back in the present volume where new evidence has changed views (see especially Chapter 23), but much of this book is dedicated to the road to Berenike and the southern zone. 13. The mineral resources of the entire desert are given their due in Sidebotham, Hense, and Nouwens 2008, based on extensive survey work carried out over two decades.

4

Rome in Egypt’s Eastern Desert

In broad terms, one could say that this volume is devoted not so much to the “what” of the Roman exploitation of the Eastern Desert, as to the “how.” How did the Romans conceive of this new domain and organize the space? How did they organize their workforce, keep them fed, and pay them? How did they manage transportation and communication across these long desert stretches and between the inhabited outposts? How did they provide security for long exposed lines of communication in an alien environment better known to indigenous peoples? How did they provide some semblance of normal life and maintain morale for the quarry workers and the garrisons of the forts? How did they administer this immensely complicated set of related enterprises? How did they extract tax revenues from this zone? All of these questions perhaps come down to a larger question of what we mean by “Rome” in talking about Rome in the Eastern Desert of Egypt. What is Roman in the way the Romans managed this area? That the evidence from the praesidia tells us little about the trade does not, however, mean that it does not have connections with the documents found in the ports. Most of the texts from Berenike in fact also belong to the “how” of the subject. They concern to a large extent the management of the customs operations that encapsulated the goods at one end of the road in Koptos, enabling them to pass through the gate at Berenike; and they tell us something (even if not enough) about how Roman Berenike got its water supply. The Egyptians who provided transportation services across the desert, whether for export goods, supplies for local consumption, or water, were organized much as for the intervening stations, and some of the same individuals appear in both Berenike ostraca and those from forts on the road from Koptos to Berenike. This book is the result of several decades of study, but it is a gathering of articles rather than a systematic synthesis. They are framed by a study of the topography and place-names of the Eastern Desert (Chapter 1) and a concluding reflection on the enterprise of writing history from these ostraca (Chapter 41).14 The topical structure of the book does reflect accurately the unsystematic nature of the material found. Some of the documents are epigraphical, showing what individuals or officials wanted to commemorate and wished others to think. Most of it is ostraca, that is, intentionally ephemeral texts. Chapter 41 stresses the virtually complete disappearance of papyrus from the forts and quarries, and with it everything that the ancient inhabitants of these sites thought was worth keeping and taking with them back to the valley, where it has not survived. We have to work with the garbage.15 There was, of course, a pre-Roman Eastern Desert, now beginning to be better known not only from Steven Sidebotham’s survey work but from the excavations carried on by Cuvigny’s successors in her field project, as listed in Chapter 41. The first volume publishing their results has just appeared, presenting the Ptolemaic gold-mining site of Samut North.16 This extractive focus was central to the pharaonic and Ptolemaic approach to the region, and in Chapter 29 Cuvigny shows how it was closely connected to the cult of the god Pan, thanked for his guidance of those who discovered valuable resources. Sarapis and Isis were far more central to the religious life of the Roman period (see Chapters 12 and 18 and Part VII), and Pan lost importance.17 The Romanization of religious life is visible in the operations of an oracle in the praesidium of Dios (Chapter 31): even if such oracles were a long-standing Egyptian practice, the close parallels that Cuvigny adduces to oracular practice in Roman Asia Minor show that the oracle’s operation was substantially fashioned to suit a different population of desert travelers. 14. Thoughts on methodology can also be found particularly in Chapters 15 and 18, as well as here and there throughout. 15. This is a recurrent theme in papyrology, to be sure. One has only to think of the fact that the Oxyrhynchos papyri come from dumps. Even in habitation areas, finds of papyri often come from dumped or discarded material. On ostraca in this respect, cf. my reflections in Everyday Writing in the Graeco-Roman East (Berkeley 2011) 117–22. 16. B. Redon and T. Faucher, Samut Nord. L’exploitation de l’or du désert Oriental à l’époque ptolémaïque (Cairo 2020). 17. Isis was the most important deity at Berenike, with her role in safe navigation. See Ast and Bagnall 2015 and further inscriptions to be published by Rodney Ast.

Introduction and Survey

5

The entire approach to the road system and structuring of the southern part of the desert by the Romans shows how great were the changes under imperial rule (Chapters 1, 15, 26). The usefulness of the ports for different sizes of ships was the key driver of the development of the roads; the trade-off between shortness of the route, to minimize land transportation costs, and the convenience of the port’s location for ships given the difficulty of sailing against the northerly winds in the Red Sea led to the increased emphasis on the Koptos to Berenike road and the development of stations along it. The world of the desert forts and quarries bears an unmistakably Roman imperial stamp, which emerges in many ways throughout the book. The military occupies a central role, as one might expect in a zone of this kind, located on the edge of a province and bordering a world not controlled by Rome. It is no surprise that what we see of the army in the Eastern Desert is similar to what we find elsewhere in Egypt and in the empire at large; parallels from as far away as Vindolanda in Britain lie ready to hand. The ostraca have, however, provided new evidence for questions previously impossible to answer with any confidence, especially in the area of rations and supply; from Chapters 26–28 we learn not only how much soldiers received but the mechanics of the processes by which they got their supplies, especially when detached from the home bases of their units, as all of the soldiers in the garrisons of the forts of the desert were. Because of the wide circulation of information inside the military structure, we also gain very valuable information about the disposition of Roman forces in the Dodekaschoinos and the way in which the concept of a river boundary functioned there (Chapter 28). The situation was not static, and the full study of relations between the Roman government and the Barbarians (as they called them) in Chapter 27 shows that the increased tension from the 70s on, with violent clashes in the early second century, is almost certainly not just a matter of increased aggression by the indigenous peoples, but their response to the changes brought by the Romans: in the Romans taking control of some mineral resources that the local population traditionally exploited themselves; in increased competition for limited resources of water and vegetation; in the imperial attempt to crack down on smuggling (Chapter 26); and perhaps most of all simply in creating such a tempting array of targets in the caravans loaded with extremely valuable commodities. And yet it is equally clear from Chapter 27 and the forthcoming O.Blemmyes that the relationship between the Roman authorities and the unincorporated population was not simply conflictual; we have here an extraordinary chance to see the workings at the everyday level, and at different periods, of Roman interaction with the peoples on the frontiers of the empire, unfiltered by the biases of the literary sources. Even more than the military, the administrative sphere provides us with a remarkable sense of the Roman balancing of standardization and pragmatic adaptation. The demonstration (Chapter 9) that the compensation of Egyptian quarry workers was essentially the same as those in European provinces, despite Egypt’s having constituted a separate currency zone with its own coinage, is truly remarkable. And the Egyptian quarry specialists were well paid and supplied (Chapter 11), as they presumably needed to be in order to attract qualified workers and keep them engaged. But standardized pay is only one side of the picture; the other is a concerted attempt to compensate for the isolation and harsh conditions of desert life by creating a settlement for the workers’ families at the new town of Kaine in the valley (Chapter 14); there the monthly rations could be turned into home-made bread by the women of the worker’s family, and the quartermaster of the month would carry out innumerable errands for the other workers so that their income would be handled properly and their bills settled. The alternative, settling the families by the quarries, as done elsewhere in the empire, was clearly impractical, given the difficulties of supplying these places with food and water. We take away the impression that the administration was acutely aware of the need to attract and retain skilled labor, leaving the grunt work to be done by the familia—a body the character of which remains somewhat enigmatic (Chapter 11). The native work-

6

Rome in Egypt’s Eastern Desert

men may not have been of any particularly elevated social status, but their skills gave them an economic advantage of which the state was conscious. The soldiers in the praesidia also had needs. These dominate the corpus of letters on ostraca from the excavations of those sites. They awaited eagerly the arrival of supply caravans from the valley, but also the more informal private provision of what official supplies did not provide, particularly fresh vegetables grown at those of the stations with adequate water supplies, a theme which recurs constantly (Chapter 22; for cabbage-family plants, see particularly Chapter 38). Non-military people also lived in the forts, primarily in order to help meet these needs. They also supplied commercial sex to the soldiers, through a system of the mainly enslaved prostitutes hired on a monthly basis as a common (“rotating”) service for the soldiers in a fort (Chapters 23–25). From Cuvigny’s acute analysis of the letters concerning prostitution we get both a clear sense of the business side of the system and insights into the complex psychology of prostitution in such an environment; it was not a business like any other, and particularly from Chapter 25 we get a strong feeling for the reactions and agency of the women, who had to deal with harassment and violence, even rape where they did not do what they were told. It was an ugly business, and Part V does not make pleasant reading. The Roman administration seems to have been content with letting the private sector fill many of the needs of both soldiers and quarry workers, in keeping with the general character of their administrative habits. But there was still need for an administration, and unlike in the Nile valley, it faced the logistical challenges posed by land travel in a difficult landscape to its ability to communicate and command over such an expanse. In particular, the mix of military and civilian activities was distinctive to the desert and required its own structures. Two aspects of this subject gain significant traction from the ostraca: the structure and composition of the senior administration, particularly the prefect of the Mons Berenicidis, often at the same time commander of a military unit and yet of procuratorial status (Chapters 2–6), and the system of postal communication (Chapters 16–17). These systems evolved over time, and it can hardly be claimed that even now everything is clear. But the picture is far more detailed than it was. The original articles that underlie the Chapters of this book were written at various times over an extended period, as the list of original versions (pp. xvii–xx) shows. As new evidence came to light, Cuvigny’s views evolved with it. In the present book, the treatments of many questions have been updated to take account of new evidence, but at the same time Cuvigny has been careful to document the development of her views (a good example can be found in Chapter 24). The reader thus gets a candid sense of how scholarship progresses in a dynamic setting with a constant flow of new information. The duty to tell readers such changes is one often stressed by Louis Robert. This is not the only respect in which many passages of this book remind one of Robert. His was a style of philology in its broadest and truest sense, in which all types of text are taken into account with care and on their own terms, and in which the texts are thoroughly confronted with the evidence of archaeology, landscape, and physical realities. This is what the reader will find here.

1 A survey of place-names in the Egyptian Eastern Desert during the principate according to the ostraca and the inscriptions The purpose of this chapter is to take stock of the progress made on the toponymy of the Eastern Egyptian Desert through the ostraca found in the excavations of Roman sites in which I took part between 1987 and 2012 (Fig. 1).1 The ostraca, published and unpublished, to which I will refer come from the quarry sites of Mons Claudianus and Domitiane/Kaine Latomia (Umm Balad), two forts (praesidia) on the road from Koptos to Myos Hormos (Maximianon and Krokodilo), and three praesidia on the road from Koptos to Berenike (Didymoi, Dios, and Xeron). I occasionally take into account the ostraca of Porphyrites and Myos Hormos.2 I also refer to names read on Greek ostraca found recently in Biʾr Samut, a fort along the road from Apollonos polis (Edfu) to Berenike; Biʾr Samut was founded under Ptolemy II or III and abandoned under Ptolemy IV.3

1. This article has developed from a lecture presented on 13 April 2013 as part of the interdisciplinary EPHE project “Lieux d’Égypte ou la toponymie égyptienne des pharaons aux Arabes” (2012–2014). I have been fortunate to benefit from a critical reading of my manuscript by Herbert Verreth, whom I thank for helping me with his many comments, and spotting many small blunders, omissions, and inconsistencies; his judicious questions also helped me to clarify my thinking and deepen my reflection on some points. Naim Vanthieghem was kind enough to suggest to me a system that is both consistent and simple for the transcription of modern Arabic place-names. Unless otherwise noted, all photographs were taken by Adam BülowJacobsen, whom I also thank, as well as Mathilde Bru, for their help with the translation into English. 2. The ostraca are designated by a publication or an inventory number, preceded according to the provenance, by the abbreviations O.Claud., O.KaLa., O.Krok., O.Max., O.Did., O.Dios, O.Xer., O.Porph., O.MyHo. I thank Wilfried Van Rengen for allowing me to quote ostraca belonging to the two last corpora. 3. Excavations 2014–2016 funded by IFAO and MAE in the program MAFDO now led by Bérangère Redon and Thomas Faucher.

7

8

Rome in Egypt’s Eastern Desert

Mediterranée Alexandreia Pelousion Klysma

Memphis

SINAI

Soknopaiou Nèsos

Oxyrhynchos Oasis mikra Akôris Alabastrinè (Baharîya) Antinoou polis Hermou polis (Farafra) Lykôn polis Ptolemaïs

(Da Oasis megalè khl a)

Myos Hormos

(Kh

ârg

a)

Koptos

Syènè 0

Berenikè

50 km

Figure 1 (above and right). The Eastern Desert in Roman times. © J.-P. Brun wâdî Belih

Mons Porphyrites

Abû Sha’ar al-Qiblî Abû Sha’ar

Domitianè/Kainè Latomia

Bâdiya Abû Qurayya Dayr al-Atrâsh Qattâr Bâb al-Mukhayniq Al-Sâqiya Wadi Safaga Abû Zawâl Al-Hâyita Tal‘t al-Zarqa Mons Claudianus Tiberianè Al-‘Aras Jidâmi Samna Quway Kainè Qarya Krokodilô Bi’r Sayyâla (Qina) Al-Matûla Persou Qusûr (Qusayr al-Qadîm) Dawwî al-Banât Al-Hamrâ’

Ptolemaïs

Myos Hormos

Maximianon

Aphroditès orous

na

Al-Kanâ’is

a dri

Bi’r Bayza

Barrâmîya

ge

Dios

Ha

Kompasi

Via

Paneion du wâdî Minayh

(Isna)

Apollônos polis hè megalè

Bi’r al-Hammâmât

Didymoi

Latôn polis

ou rR

Dios polis

Phoinikôn

Me

Koptos

Xèron Pelagos

Mersa Nakari

(

Syènè 0

Berenikè

50 km

A survey of place-names in the Egyptian Eastern Desert wâdî Belih

Mons Porphyrites

9

Abû Sha’ar al-Qiblî Abû Sha’ar

Domitianè/Kainè Latomia

Bâdiya Abû Qurayya Dayr al-Atrâsh Qattâr Bâb al-Mukhayniq Al-Sâqiya Wadi Safaga Abû Zawâl Al-Hâyita Tal‘t al-Zarqa Mons Claudianus Tiberianè Al-‘Aras Jidâmi Samna Quway Kainè Qarya Krokodilô Bi’r Sayyâla (Qina) Al-Matûla Persou Qusûr (Qusayr al-Qadîm) Dawwî al-Banât Al-Hamrâ’

Ptolemaïs

Myos Hormos

Phoinikôn

Maximianon

Bi’r al-Hammâmât

Didymoi

an a

(Idfu)

dri

Barrâmîya Al-Kanâ’is

e

Dios

Bi’r Bayza

Apollônos polis hè megalè

ug

Kompasi

(Isna)

Ha

Latôn polis

Aphroditès orous

Via

Paneion du wâdî Minayh

Ro

Dios polis

r Me

Koptos

Xèron Pelagos

Bi’r Samût

Mersa Nakari

Mons Smaragdus

Phalakron Apollônos Hydreuma Cabalsi ?

Wâdî Lahma Kainon Hydreuma

0

50 km

Syènè

Jean-Pierre Brun

Agglomération antique

Fort antique

Station antique non fortifiée

Berenikè Piste antique

Piste hypothétique

These documents, many of which are still unpublished, have yielded a small corpus of new or already known toponyms, but some of the latter were already known in a form distorted by the Medieval manuscript tradition. The main classical sources on place-names in the Eastern Desert up to now were: – The description of the road from Koptos to Berenike by Pliny the Elder, informed by negotiatores, in the state in which it was found in AD 50, i.e., before the wells (hydreumata) were fortified into praesidia under Vespasian (Nat. 6.102–3); – The list of stages on this route in three itineraries taken from the manuscript tradition: Antonine Itinerary (172–73 ed. Parthey and Pinder), Peutinger Table, and the Anonymous of Ravenna (2.7.4 ed. Schnetz). – Ptolemy’s Geography for the coast of the Red Sea (4.5.14–15) and the desert (4.5.27).4

4. I use the reference system of the latest edition of the Geography, to which Germaine Aujac kindly drew my attention: Klaudios Ptolemaios Handbuch der Geographie, Basel 2006.

10

Rome in Egypt’s Eastern Desert

I. The administrative districts of the Eastern Desert under the early Roman empire Our ostraca come from a geographical area which, in the Roman period, seems to have consisted of two distinct administrative sectors with different purposes. The more southerly is the better defined and the one for which the administrative structure is the most stable and best known. The Romans called it Mons Berenicidis or Mons Berenices, which does not mean “Mountain,” but “Desert of Berenike,” the Latin mons being, through the Greek ὄρος, the loan translation of an Egyptian word, ḏw, which means both desert and mountain.5 The Desert of Berenike, named after its most active port, was crossed by the roads to Myos Hormos and Berenike, both of which departed from the Nile valley at Koptos. They represented a terrestrial segment of one of the main trade routes with the Red Sea world. From Vespasian’s time, they were equipped with fortified wells, each headed by a curator praesidii. The curators were placed under the direct authority of the prefect of Berenike, who was an imperial procurator (Chapters 3 and 6). These territorial prefects, whose prosopography has been enriched by our ostraca, often added to their procuratorship a military command, the prefecture of the cavalry wing stationed in Koptos. In the ostraca of the aforementioned sites, the road to Myos Hormos seems to mark the northern limit of the Berenike Desert. Nevertheless, there is a reference to a prefect of Berenike north of this road: this is the dedication of the Paneion of Ophiates, a Roman granite quarry in the Wadi Umm Wikala, a tributary of the Wadi Samna (I.Pan 51);6 it dates from year 40 of Augustus’s reign (AD 11) and specifies the name of the prefect of Berenike in power, Publius Iuventius Rufus, who combined this function with that of archimetallarches, that is to say, director in chief of mines and quarries. In AD 11, Mons Claudianus and Porphyrites had yet to open, and the mineral resources specifically mentioned (and identified) in the prefect’s titulature, Smaragdos and Bazion,7 lay south of the Wadi Samna. The inscription, therefore, contains the most northerly mention of a prefect of Berenike: in fact, there is never any mention of this official in the ostraca of the granite and porphyry quarries that the Romans opened later: Porphyrites, Tiberiane, Mons Claudianus, Domitiane/Kaine Latomia. Was this northern area of the Eastern Desert even an administrative entity? This is not clear, as we shall see. The exploitation of Mons Claudianus granite and the porphyry of the Porphyrites caused the Romans to reorganize the road system in this region, which had been explored before the Romans. Gold deposits had been exploited under the New Kingdom and under the Ptolemies in the area of what would become Mons Claudianus, especially between the current Qena-Safaga road and Myos Hormos road;8 in this segment of the Eastern Desert there are also the amethyst mines of Abu Diyayba, exploited under Ptolemy VI and still active at the beginning of the Empire. Prior to the founding of Kaine (Qena), which served as a terminus for the Claudianus and Porphyrites roads, the mining sites north of Myos Hormos road must have been administered and provisioned from Koptos, and they belonged presumably to what was called in Egyptian ḏw Gbtyw, the Desert of Koptos. This is the reason why, in my opinion, at the beginning of the Empire, and before the founding of Kaine, the Berenike Desert included Ophiates, an early Roman granite quarry north of what later appears as the northern border of Mons Berenicidis. 5. On the misinterpretations that were driven by the misunderstanding of mons and of the suffixed form Berenicis, see infra pp. 69 f. 6. On this metallon, see Sidebotham, Barnard, Harrell, and Tomber 2001. 7. On both sites, see pp. 16 f. 8. These gold mining sites that often betray activity in the Ptolemaic period have never been explored. They are conveniently catalogued and described by Klemm and Klemm 2013, in their section entitled “Middle central group,” pp. 68–146. Nowadays, they are threatened by a project of intensive mining exploitation which will concentrate on the so-called Gold Triangle, that is to say the part of the Eastern Desert comprised between the Qena–Safaga road and the Quft–Qusayr road.

A survey of place-names in the Egyptian Eastern Desert

11

The opening of quarries at Porphyrites and Claudianus caused new logistical problems: it was not a question of taxing and transporting valuable products, but of extracting multi-ton monoliths and transporting them over a hundred kilometers to the Nile. The development of a closer embarkation site was necessary, and the area, served by its own road system, had no reason to be commanded by a prefect of Berenike exercising his authority from Koptos. It was no longer about placing “all the mines and quarries of Egypt” under his authority. The ostraca from the four metalla that have been excavated, Claudianus, Porphyrites, and their satellites (Tiberiane and Domitiane/Kaine Latomia), show that these quarries functioned as a network; the staff and labor force were moved, as needed, from one site to the other. But did this region have a name? A series of ostraca from Mons Claudianus, receipts for advances to the familia,9 allow us to suggest a hypothesis. The familia is one of two major categories of labor in the quarries of the area. The other consists of the pagani, who were highly qualified stone-carvers and blacksmiths of indigenous origin and free status. Of the two categories, the familia is the more enigmatic. This is, most likely, an imperial familia, so in principle slaves of the emperor, but some of these individuals, who have patronymics or gentilicia, cannot have had servile status. Their variegated onomastics often denote an origin outside Egypt. In any case, the familia was employed in tasks that demanded more strength than technical knowledge. Some of these familiares lived on credit and received advances of food, for which they were made to sign receipts, on which appeared their administrative affiliation with a numerus and an arithmos. These two words, the first Latin, the second Greek, are synonymous in principle (they mean “number”), but in this case, the arithmos represents a subdivision of a numerus. In the ostraca from Mons Claudianus, members of the familia almost all belong to the numerus of Porphyrites and the arithmos of Claudianus; but a few, registered in the arithmos of Tiberiane, worked in this satellite of Claudianus. Two receipts for advances are exceptional: they are issued by individuals belonging to another numerus, that of Alabastron:10 they were probably registered in the rosters of the alabaster quarries of the Hermopolite, which calls to mind the existence at Hermou polis of an “office of the accountants of Porphyrites and other metalla.”11 If we project the administrative structure of the familia on the map of the region, it seems that Porphyrites was not only the name of the metallon where porphyry was extracted, but referred also to the entire area that was penetrated by the two roads coming from Kaine. This region of Porphyrites (which remains a hypothesis) did not have the same administrative structure as the Berenike Desert. It was not under the authority of a territorial prefect who would have been the counterpart of the prefect of Berenike. However, in some late ostraca from Claudianus, there appears the word ἔπαρχος, “prefect.” It is, unfortunately, not clear whether this is a territorial prefect or the officer commanding a military unit, a wing (ala) or cohort. This mysterious prefect forms a pair with an ἐπίτροπος (epitropos) who is hierarchically lower. This is presumably an ἐπίτροπος τῶν μετάλλων, thus a procurator metallorum, who was an imperial freedman. We know the names of two of these prefects. One is called Vibius Alexandros. He receives a pessimistic letter addressed to him by a noncommissioned officer, left in charge of Mons Claudianus with the title of vice-curator and many logistical problems on his hands.12 The ostracon contains the draft of two letters from the vice-curator, one to the prefect, and a second one on the same subject addressed to the procurator metallorum Tertullus. The letters date from Phamenoth 5 of year 29 of the reign of Commodus, or 1 March 189. Vibius Alexandros happens to be known through a papyrus in Leipzig as the epistrategos of Heptanomia. In this capacity, 9. I published them in O.Claud. III. 10. O.Claud. III 528 and 587. 11. Cockle 1996. 12. Chapter 5.

12

Rome in Egypt’s Eastern Desert

he receives a petition dating from the first months of 189. The cumulation of an epistrategia—a procuratorial post specifically Egyptian—with a prefecture is not unique: there are two parallels, but they do not help us to decide whether Vibius Alexandros owes his title of prefect to a unit command or to a territorial prefecture, since both cases are represented. The name of the other prefect who has authority over Mons Claudianus is Antonius Flavianus. He is the recipient of two draft letters written by the native quarrymen to announce that two columns are ready (with the help of Sarapis) and that they need to be sent steel and coal so that they can complete the third.13 This prefect is also assisted by a procurator. I was surprised to find Antonius Flavianus on an ostracon of the Berenike Desert, coming from the fort of Dios. The name of Antonius Flavianus in the dative fills the first line of the copy of a letter addressed to him by the curator of Dios.14 The sherd is a fragment of an amphora on which the curator copied official correspondence, or perhaps just the letters he sent. In the latter case, it would be a liber litterarum missarum, but the state of the document does not allow us to be certain. The title of Antonius Flavianus does not appear, unfortunately. All I can say is that, when they are detached to the Berenike Desert, the curators have as their direct hierarchical superior the Prefect of Berenike, and that it is with this official that they exchange correspondence, except when it is of a purely local nature. Is it possible that the two areas I have distinguished in the Eastern Desert were, at a given moment, at the end of the second century or the beginning of the third, under the authority of a single Roman equestrian official? The possibility remains open, but the state of the documentation does not allow us to say more. It may simply be a temporary situation, in which Antonius Flavianus acted as an interim prefect of Berenike.

II. Analytical classification of Greek and Latin toponyms To present the toponyms I will use a system of classification by topographical features. The rules for the formation of toponyms, their syntactic behavior, and their thematic typology vary according to these features (this is not unique to Greek): metalla; quarries in the narrow sense of extraction sites (λατομίαι); praesidia; wells (ὑδρεύματα); roads; Red Sea ports. The names bestowed by today’s desert dwellers on the praesidia are sometimes those of the wadis in which they stand. The ostraca never mention names of features belonging to physical geography, such as wadis or mountains. Is this a bias of this kind of source, or did the men sent from the Nile valley, contrary to the Beduins of today, not bother to bestow names on such features? For these sedentary people, the desert may have been only a network of human settlements, joined by a limited number of much trodden roads. The same indifference to geography is reflected in the fact that no desert place-name refers to geographic directions and is called Northern, Eastern, etc. The names used in the Eastern Desert in the imperial period are mainly known to us from Greek texts, more rarely from Latin ones. They were almost all created under the Ptolemies or under Roman rule, so that some are Latin. Several are in other languages which we cannot always determine, and we do not know if they were assigned by the Romans or if they belong to an earlier toponymic substrate.15 For the analytical description of toponyms, I have used Dorion and Poirier 1975 and Löfström and Schabel-Le Corre 2005. In particular, I borrow from the latter the distinction between appellative and proprial, applied both to nouns and adjectives; for example: 13. O.Claud. IV 848 and 850. 14. O.Dios inv. 514. 15. List p. 78.

A survey of place-names in the Egyptian Eastern Desert

13

Table 1.1. Distinction between appellative and proprial Noun

Adjective

ὅρμος, λατομία, ὕδρευμα, ἄκανθα, σμάραγδος, φοινικών, ἀλαβάρχης, πορφυρίτης, καμπή, πέλαγος

μέλας, μέγας, ξηρός, φαλακρός

appellative proprial

Βερενίκη

Αὔγουστος, Κλαυδιανός

Another distinction, between generic and specific constituents, is essential in toponomastics. In the place-names of the Eastern Desert, the generics are common nouns referring to topographic features: ὁδός, λατομία (and κοπή), μέταλλον, ὅρμος, ὄρος, πραισίδιον, ὕδρευμα. Toponymists are divided on the status that should be given to the generics which designate features: do they, or do they not, form part of the toponym? In toponymy, the status of the generic is, indeed, floating. It depends on the feature, on the context of enunciation, and on the specific element: the nature of the latter (locative/descriptive, proprial/appellative), its singular or common character, its formal characteristics, and in particular the number of syllables. In the Sargasso Sea, Sea is part of the name, but it may be abbreviated to Sargasso, which is not the case for the South China Sea, where the specific is locative. Mont Blanc cannot be reduced to the specific, but Mount Kilimanjaro is generally abbreviated to Kilimanjaro and Rocky Mountains to Rockies. For some toponymists, in the phrase the Town of Manchester, Town is an element of the name, which is less questionable for Georgetown. In the case of the Eastern Desert, it seemed more effective to consider the generics listed above as being part of the toponym: this makes it possible to understand better the differences in their behavior and in particular the question of solid compounds (such as Georgetown). Greek city names are often open compounds where the generic element πόλις is postpositioned and sometimes elided, whereas the first item, an adjective or a noun in the genitive, retains its ending: cf., e.g., Ἡρακλέους πόλις. But the derivative noun Ἡρακλεοπολίτης, which refers to the corresponding nome, is a solid compound. We will see, in the toponymic corpus of the Eastern Desert, that certain generics are never postpositioned, which is the first stage towards the formation of solid compounds. Finally, we should note the phenomenon of “transfer” of the generic element, when it refers to another feature than the one which it names: thus, the place-name Μέλαν Ὄρος, “Black Mountain,” is not the name of a mountain, but of a praesidium.

III. Metalla In Roman Egypt, μέταλλον signifies a geographical and administrative entity including extraction sites (λατομίαι) and all the necessary facilities for the work and daily life of the workers and the military and administrative staff (dwelling places, offices, wells, stables, granaries, sanctuaries, forges, baths, etc.). When metallon designates one of these entities, it is, in the ostraca from Mons Claudianus, in the singular: ἐν μετάλλῳ Κλαυδιανῷ, μέταλλον Τιβεριανόν,16 ἐν μετάλλ(ῳ) Πορφ[υ]ρίτ(ου),17 ἀπὸ μετάλλου Ἀλαβαστρίνης.18 In the papyri, however, this generic is usually in the plural, even when referring to a specific metallon: τοῖς Πορφυρειτικοῖς καὶ Κλαυδιανοῖς μετάλλοις (P.Oxy. XLV 3243.14 [214/5]), τῶν κ̣α̣τ̣ὰ τὴν Ἀλαβαστρίνην μετάλ[λων] (P.Sakaon 24.2–3 [325]). But, in general, μέταλλον is omitted. 16. O.Claud. inv. 6179. 17. O.Claud. IV 854.3. 18. O.Claud. inv. 6366.

14

Rome in Egypt’s Eastern Desert

The distinction between μέταλλον and λατομία is not always clear-cut: the μέταλλον Τιβεριανόν quoted above is alternatively designated as Τιβεριανή; the metallon of Umm Balad is called Καινὴ Λατομία and, not far from there, is Γερμανικὴ Λατομία. The gender of the specifics Δομιτιανή and Τιβεριανή shows that the implicit generic is λατομία, and one observes that if the dedication of Publius Agathopous in Ophiates evokes πάντων τῶν μετάλλων τῆς Αἰγύπτου (I.Pan 51 [AD 11]), the parallel inscription that he made at Wadi al-Hammamat seven years later gives λατόμων πάντων τῆς Αἰγύπτου (I.KoKo. 41), where the abnormal phrase λατόμων πάντων is considered by Dittenberger, correctly in my view, as an error for λατομιῶν πασῶν. Table 1.2. Names of metalla in the Eastern Desert: semantic classification Imperial Reference Γερμανικὴ Λατομία Δομιτιανή Κλαυδιανόν Τιβεριανή

Name of Material Βάζιον Μαργαρίτης Ὀφιάτης Πορφυρίτης Σμάραγδος

Descriptive Καινὴ Λατομία

Anthropophoric? Ἀλαβάρχης Πέρσου*1

Uncertain Ταμόστυμις*2

* Commented on in the section Praesidia (pp. 46 f).

1. A ghost: the complex toponym “Mons Porphyrites” The familiar names of Mons Claudianus and Mons Porphyrites are in fact poorly documented. For Mons Claudianus, there is only one example: the dedication in Latin, on an altar at Mons Claudianus by a centurion appointed directly by the emperor to be in charge of the metallon.19 The Greek equivalent, which would be τὸ Κλαυδιανὸν Ὄρος, is not attested. The neuter adjectival noun Κλαυδιανόν is normally used alone, sometimes preceded by the article; in the rare cases where the generic, normally implicit, is expressed, it is μέταλλον. Thus, in O.Claud. IV 853, a letter addressed collectively to Probus, procurator metallorum, the quarrymen working in the metallon of Claudianus call themselves ἐργαζομένων ἐν μετάλλῳ Κλαυδιανοῦ (l. Κλαυδιανῷ20). Exceptionally, the generic πραισίδιον replaces μέταλλον: when, probably from the time of Antoninus Pius onwards,21 the commander of Mons Claudianus on site is no longer a centurion, but instead a curator, his title is in most cases κουράτωρ Κλαυδιανοῦ, seven times κουράτωρ μετάλλου Κλαυδιανοῦ, and only twice κουράτωρ πραισιδίου Κλαυδιανοῦ.22 On the other hand, the Latin phrase Mons Porphyrites is not attested in any ancient sources. In papyrus documents, what we commonly call Mons Porphyrites is simply called “the Porphyrites,” ὁ Πορφυρίτης (remember that ὁ πορφυρίτης is a masculine noun, meaning “porphyry”: we shall return to the toponymic use of names of materials). It is the same in the only Latin inscription that mentions it, despite its solemn character. It adorned the facade of the residence, situated at Hermou polis, of the accountants of the mines and quarries: hosp(itium) tabula(riorum) Porphyr(itae) et aliorum metal19. I.Pan 39: Annius Rufus (centurio) leg(ionis) XV Apollinaris praepositus ab Optimo Imp(eratore) Traiano operi marmorum Μonte Claudiano (…). 20. This casual error is not rectified in the apparatus criticus of the edition. 21. Chapter 15, pp. 238–40. 22. O.Claud. inv. 8094: a letter from the κουράτορ πρεσιδ⟨ί⟩ω Κλαυδιανῶ (sic), where the writing betrays a Latin speaker; and O.Claud. II 372, a letter of Aelius Serenus, who refers to himself as κουράτωρ πραισιδίου Κλαυδιανοῦ, while in his letter O.Claud. II 371, which is in another hand, he is κουράτωρ Κλαυδιανοῦ μετάλλου.

A survey of place-names in the Egyptian Eastern Desert

15

lorum.23 The complex toponym Mons Porphyrites has obviously been created in modern times, either from Latin translations of Ptolemy’s Geography, dating from the Renaissance, or by analogy with Mons Claudianus (a toponym regularly formed as: generic noun + proprial adjective as specific element). As for the matching Greek Πορφυρίτης Ὄρος, which is in the list of Greek place-names compiled by Redard,24 this is also not attested, at least not in the documentary sources (papyri and inscriptions). No wonder, since it contravenes the rules of the composition of complex toponyms in Greek: the specific element, if a noun (common as well as proper), is in the genitive. See Μυὸς Ὅρμος,25 Ἀπόλλωνος Ὕδρευμα, Ἡρακλέους πόλις. In Greek as in Latin, complex toponyms never consist of two nouns of which one is in apposition to the other. But, what can we say about two passages, by Ptolemy and Palladius, where Πορφυρίτης is used with ὄρος? At first, it should be observed that the phrase *Πορφυρίτης Ὄρος would be all the more surprising in Greek as the two nouns are of different genders: a phrase indicating that ὁ Πορφυρίτης is to be understood as the mountain, not the stone, should be ὁ Πορφυρίτης τὸ ὄρος (“the Porphyrites mountain”). Ptolemy’s Geography, as it came to us, more or less respects this rule in constructed sentences or after a preposition, not—at least in appearance—in the table of ground coordinates. For example: ὁ Παρνασσὸς ὄρος, but παρὰ τὸν Καρπάτην τὸ ὄρος ... ὁ μὲν Αἶμος τὸ ὄρος κεῖται. In fact, ὁ Παρνασσὸς ὄρος is not a phrase, but ὄρος is only a gloss directed to the person who is drawing the map. Therefore, ὁ Παρνασσὸς ὄρος is to be understood, not as “Mount Parnassus” but as “The Parnassus (mountain).” The same reasoning applies to Ptolemy’s Geography 4.5.15, Σμάραγδος ὄρος [coordinates], which should be translated “Emerald (mountain).” Coined by modern scholars, the toponym “Mons Smaragdus” is another incorrect construction.26 Πορφυρίτης and Σμάραγδος are simple toponyms. In the phrase ὁ Πορφυρίτης τὸ ὄρος, ὄρος is just an apposition, not the generic element of a complex toponym. Here is the passage, which is not necessarily free of corruption, where Ptolemy mentions the Porphyrites (Geogr. 4.5.27): τὴν δὲ παρὰ τὸν Ἀραβικὸν κόλπον ὅλην παράλιον κατέχουσιν Ἀραβαιγύπτιοι ἰχθυοφάγοι, ἐν οἷς ὀρειναὶ ῥάχεις· ἥ τε τοῦ Τρωϊκοῦ λίθου ὄρους (coordinates) καὶ ἡ τοῦ ἀλαβαστρηνοῦ ὄρους (coordinates) καὶ ἡ τοῦ πορφυρίτου ὄρους (coordinates) καὶ ἡ τοῦ μέλανος λίθου ὄρους (coordinates) καὶ ἡ τοῦ βασανίτου λίθου ὄρους (coordinates) “The entire coastline along the Arabian Gulf is inhabited by Arabegyptian fish-eaters, in whose territory are mountainous massifs: the massif of the mountain of Trojan stone (...); the one of the mountain of alabaster (...); the one of the mountain of porphyry (...); the one of the mountain of black stone (...); the one of the mountain of bekhen stone (...).”

23. Cockle 1996. 24. Πορφυρίτης Ὄρος (Redard 1949: 149). 25. The particular case of the “solid composition” Μύσορμος is not taken into account. It confirms a fact of popular pronunciation which removes the o of Μυός. 26. The emerald mines are called σμαράγδεια μέταλλα in Heliodoros, Aethiopica 10.11.1.

16

Rome in Egypt’s Eastern Desert

The other lines, particularly those where the word λίθος appears, show that πορφυρίτου is here the name of the material, not the specific element of a complex toponym of which the nominative would be ὁ Πορφυρίτης Ὄρος; πορφυρίτου is a genitive of material qualifying τοῦ ὄρους. So, we translate this line: “the (mountain range [ῥάχις]) of the mountain of porphyry.” In Palladius, however, it is difficult to deny the existence of an abnormal phrase of which the nominative would be ὁ Πορφυρίτης ὄρος: ὃς τὰ μὲν πρῶτα ἔξω πάσης Αἰγύπτου καὶ Θηβαΐδος ἐν τῷ Πορφυρίτῃ ὄρει μόνος ἀναχωρήσας κτλ. (Dialogus de vita Joannis Chrysostomi, ed. Sources Chrétiennes 341, XVII 82). Nevertheless, this formulation is incorrect. Several explanations can be proposed, among which it is difficult to make a choice: (1) The text is corrupt: one should restore ἐν τῷ Πορφυρίτῃ ⟨τῷ⟩ ὄρει. (2) ὄρει is a gloss aiming at disambiguating Πορφυρίτῃ, but it has migrated into the text. (3) As neuter and masculine items have the same form in the dative, it was perhaps less necessary to apply the rule ὁ Ὄλυμπος τὸ ὄρος. (4) Πορφυρίτης is clearly treated as an adjective. This adjectival use is only attested in Greek in an inscription from Smyrna, which lists κείονας εἰς τὸ ἀλειπτήριον Συνναδίους οβʹ, Νουμεδικοὺς κʹ, πορφυρείτας ϛʹ (IGRR IV 1431 = IK 24/1, 697 [124–138], lines 40–42). But at least this masculine word qualifies a masculine noun, which is not the case with Palladius. I also noted an example in Latin, where the noun (if the reconstruction l]agonam is correct) is feminine:27 ]agonam / porphyriten / cum basi d(e) s(uo) d(edit) (CIL XIII 4319). In documents from Egypt, πορφυρίτης is never used as an adjective. The adjective derived from it is πορφυριτικός (“in porphyry” or “relating to Porphyrites”).

2. Toponyms from materials: a pearl fishery in the Red Sea The Greek word πορφυρίτης was created one day in July AD 18 from the base πορφύρα (i.e., the purple dye which comes from two species of Murex) and the suffix -ίτης often used, from the Hellenistic period onwards, for the formation, among other things, of masculine names of stones.28 Gaius Cominius Leugas, a prospector who roamed the Eastern Desert in the reign of Tiberius, had just discovered this purple-colored rock, as well as several others, and as a thanksgiving he dedicated a chapel on-site to Pan and Sarapis (AE 1995, 1615). When the massif began to be exploited, becoming a metallon, ὁ πορφυρίτης was used as a placename. One can observe at this time the same semantic shift in meaning, from material to placename, in the dedication of a Paneion situated in another metallon, in Wadi Umm Wikala (I.Pan 51 [AD 11]). This inscription, seven years prior to the discovery of Porphyrites, offers a series of place-names drawn from materials which has not been unanimously recognized as such: ἐπ{ε}ὶ Ποπλίου Ἰουεντίου Ῥούφου χιλιάρχου τῆς τερτιανῆς λεγεῶν(ος) καὶ ἐπάρχου Βερνίκης καὶ ἀρχιμεταλλάρχου τῆς ζμαράγδου καὶ βαζίου καὶ μαργαρίτου καὶ πάντων τῶν μετάλλων τῆς Αἰγύπτου, ἀνέθηκε ἐν τῶι Ὀφιάτηι ἱερὸν Πανὶ θεῶι μεγίστωι (l. 2–12). Ὀφιάτηι presents itself clearly as the name of the place, 27. P. Flotté (Carte Archéologique de la Gaule 57/2. Metz, Paris 2005, 285) is hesitant about the reconstruction of l]agonam, probably because epigraphic references to lagona (pitcher) are always graffiti on an object. But why couldn’t it be the effigy in porphyry of a lagona? 28. It belongs to a group of masculine and feminine nouns formed on a nominal base, and characterized by the suffix -ίτης for masculine, -ῖτις for feminine. These denominational derivatives, which have proliferated from the Hellenistic period, are frequently trade-names, terms of botany, zoology, geology, and geography. Just think of the names of the Egyptian nomes: ὁ Ὀξυρυγχίτης νομός. The classic book about them is Redard 1949.

A survey of place-names in the Egyptian Eastern Desert

17

and it is, as in the case of Porphyrites, a direct borrowing from the extracted material. Indeed, ὀφιάτης is a dialect form of the stone-name ὀφίτης,29 mentioned by Pliny (Nat. 36.55): Pliny explains the difference between Augusteum and Tibereum, discovered in Egypt under Augustus and Tiberius, and ophite, the source of which he does not indicate, but it is likely the “Theban ophite” colored with small flecks, evoked by Lucan: parvis tinctus maculis Thebenus ophites (Pharsalia 9.714). The fact that the words ζμαράγδου καὶ βαζίου καὶ μαργαρίτου are syntactically on the same level as πάντων τῶν μετάλλων τῆς Αἰγύπτου shows that these nouns are used as toponyms; they should therefore be provided with an uppercase letter:30 ἀρχιμεταλλάρχου τῆς Ζμαράγδου καὶ Βαζίου καὶ Μαργαρίτου καὶ πάντων τῶν μετάλλων τῆς Αἰγύπτου, “chief director of the Emerald, the Topaz, the Pearl, and all the mineral resources in Egypt.” The same enumeration appears seven years later, in another inscription made by the same dedicant, Agathopous, freedman of the prefect Iuventius Rufus, engraved on a naos in the stone quarries of bekhen in Wadi al-Hammamat (I.KoKo. 41 [AD 18]). While Smaragdos31 and Bazion (St. John’s Island, where topaz was mined)32 have been identified, the location of Margarites is unknown. I agree with Dittenberger, who believes it to be somewhere in the Red Sea, where conditions are favorable to the biology of pearl oysters.33 They are, in this region, Pinctada radiata and Pinctada margaritifera. According to G. Ranson, Pinctada radiata thrives in the tropics and a little beyond, in areas where freshwater inflows mitigate salinity.34 It would be plausible that the Romans had imposed an institutional framework on a fishery traditionally practiced by a local population of Ichthyophagoi. Likewise, according to Strabo, in the early years of the principate, the Smaragdos was exploited by Beduins whom he calls Ἄραβες,35 as it would be again in Late Antiquity by the Blemmyes.36 If the pearls of the Indian Ocean and the Persian Gulf (especially those of Tylos, now Bahrain)37 are mentioned by ancient authors (but after the conquests of Alexander), the two dedications made by 29. In principle, it should be derived from a base ὀφια. But cf. Chantraine 1933: 311: “The suffix [sc. -ατας, -ητης, -ατης] was sometimes extended to derivatives, although no base in long α was independently attested: πολιήτης, “citizen,” is the normal derivative of πόλις in dialects other than Ionian-Attic, cf. πολιάοχος.” The form ὀφιῆτις is attested. 30. Dittenberger had previously proposed this idea (OGIS II 660, note 4). The toponymic use of the name of the material probably comes from a Greek usage: Pliny (Nat. 37.73.3) mentions an emerald deposit in Chalcedon, called Smaragdites (ὁ σμαραγδίτης is a duplicate of ἡ σμάραγδος): mons est iuxta Calchedonem, in quo legebantur, Smaragdites vocatus. 31. This was already the name of this metallon in the third century BC, according to an ostracon from Biʾr Samut, which reads ἐ]π̣ὶ̣ τὴν Μάραγδον (O.Sam. inv. 303). 32. On this deserted island off Berenike, today Jazirat Zabarjad, see Fournet 2018: http://books.openedition.org/cdf/5242, § 14. 33. OGI II 660, note 6. 34. Ranson 1961: 11–12. On the distribution of pearl oysters in the Red Sea, see Donkin 1998: 29–36. 35. Probably in the anthropological sense of Beduins, which the word also has in Greek, but not in the ethnic sense. 36. Has exploitation by the people of the desert ever stopped in between? The presence of the Roman army in the first and second centuries in the emerald mines is attested by the discovery of elements of lorica squamata (Sidebotham et al. 2008: 299). Have the Romans, when they took over the emerald mines, employed a Beduin task-force? Unfortunately, no ostracological discovery gives us clues about the composition of the workforce or the administrative framework. Local architecture is also a difficult issue: no fortified square building, but a multitude of small huts and several imposing official buildings. S. E. Sidebotham hypothesized that the large praesidium of Apollonos Hydreuma could have housed the garrison that controlled the exploitation of emerald mines (Sidebotham et al. 2008: 301); but it is at a distance of 20 km from the village of Sikayt, a central inhabited area of one of the mining districts of Smaragdos, which consists of several such. The modern name of Sikayt was connected by Letronne with the epithet of Isis read by nineteenth-century travellers on a rock inscription of the site, now destroyed (I.Pan 69). The best facsimiles, those of Nestor Lhôte and of Wilkinson are, respectively παρατηκυριιϲκαιτηϲενεκειτνει and παρατηκυριιϲιδιτηϲενϲκειτηϲ. The conjecture παρὰ τῇ κυρίᾳ Ἴσιδι is reasonably safe; from the epiclesis which comes next, modern scholars deduced the toponym Σενσκις, which would be the name of one of the exploitation areas of Smaragdos. 37. In the Erythrean area, the largest concentrations of Pinctada radiata are in Bahrain and Ceylon.

18

Rome in Egypt’s Eastern Desert

Figure 2. Pliny’s marmor Tibereum: column in granodiorite from Tiberiane in Zenon’s chapel at St. Praxedes. © A. BülowJacobsen

A survey of place-names in the Egyptian Eastern Desert

19

P. Iuventius Agathopous are the only ancient testimony to the possible presence of a pearl fishery in the Red Sea. Neither Pliny nor the Periplus Maris Erythraei says anything about it; and yet, the latter mentions the Persian Gulf fisheries, though outside the route described.38 We cannot exclude the possibility that the Margarites, operated by the State in the early years of the province, quickly found itself competing with foreign pearls, so that it would not have been considered worth the trouble to organize exploitation when Indian pearls and those from the Persian Gulf were flooding the market. But this operation may simply have left no written record other than the two early inscriptions of Agathopous, for shells of Pinctada radiata and Pinctada margaritifera were found in the excavations at Myos Hormos and Berenike, and the locals obviously liked to rework the shell.39 One would expect, however, in the case of a fishery, large shell dumps40—unless of course the fishermen threw them back into the sea.41 Such dumps do exist in the lagoon of Berenike, but local constraints (minefields) have not allowed archaeologists to go and look at the composition.42 The traditional interpretation of the name Margarites has however been challenged by Pierre Schneider:43 μαργαρίτης in the inscriptions of Agathopous might not be a pearl, but the gem called χερσαῖος μάργαρος/μαργαρίτης by Aelian (De Natura Animalium 15.8.30) and by Origen (Commentarium in Evangelium Matthaei 10.7). These two sources place the “land-pearl” in India, a notoriously vague geographical concept that could include any region connected with Red Sea trade; Aelian curiously states that this gem has no inherent nature, but is generated by the rock crystal (ἀπογέννημα εἶναι κρυστάλλου, οὐ τοῦ ἐκ τῶν παγετῶν συνισταμένου, ἀλλὰ τοῦ ὀρυκτοῦ). Schneider connects this passage to Plin. Nat. 37.23: citing Juba, Pliny reports that crystallum is found—among other sources— on the island in the Red Sea which also produces topaz, i.e., St. John’s Island. For Schneider, μαργαρίτης would thus be the Greek name for rock crystal extracted, according to Juba, from St. John’s Island and a nearby island called Necron (= Νεκρῶν).44 A final example of the name of the material used as a toponym (not to designate a metallon, but a latomia) is found in an archive of ostraca from Porphyrites, dating to the late third–fourth century. A number of these notes are orders to send bread to various microsites of the metallon, among them “the Batrachites” (εἰς τὸν Βατραχείτην). W. Van Rengen45 recalls in this respect a passage of Pliny, according to which “Koptos also exports batrachitai”; the naturalist distinguishes two types of these frog-colored minerals.46 Is there a confusion with the batrachites later attested in Porphyrites? 38. Peripl.M.Rubr. 35: ἐκδέχεται μετ’ οὐ πολὺ τὸ στόμα τῆς Περσικῆς καὶ πλεῖσται κολυμβήσεις εἰσὶν τοῦ πινικίου κόγχου. 39. S. Hamilton-Dyer, “Faunal Remains,” in Peacock and Blue (eds.) 2006b: 273. 40. Especially when you know that an average of 500 oysters have to be sacrificed to get a few pearls (Strack, “Introduction,” in Southgate and Lucas 2008: 13). In the Red Sea, we find such shell dumps on the islands of Dahlak and Farasan, at a more southern latitude (Sharabati 1981: 53). We know that there was, under Antoninus Pius, a Roman garrison at Farasan, but the oyster deposits there are not ancient. 41. The two scenarios are possible: Schörle 2015: 48 f. 42. Schörle 2015: 48. 43. Schneider 2016. Unlike Schneider, I have no problem admitting that the ancients were able to classify the pearl fisheries in the category μέταλλον, because pearls are often likened by the authors to stones (λίθοι). 44. Aelian gives the impression that it is a product derived from crystal. For other hypotheses about the nature of this terrestrial Indian pearl, see RE XIV.2, 1700 (which favors the hypothesis that it must be bamboo resin tears). 45. The Roman Imperial Porphyry Quarries, Gebel Dokhan, Egypt, Interim Report 1998, 26 (unpublished). Βατραχίτης is the only name of a latomia found in the O.Porph. 46. Batrachitas quoque Coptos mittit (Nat. 37.149). But in this passage, Pliny mentions gems, not an architectural material. Quoque is in reference to another gem exported by Koptos, which Pliny names the balanites: Balanitae duo genera sunt, subviridis et Corinthii aeris similitudine, illa a Copto, haec ab Trogodytica ueniens, media secante flammea uena, “As to the ‘balanites’, or ‘acorn-stone’, there are two varieties, of which one is greenish and the other like Corinthian bronze in its color.

20

Rome in Egypt’s Eastern Desert

While materials (whose names existed before the Romans, such as σμάραγδος, βάζιον, and μαργαρίτης, or were invented by them at the time of discovery, such as βατραχίτης, ὀφιάτης, and πορφυρίτης) became place-names, conversely names of materials are derived from toponyms: the granite of Mons Claudianus was called marmor Claudianum,47 that of Tiberiane marmor Tibereum (Plin. Nat. 36.55; Fig. 2).

3. Metalla in the area of Umm Balad Presumably the number of occurrences of toponyms on ostraca is more or less inversely proportional to the distance between the named sites and the place of discovery (see Table 5, where all the toponyms found on ostraca from Umm Balad appear, whatever the topographic feature). A. Δομιτιανή vs. Καινὴ Λατομία48 The Umm Balad ostraca, excavated in 2002 and 2003, revealed several completely new names referring to metalla, wells, and praesidia. Unfortunately, we do not know which sites they correspond to on the ground. There is even an uncertainty about the name of the metallon of Umm Balad, since two names appear concurrently, Δομιτιανή and Καινὴ Λατομία. What I have long considered as the most likely scenario is that the metallon was opened under Domitian between AD 89 and 9149 with the name Δομιτιανή, which would have been replaced by Καινὴ Λατομία after the damnatio memoriae of the emperor (but why call a quarry which had existed for several years “new”?). However, analysis of the stratigraphy of the midden made by J.-P. Brun does not fit with this interpretation as neatly as would be hoped: the name Καινὴ Λατομία appears already in SU 4, an early layer of the dump, in which Καινὴ Λατομία and Δομιτιανή are found concurrently. Δομιτιανή, the attestations of which are considerably fewer than those of Καινὴ Λατομία, also appears in layers with Antonine material. Table 1.3. Δομιτιανή and Καινὴ Λατομία. Distribution of occurrences according to stratigraphy1 Καινὴ Λατομία

Δομιτιανή

Phase 1 (c. 91–100)

27

5

SU 4 (= earliest layer of phase 1 containing ostraca)

16

1

Phase 3 (Antoninus)

23

9

1. Occurrences of which the stratigraphy is not certain have been excluded.

The former comes from Coptos and the latter from the Cave-dwellers’ country, and both are intersected through the middle by a bright red layer” (trans. D. E. Eichholz, Loeb). The description of Pliny, except for the fire-colored vein, perfectly matches the graywacke of Wadi al-Hammamat, probably exported through Koptos, whose tones range from dark green to dark gray and, once polished, has a bronze patina. The “balanites” might be a ghost-word and a ghost-rock. Pliny knows graywacke under its correct name basanites: quem (lapidem) vocant basaniten, ferrei coloris atque duritiae (Nat. 36.58). This comparison with the color of iron, not bronze, is less felicitous. Also Nat. 36.147. 47. Edictum Diocletiani de pretiis rerum venalium 33.6 (Lauffer); 31.6 (ZPE 34 [1979] 163–210). 48. Gnoli 1971: 133 called the diorite extracted at Umm Balad granito verde fiorite di bigio; it appears in the Palatine palaces. Gnoli especially mentions the slabs of the pavement in the triclinium of the Domus Flavia and adds that this material was used to make floor or wall slabs and small objects such as columnettes. 49. The fragmentary dedication of the fort has an erased line where the name of the praefectus Aegypti would be expected: it must thus be Mettius Rufus, prefect between 89 and 92. Furthermore, the oldest dated ostracon is from 91.

A survey of place-names in the Egyptian Eastern Desert

21

J.-P. Brun used to explain these paradoxes as follows. The site had from the start two names: Καινὴ Λατομία referred to the anteriority of Porphyrites, Δομιτιανή was the honorific name. The approximately eight years of work under Domitian during which it was officially used must have left a very thin layer at the bottom of the dump. Then, the damnatio memoriae did not prevent people continuing to use Δομιτιανή locally in private texts, unless the ostraca mentioning Δομιτιανή in phase 3 are residual material. J.-P. Brun’s explanation was based exclusively on the stratigraphy, but two names for one site is a most uncommon occurrence in the desert: the only other instance is another metallon, Mons Claudianus, which had an honorific name as well, at least according to a hypothesis of mine; but this name, Fons Traianus Dacicus, was never used by ostraca writers.50 A solution to the problem occurred to me recently when preparing the publication of Umm Balad. I realized that the layers in phase 3 are actually mixed and contain a lot of Domitianic material: the metallon must have originally been called Καινὴ Λατομία Δομιτιανή (or perhaps in another order), its supposed two names being only two ways of abbreviating, one of which of course disappeared after Domitian. B. Small Metalla Listed in the Area of Umm Balad, Orphan Toponyms Besides Porphyrites and Claudianus, which we have treated above, the O.KaLa. mention two other metalla, Germanike Latomia and Alabarches. To identify their location, we have three possible candidates,51 which are all equally viable: (1) The area formed by the two quarries and the workers’ village52 located at the bottom of the wadi, access to which was controlled by the praesidium of Umm Balad; the distance between the fort and the extraction zone is, meandering along this wadi, c. 1.20 km. Letters mentioning Alabarches suggest, without certainty, that it could be this site; Alabarches would, therefore, be a microtoponym within the metallon of Domitiane/Kaine Latomia. (2) The quarries of blue porphyry at Umm Tuwat (27° 10’ 12” N/33° 14’ 25” E).53 The Romans extracted, in the first to second centuries, a blue-gray porphyry (trachyandesite porphyry)54 that has stick-like inclusions reminiscent of the green porphyry from Sparta (Fig. 3). Umm Tuwat is c. 6 km as the crow flies northwest of Umm Balad (there is no direct route between the two sites, but we have not had time to check if there were mule tracks). The site includes no agglomeration;55 Bagnall and Harrell only indicate three extraction points (Fig. 4) and two small buildings of 3.5 m2 and 5.5 m2. There are very few ceramics, but the wide and beautiful stone-free track that leads there suggests that the Romans had ambitions for this site (Fig. 5). It has been suggested that the blue porphyry of Umm Tuwat was the knekites, one of the materials listed by Gaius Cominius Leugas, the discoverer of Porphyrites.56 Umm Tuwat porphyry was spotted in the palaces of the Palatine, but objects made of this material are extremely rare; the main one is the right column at the entry of the chapel of St. Zenon at St. Praxedes (Fig. 6).57 50. Cf. p. 65. 51. I do not take into account isolated mining, as on the walled rock of Badiya or near the praesidium of Qattar, which are mere exploratory tests (Brown and Harrell 1995: 224). 52. 27° 9’ 11.71” N/33° 17’ 0.24” E. 53. Coordinates according to Brown and Harrell 1995: 224. 54. http://www.eeescience.utoledo.edu/faculty/harrell/egypt/Quarries/Hardst_Quar.html 55. I went there with A. Bülow-Jacobsen in January 2004. 56. Bagnall and Harrell 2003. 57. Gnoli 1971: 113. While Bagnall and Harrell perceive the color of this material as pale, for Gnoli, who calls it porfido serpentino nero, it is a dark rock.

22

Rome in Egypt’s Eastern Desert

Figure 3. A porphyry plate from Umm Tuwat polished on one side, found in the midden at Umm Balad. © Rights reserved by H. Cuvigny

Figure 4. A latomia at Umm Tuwat. © A. Bülow-Jacobsen

A survey of place-names in the Egyptian Eastern Desert

23

Figure 5. The track towards Umm Tuwat. © A. Bülow-Jacobsen

Κνηκίτης is derived from κνῆκος, “saffron,” Carthamus tinctorius, which, used as a dye plant, produces yellow. Now the porphyry of Umm Tuwat has no yellow in it, but if one accepts the idea that the ancients perceived colors more in terms of their light intensity than their hue, knekites could mean “pale stone” and, of all the materials extracted in the Eastern Desert, only the porphyry of Umm Tuwat could, according to the authors of this hypothesis, fit that description. (3) The diorite quarry of Umm Shejilat,58 which is c. 18 km, almost in a straight line, south of the praesidium of Qattar, was quarried during the first/second century AD.59 We did not go there, but the satellite image shows access to it from Qattar. The built-up area is not shaped like a praesidium. Mons Claudianus is not far away, but the mountainous terrain prevents direct communication between the two sites. Umm Shejilat had to be in the orbit of Porphyrites and perhaps more precisely of Umm Balad. However, caravans carrying provisions must have come from Kaine and left the hodos Porphyritou at the small station of Bab al-Mukhaniq, from where 58. 26° 56’ 30” N/33° 14’ 39” E. On the diorite from Umm Shejilat, cf. Gnoli 1971: 126; Brown and Harrell 1995: 224. It is called granito della colonna, the most famous object made in this material being a small column (actually a baluster) brought from the Holy Land in the thirteenth century by Cardinal Giovanni Colonna; assumed to be the column of the scourging of Jesus, it is kept in the church of St. Praxedes in Rome (one easily finds a picture on the web by searching Colonna della flagellazione). The discoverer of this quarry, an Egyptian engineer, reports a Roman well at Umm Shejilat (Gnoli 1971: 126, n. 2). 59. http://www.eeescience.utoledo.edu/faculty/harrell/egypt/Quarries/Hardst_Quar.html

24

Rome in Egypt’s Eastern Desert

Figure 6. Porphyry column from Umm Tuwat in Zenon’s chapel at St. Praxedes. © A. Bülow-Jacobsen

A survey of place-names in the Egyptian Eastern Desert

25

Figure 7. Map of the Tabula imperii romani, Sheet Coptos, published by Meredith 1958. © Rights reserved

Meredith’s map shows a road leading to the gold mine of Wadi Ghazza60 and to an anonymous granite quarry that corresponds, according to its location, to Umm Shejilat (Fig. 7). To name these three sites, two toponyms, as we have seen, are available: Γερμανικὴ Λατομία The name of this metallon appears only in O.KaLa. inv. 765, a receipt issued by a sklerourgos to a camel driver for a delivery ἐν Γερμανικῇ Λατομίᾳ (the delivered object is a load and a half of monthly rations for quarry workers).61 The document is dated Phaophi 2 of Domitian’s 16th regnal year, or 29 September 60. Or Ghozza. At this place, Meredith’s map shows also the symbol of a “probable fort.” It was excavated in January 2020 by the MAFDO, under the direction of Thomas Faucher and Bérangère Redon, and was shown to be a Roman praesidium called Berkou (see below p. 63). 61. The fact that Germanike Latomia was a delivery address for a camel driver bringing supplies indicates that the feature to which this place-name refers is a metallon, not a latomia in its usual meaning of a specific quarry-site.

26

Rome in Egypt’s Eastern Desert

Figure 8. O.KaLa. inv. 269. © A. Bülow-Jacobsen

96 (the news of the Emperor’s death, which had occurred on September 18, had not reached the site yet). Γερμανική refers to the honorific victory title, or cognomen ex virtute, Germanicus, to which Domitian was very attached, and shows that the metallon had been opened under this emperor, in the same way as Δομιτιανή. The Germanike Quarry should not be very far from Umm Balad, as the camel driver, who had probably loaded his animals at Umm Balad, had brought back the receipt as proof. This could be Umm Tuwat. The fact that two pieces of porphyry from Umm Tuwat were found in the midden at Umm Balad (one was polished, Fig. 3) shows that in any case the two metalla worked at the same time, that is to say under Domitian and/or Trajan. Isolated and poorly equipped, Umm Tuwat probably depended for its supplies on Umm Balad. Ἀλαβάρχης / Ἀραβάρχης This name is attested in seven ostraca from Umm Balad, six private letters, which are rather uninformative, and an amphora titulus published below (Fig. 8). Only the latter suggests that Alabarches/ Arabarches62 must have been a metallon. The amphora belonged to the architect Sokrates. Does the place-name refer to an arabarches63 or an individual called Arabarches, since this title was used—especially in Thebes and Elephantine—as a personal name?64 Since the question of customs duties is prob62. In the ostraca from Umm Balad, we have five occurrences of the spelling Ἀλαβ- and four of Ἀραβ-. 63. On the alabarchai, tax farmers who were sometimes fabulously rich, see Burkhalter 1999 and Kramer 2011: 175–84. 64. J. Gascou, pers. comm., suggests that Arabarches falls into the category of auspicious anthroponyms, the rich arabarches being proverbial.

A survey of place-names in the Egyptian Eastern Desert

27

ably irrelevant in this part of the Eastern Desert, I am inclined to favor the second hypothesis, and to relate Arabarches to the anthropophoric names of some latomiai at Mons Claudianus. The choice of a personal name may reflect the fact that Arabarches was a small metallon. For A. Bülow-Jacobsen, it could be the workers’ village at the foot of the quarries of Umm Balad (see p. 21). O.KaLa. inv. 269 (Fig. 8) Phase 1—SU 4 Titulus on amphora AE3.



10.5 × 3.5 cm

Domitian/Trajan (c. 91–100) Nile silt clay

Σω]κ̣ράτ( ) ἀρχιτέκ̣{κ}(τ-) Ἀραβάρχ(ου?) ][

“Sokrates, architect at Arabarches.” 1.

The absence of a preposition dictates the restoration of the genitive. This is the only case in the ostraca from the metalla where the trade-name ἀρχιτέκτων is determined by a place-name.

Ἀλαβάρχης is most often used without an article,65 but the paucity of occurrences does not allow us to make this a general rule: in the epistolary ostraca from Umm Balad, Πορφυρίτης preceded by a preposition is used interchangeably with and without the article. But πορφυρίτης originally is a common noun and, therefore, more likely to take the article, while Ἀλαβάρχης, as we have seen, may be an anthroponym.

IV. Latomiai: quarry-names at Mons Claudianus If the ostraca of Porphyrites have only revealed the name of a single quarry, Βατραχίτης,66 the Claudianus corpus has yielded a rich harvest of these microtoponyms. I have taken almost all of them from O.Claud. IV. The infrared photos that A. Bülow-Jacobsen took after publication of the book allowed me to make some corrections. In the inventory below, I only quote the instances that allow us to observe the morphological and syntactic behavior of these names. Of the 130 quarries reported by David Peacock, only seven, the names of which are known from the ostraca, were identified through inscriptions found on the spot: Epikomos, Harpocrates, Hieronymus?, Kochlax, Myrismos, Nikotychai, and Philok( ). I indicate, in this case, the number given by David Peacock.67

1. Inventory Ἄμμων? (Antoninus?) λατομ(ία) Ἄμμ̣[ωνος]: Ο.Claud. IV 719.6 (much restored). Ἄνουβις (Trajan) This quarry is mentioned once in the opening words of a list of experts in the form Ἀνούβι (Ο.Claud. IV 65. In eight cases, six without the article (among which are three examples of the phrase εἰς Ἀλαβάρχην) and two with (including P.Worp 20). 66. See p. 19. 67. List of quarries: Peacock and Maxfield 1997: 178–89. The numbers are shown on the plan published in O.Claud. IV, p. 10.

28

Rome in Egypt’s Eastern Desert

Figure 9. O.Claud. IV 816.1 (detail). © A. Bülow-Jacobsen

Figure 10. O.Claud. IV 866.4 (detail). © A. Bülow-Jacobsen

632.1). Although both lines of this incipit pose insuperable difficulties in reading and interpretation, I think that this is the dative Ἀνούβι, well-attested in the inscriptions (there are many examples outside Egypt, especially in Delos; in Egypt: I.Alex.impér. 124). Ἄπις (Trajan) Mentioned in the dative Ἄπιδι with other names of quarries and of other features that are allocated water rations in the large organizational chart O.Claud. inv. 1538.2 and 6 (Chapter 11). Ἀπόλλων (Trajan) Six certain occurences. Probably a different quarry from Apollon Epikomos, called simply Epikomos in the ostraca. – Ἀ]π̣όλλωνος: O.Claud. inv. 2853.5.68 In this catalogue of water distribution, quarry names are in the nominative when they are adjectives, and in the genitive when they are theonyms or anthroponyms. – λατομίᾳ Ἀπόλλωνο(ς): Ο.Claud. IV 634.2. – Ἀπόλλωνι: Ο.Claud. IV 741.1. – εἰς λατομίαν Ἀπόλλων[ος]: Ο.Claud. IV 786.3. – εἰς τὴν λατομίαν τοῦ Ἀ[πόλ]λωνος: Ο.Claud. IV 819.4–5. – ἰς λατομίαν Ἀπόλ(λωνος): Ο.Claud. IV 867.5. – [εἰς Ἀπόλ?]λ̣ωνος λατομίαν: Ο.Claud. IV 816.1 (Fig. 9). But the lambda is strange. Perhaps rather Νειλάν]μωνος (cf. below, Νειλάμμων)?

68. Organizational chart belonging to the same series as the document I published in Chapter 11.

A survey of place-names in the Egyptian Eastern Desert

29

In O.Claud. IV 866.4, the infrared image shows that Ἀπολλωνου was corrected by the scribe to Ἀπόλλωνος, a sigma having been written in the line space above the upsilon (Fig. 10). Given the context, it seems certain that this is a city, not a quarry, and that we can reconstruct this with confidence as ἀπῆλθεν εἰς Ἀπόλλωνο⟦υ⟧⸌ς⸍ | [πόλι]ν̣. Ἁρποκράτης (Trajan and Antoninus) The Harpocrates quarry is marked with a Latin inscription, where its name is preceded by the mark CEP, suggesting it was part of the caesura of Epaphroditos69 (= Peacock no. 109; cf. Peacock and Maxfield 1997: 188 and 220). This latomia is attested in three ostraca only: – [λατ]όμου Ἁρποχράτο̣[υ]: Ο.Claud. IV 635.1 (Trajan); – Arpochrate: Ο.Claud. IV 843.6 (Trajan?); – λατ(ομίᾳ) Ἁρπ̣ο̣(κράτου): Ο.Claud. IV 841.6 and 23 (c. 150). Αὐγούστη (Trajan) Αὐγο(ύστῃ): Ο.Claud. IV 775.8 and 776.11. Ἀφροδίτη (Trajan) λατομί[ᾳ] Ἀφροδε̣[ίτης]: Ο.Claud. IV 637.1–2. Βάρβαρος (end of the reign of Hadrian) λατομ(ίᾳ) Βαρβάρου: Ο.Claud. IV 730.1. Διόνυσος (Trajan and Antoninus) A dozen instances, including: – Διονύσῳ: Ο.Claud. IV 699.15. – Διονύσου πλάκες: Ο.Claud. IV 844.3. – δὸς εἰς τὸν Διόνυσον: Ο.Claud. IV 808. The theonym is preceded by the generic in two Antonine texts: λατο(μίᾳ) Διον(ύσου) (O.Claud. IV 841.13, where we learn that this quarry is part of the caesura of Epaphroditos) and ex lat(omia) Dionysu (Ο.Claud. IV 845.2). Διόσκορος This supposed quarry is mentioned only in O.Claud. IV 748.1, but there is, in my view, no reason to consider this anthroponym as a toponym. Ἐπίκωμος (Trajan) Five examples, the generic being always implicit, e.g., Ἐπικώμῳ (Ο.Claud. IV 776.3) and possibly δὸς εἰς τὸν Ἐπ[ίκωμον] in O.Claud. IV 817.1 (Ἐπί̣κ̣[ωμον] ed., a misreading due to a misplaced chip). Epikomos should probably be identified with quarry no. 7,70 where one can read the inscription Ἀπόλ(λων) Ἐπίκωμος71 (“Apollo presiding over the festivities”). 69. See p. 43. 70. Peacock and Maxfield 1997: 225. 71. I.Pan 45 = SEG XLVII 2122 (4), where the unfortunate resolution Ἀπολ(λώνιος) is corrected.

30

Rome in Egypt’s Eastern Desert

Figure 11. O.Claud. IV 841.11 (detail). © A. Bülow-Jacobsen

Ἐπιφανής? (c. 150) Attested only in Ο.Claud. IV 841.8, where the restoration λατ(ομίᾳ) Ἐπιφαν̣[ή]ς̣ is possible, but not certain. For the nominative after λατ(ομίᾳ) see s.n. Εὔπλοια. This name would be in the category of divine epicleses. Εὔπλοια (c. 150) I read λατο(μίᾳ) Εὐπλοίᾳ [ in Ο.Claud. IV 841.11 (Fig. 11, Ευπτικ̣[] ed.). We cannot exclude Εὐπλοία[ς], but it should be noted that the text presents two certain examples of λατομίᾳ followed by a specific in the nominative: λατ(ομίᾳ) Φιλοσέραπις and λατ(ομίᾳ) Κόχλαξ. The quarry-name “Good navigation” is not irrelevant, since the monoliths extracted came down the Nile and across the Mediterranean. Εὐτύχης (Trajan) Eight occurrences, all in the dative and without the generic. This is the anthroponym Εὐτύχης, gen. -ου, not the adjective εὐτυχής. Ζεύς (Trajan) One example: λατομίᾳ Διός (Ο.Claud. IV 638.1). Ἥρα (Trajan) The quarry of Hera appears in several listings of quarries in the dative (Ἥρᾳ) and in the genitive Ἥρας (depending on an implied λατομία) which serves as a header in the list of experts in O.Claud. IV 640. Also note the expression ἐκ τῆς Ἥρας (O.Claud. IV 743.4). This quarry is mentioned in the chart inv. 1538 (Chapter 11), along with another work-site called Κρηπ(ὶς) Ἥρας, perhaps the platform to allow loading of the blocks extracted from the Hera quarry onto chariots (on krepides, see pp. 42 f.). Ἱερώνυμος (Trajan) This name from a family of architects appears twice in a quarry name, as the modifier of a generic which is not λατομία, as read in the edition, but λουτήρ, also known by itself as a quarry name. – Ο.Claud. IV 710.3 (Fig. 12): λα̣τ̣ο̣μ(ίᾳ) Ἱερων(ύμου) ed. → Λουτ(ῆρι) Ἱερων(ύμου). – Ο.Claud. IV 779.2 (Fig. 13): λ̣α̣{υ}τ(ομίᾳ) Ἱερων(ύμου) ed. → Λουτ(ῆρι) Ἱερωνύμ(ου). To these two instances, one should add the Latin account Ο.Claud. IV 843, where the quarry is simply called Hieronymi. It is impossible to establish with certainty whether Λουτήρ, Ἱερώνυμος, and

A survey of place-names in the Egyptian Eastern Desert

31

Figure 12. O.Claud. IV 710.3 (detail). © A. Bülow-Jacobsen

Figure 13. O.Claud. IV 779.2 (detail). © A. Bülow-Jacobsen

Λουτὴρ Ἱερωνύμου are three names for one quarry; at least Λουτήρ (infra) could be an abbreviation for Λουτὴρ Ἱερωνύμου. Hieronymos may be identified with quarry no. 83, north of the cemetery: on one side the letters ιερω are engraved (Peacock and Maxfield 1997: 187 and fig. 6.57 = SEG XLII 1575). This case of an architect used as eponym of a quarry is unique. The name of the architect Herakleides appears on a quarry face in quarry no. 129, but is then preceded by διά.72 Καινὴ λατομία (Trajan) The namesake of the metallon at Umm Balad, it is attested among other latomiai names in four related lists of quarry names (O.Claud IV 700 and comm ad. 2; 702; 704; 777). Κάνωπος (Trajan) Five examples, including: – λατομ(ία) Κανόπο̣υ̣ (Ο.Claud. IV 641). – Κανώπου in Ο.Claud. IV 704 and 779, quarry lists, the names of which are in the dative. In these lists, anthropophoric quarry names are usually in the dative even if they are known elsewhere as λατομία + anthroponym in the genitive. However, I think we can read Κανώπῳ (Fig. 14) in the list Ο.Claud. IV 783.1 (Κανώπ⟨ο⟩υ [ ed.). 72. I.Pan 40, cf. O.Claud. I, p. 48; Peacock and Maxfield 1997: 189 and 221.

32

Rome in Egypt’s Eastern Desert

Figure 14. O.Claud. IV 783.1 (detail). © A. Bülow-Jacobsen

Κόχλαξ (Antoninus) Ο.Claud. IV 841.63, 64: λατ(ομίᾳ) Κόχλαξ; Ο.Claud. IV 842.5, 6: •Κόχλ[αξ; Ο.Claud. IV 843.3 (Trajan?): Cochlax. Κόχλαξ as a quarry name never inflects; this rare and onomatopoeic word means “gravel.” The Kochlax quarry, identified by an inscription with this name, is no. 120 in the inventory of D. Peacock. Λέων (Trajan and c. 150) The quarry is simply called Λέοντι in several lists of quarry names in the dative, but λ̣α̣τ̣(ομίᾳ) Λέοντος in Ο.Claud. 841 IV.40, which is a later text. Λουτήρ (Trajan) Five ostraca mention this quarry, which seems to take its name from the object, bath or basin, that was being (had been?) extracted. The place-name appears in the dative Λουτῆρι and in Ο.Claud. IV 814, preceded by the article (δὸς εἰς τὸν Λουτῆρα). It is probably the same quarry as that referred to as Λουτὴρ Ἱερωνύμου (supra), for, in the quarry lists O.Claud. IV 770 and 774, Λουτῆρι immediately precedes Μέσῃ, while in two lists pertaining to another series, Ο.Claud. IV 710 and 779, Μέσῃ immediately precedes Λουτῆρι Ἱερωνύμου. Μάρων (Trajan and c. 150) As with the Leon quarry, the anthroponym refers to the quarry itself and not, as it seems, to a person with a relationship to it: many occurrences of Μάρωνι appear in lists of quarry names in the dative, but this feature was forgotten by the time of Ο.Claud. IV 841.41, where one has λ̣ατ(ομίᾳ) Μάρωνος distinct undoubtedly from λατ(ομία) Μάρωνος Μακρά on the same ostracon (l. 31). Μάρωνος Μακρά (c. 150) The “Long quarry of Maron” is only attested in Ο.Claud. IV 841.31. The edition gives the impression that it belonged to the caesura of Epaphroditos, which appears as a header to a text block on line 30. This is curious, given that Maron’s quarry is in the caesura of Enkolpios according to Ο.Claud. IV 841.41. However, the infrared image shows that κοπ(ῇ) Ἐπαφροδ(ίτου) on line 30 is intentionally deleted (Fig. 15).73 73. As the beginning of line 31, which should be in double straight brackets: ⟦ν̅ ι̣β⟧̣ .

A survey of place-names in the Egyptian Eastern Desert

33

Figure 15. O.Claud. IV 841.30–33. © A. Bülow-Jacobsen

Μεγάλη λατομία (Trajan and c. 150) Just as with Καινή, but not Μέση, the generic element λατομία is always expressed when qualified by Μεγάλη (see Ο.Claud. IV 782.2–3). Two occurrences under Trajan, then in O.Claud. IV 841.65. Μέση (Trajan and c. 150) Except once, in the opening words of a staff list, Ο.Claud. IV 644 (Μέσῃ{ς} λατομίᾳ), the “Middle” quarry, frequently mentioned, is referred to simply as Μέση, unlike Καινή or Μεγάλη which are always followed by the generic (e.g., Ο.Claud. IV 812: δὸς εἰς τὴν Μέσην σφυρίδας δέκα). The later account Ο.Claud. IV 841.54 suggests that this important quarry could have been placed under the protection of Isis: λατ(ομίᾳ) Μέσῃ Εἴσειδι. It is abnormal that the theonym after λατομία is not in the genitive. Μίθρας (Trajan) Attested only in the incipit of a headcount written in charcoal, Ο.Claud. IV 646: λατομίᾳ Μίθρα. Other examples of λατομία + name of god in the genitive suggest the interpretation of Μίθρα as a genitive, not as a dative in apposition. Μυρισμός (Trajan [and Antoninus Pius?]) The four certain examples of the Myrismos quarry (Peacock no. 22) date to Trajan. The quarry could have functioned under Antoninus Pius if we accept the reading ex lat(omia) Ṃuṛ[ in the Latin account Ο.Claud. IV 845. The name of this quarry appears in three ostraca abbreviated as Μυρισμ( ). It is preceded by the generic in the inscription that we found in the rubble of the quarry, after its destruction by the Zamzam Company in 1989 (Fig. 16). Jean Bingen published the Greek text of the inscription and translated it “The (column) no. 3 of the Myrismos quarry. The one who loves Trajan.”74 The epithet Φιλοτραιανός, because it is in the nominative, cannot be a second name of the λατομία, hence Bingen’s idea of linking it to an implied στῦλος (column). Bingen considers that Myrismos, the eponym of the quarry, is “perhaps an imperial slave or a Greek contractor.”

74. BIFAO 93 (1993) 64 f. = SEG XLIII 1121.

34

Rome in Egypt’s Eastern Desert

Figure 16. Column base inscribed on the underside, from the quarry of Myrismos. © A. Bülow-Jacobsen

Νειλάμμων (Trajan or Hadrian) This quarry is mentioned only in five headcounts written by the same scribe, who invariably expresses the generic and dissimilates the first μ to ν: λατομίᾳ Ν(ε)ιλάνμωνος (O.Claud. IV 734–738). This spelling can be explained as a fault of hypercorrection and, perhaps, through the novelty of this polytheophoric name: as a personal name, Neilammon does not spread in the papyri before the end of the second century AD (the earliest example, P.Oxy. III 477, is dated to 132/133). At Mons Claudianus, it is difficult to decide if it is the well attested personal name, or the name of the god Neilammon, a late deity in Elephantine, known through a proskynema on the terrace of Khnoum’s temple at Elephantine (SEG XXIX 1637 = I.ThSy. 277), as well as in several unpublished ostraca from Syene-Elephantine, one of them dating to 193 (R. Duttenhöfer, pers. comm.). The theonym may occur also in the temple inventory P.Erl. 21.32 (prov. unknown, c. 195): N]ειλάμμων̣ θεοῦ μ̣( ), which one might consider turning into N]ειλάμμων̣(ος) θεοῦ μ̣(εγίστου). Νερωνιανή (Trajan) Two instances, λατομία always being implied (Ο.Claud. IV 776.16; 777.2). There is also the beginning of this toponym, but erased, in Ο.Claud. IV 841.41, of c. 150. Νικοτύχαι75 (Trajan) The name of this quarry is written in scriptio plena on a granite bed that was cleared of overburden, but which was not worked (Fig. 17, Fig. 18). We probably recognize it in two ostraca: λατ]ομ(ίᾳ) 75. Chapter 8 (= SEG XLII 1576).

A survey of place-names in the Egyptian Eastern Desert

Figure 17. Nikotychai quarry, remained untapped. © A. Bülow-Jacobsen

Fig. 18. The signpost of Nikotychai quarry, with the name of the foreman. © A. Bülow-Jacobsen

35

36

Rome in Egypt’s Eastern Desert

Figure 19. O.Claud. IV 747.6: [Νικ]οτυχ( ) ἀκισκ(λάριοι) β. © A. Bülow-Jacobsen

Νικο̣τ̣(υχῶν) (Ο.Claud. IV 651.1) and Νικ]οτυχ( ) (Ο.Claud. IV 747.6, Μάρ]ονα ed., see Fig. 19). The polytheophoric name Νικοτύχη, otherwise unknown in Egypt, is attested as an anthroponym (with a masculine counterpart Νικότυχος). Πλωτίνα (Hadrian?) This name appears only in Ο.Claud. IV 739, a note assigning 16 men “to Plotina’s column,” στύλῳ Πλωτίνα[ς?] (Fig. 20). Does that mean they were sent to work specifically on a column in a quarry called Plotina, or was this quarry, as for Λουτήρ, named after the object that was extracted, as would happen with a column which would itself have a name, like the column Φιλοτραιανός in Myrismos quarry? The parallel offered by Ο.Claud. IV 657.1–2 (infra, s.n. Χρησμοσάραπις) suggests that we should rather consider Plotina as the name of the quarry.

Figure 20. Ο.Claud. IV 739.1. © A. Bülow-Jacobsen

Πορφυρίτης (Trajan) This name appears in two lists among quarry names (O.Claud. IV 705 and 706) and in a delivery order belonging to a series of notes about sending quarry equipment (O.Claud. I 17, cf. O.Claud. IV, p 135). It must, therefore, probably refer to a λατομία which is the namesake of the metallon. The restitution of this quarry name in the Antonine ostraca O.Claud. IV 841.49 and 842.4 is very hazardous. Ῥώμη (Trajan) It more likely refers to the goddess Roma than to the city. Ῥώμη is never accompanied by the generic λατομία. Sometimes, it is preceded by the article after a preposition ἐν τῇ Ῥώ[μῃ] (Ο.Claud. IV 652), δὸς εἰς τὴν Ῥώμην (Ο.Claud. IV 804; 813). In Ο.Claud. IV 742.1, one should probably read, after ἰς Ῥώμην, οὐί̣γ̣λ̣η̣ς (Ουι̣ϲ̣μ̣ο̣ϲ̣ ed.): see Fig. 21. Σελήνη (c. 150) This quarry appears only in two ostraca where its name had been misread or wrongly resolved: – ⟦λατ(ομίᾳ) Σελήνη̣ς⟧̣ (Σερ̣ηνου ed.): O.Claud. IV 841.51 (Fig. 22). – Σελήν[η(ς?) (Σελην[ου ed.): O.Claud. IV 842.2.

A survey of place-names in the Egyptian Eastern Desert

37

Figure 21. O.Claud. IV 742.1. © A. Bülow-Jacobsen

Figure 22. O.Claud. IV 841.51. © A. Bülow-Jacobsen

Σέραπις? (c. 150) λατ(ομίᾳ) Σερα[, Ο.Claud. IV 84.24: a single and uncertain example. Σῴζουσα (Trajan) Evidenced only in a Latin fragment (ex latomia Sozusa, O.Claud. IV 846.3). This epiclesis is not common in Egypt: we only know a cult of Isis Sozousa at Ekregma in northern Sinai (P.Oxy. XI 1380.76), and an Arsinoe Sozousa Street in Alexandria. Τραιανή (Trajan) Common, this name is always used alone except in O.Claud. IV 653.1–2 (λατομίας Τραϊανῆς). Φιλάμμων (Trajan) Philammon appears, every time more or less abbreviated and without the generic, in three lists of quarry names in the dative (Ο.Claud. IV 775.10; 776.7; 777.3). W. Swinnen has demonstrated that Philammon was an old Greek anthroponym, probably a hypocoristic of φιλάμενος.76 Widespread in Cyrenaica because of a homophony with the Libyan names in -άμ(μ)ων and with the name of the local god Ammon, it is also well attested first in Ptolemaic Egypt, then in Roman Egypt, where it benefited from the second century AD fashion for anthroponyms ending in -άμμων. Swinnen thinks that the popular etymology associating Φιλάμμων with the theonym Ἄμμων is late, since genealogies of the type Philammon son of Ammonios do not occur before the end of the first century BC (BGU IV 1163.3 [16–13 BC]; SEG XXVI 1839, col B.13 [first century BC–beginning of the first century AD]); secondly, Fr. Dunand, as Swinnen reminds us, has established that there are no theophoric names -άμμων in the Ptolemaic era. How did popular etymology understand the name Philammon? We know that in Greek compounds, the element φιλο- has an active sense (“who loves”) when in first position, but passive (“beloved”) in second; however, this rule does not apply strictly to anthroponyms (E. Schwyzer, Griechische Grammatik, [Munich 1939] 635: Agelaos, “leading the people,” has the same meaning as Laagos). This latter consideration would allow us to avoid the aporia of a pagan name meaning “who loves Ammon,” the concept of loving a god being deemed foreign to pagan thought: so, Philotheos, so widespread in the Christian 76. Swinnen 1968, esp. 260.

38

Rome in Egypt’s Eastern Desert

period where it means “loving God” would be in pagan times (where it is also rare), the equivalent of Theophilos (“beloved by the god”). But how should the name Φιλάμμων be understood when bestowed on a quarry in the time of Trajan? The meaning “Beloved of Ammon” would be satisfactory for a quarry, presented as being under the protection of the god, but the absence of the equivalent *Ἀμμωνόφιλος makes this hypothesis weak. We can compare Φιλάμμων to another similar quarry name, Φιλοσάραπις, which has a clear etymology. It was originally a title that falls into the category of loyalist epithets from the imperial era as in φιλο- + name of a distinguished person (e.g., φιλόκαισαρ). In this case, there is no doubt that φιλο- has an active sense. But is it the same when the second element is a theonym? M. Malaise, rallying to the position of Swinnen, considers that Φιλοσάραπις means “Who is beloved by Serapis.”77 But would the element φιλο- have two different meanings when, in 193, the archiereus Ulpius Serenianus tacked onto his name the epithets φιλοκόμμοδος καὶ φιλοσάραπις?78 I do not think so. Similarly, in Ephesus Vibius Salutaris, an official of equestrian rank who had been generous with the temple of Artemis, is honored with the title philartemis, which is added to that of philocaesar.79 In an inscription from Didyma, the philodionysoi who question the oracle are an association of “Friends of Dionysus.”80 Conversely, would these compounds φιλο- + theonym have taken a passive sense when employed as anthroponyms, which seems to have been attested (except for Φιλοσάραπις) only in Egypt (where we also encounter the rare Φιλαπόλλων, Φιλέρμης, Φιλοδιόσκορος and, more frequently, Φιλαντίνοος)? Swinnen considers Φιλαπόλλων and Φιλέρμης to be Hellenized counterparts of Μαίευρις and Μαιθωυτ/Μαιθώτης, which he interprets as transposing mry + divine name, so “Beloved of Horus/Thoth.” But this etymology, proposed by Vergote, is not authoritative; it has not been retained in the Demotisches Namenbuch, which assigns another etymology to these names (mȝʿ-, “Truthful Horus/Thoth”).81 The names Φιλαπόλλων, Φιλέρμης, Φιλοδιόσκορος, etc., therefore, rather express devotion to a pagan god, and became popular because of the fashion of the imperial period for loyalist epithets. Although semantically it seems more satisfactory to think that the quarry names Philammon and Philosarapis meant “Beloved of Amon, of Sarapis,” who consequently, watched over the success of the work, I think that these names meant “devotee of ” these gods. They are paralleled by the quarry name Philoc(aesar?) and the column Philotraianos (supra, s.n. Μυρισμός). Philoc( ) The name of this quarry (Peacock Nos. 11–13) is known only by a dipinto written in red ink under the base of the giant column which remains in place at the head of the Pillar-Wadi. It is so faded that we had never been able to read it, until 14 January 2017, when a visit to Mons Claudianus gave us the opportunity to take a digital photo that was then treated with the software DStretch (Fig. 23). The dipinto is formulated on the same model as the inscription at the Harpocrates quarry, which also belongs to the caesura of Epaphroditos. [c(aesura)]? Ep(aphroditi) [e]x lat(omia) [P]ḥiloc( ) 77. Malaise 1972: 44, n. 4 and 442. 78. SB XIV 11342.6; SB XXVI 16726.2. 79. IK XI.1, 33.4–5. 80. A. Rehm, Didyma, II. Die Inschriften, Berlin 1958, no. 502. 81. I thank Marie-Pierre Chaufray, Willy Clarysse, and Françoise Dunand for helping me unravel this tangled question.

A survey of place-names in the Egyptian Eastern Desert

39

Figure 23. Dipinto under the base of the giant column. The bad light forced us to cobble together a makeshift umbrella. © A. Bülow-Jacobsen

The reconstruction of c(aesura) is derived from the offset position of Ep(aphroditi). No ostracon gives us a clue for completing the quarry name: was it a loyalist epithet (Philocaesar rather than Philocommodus, since no major works at Mons Claudianus can be placed under Commodus)? Or just a personal name? Φιλοσάραπις (Trajan and c. 150) Φιλοσάραπις (16 occurrences of this quarry) is never used with λατομία except in Ο.Claud. IV 841.47 (c. 150), a block count where the scribe systematically uses the generic, and which reads λατ( ) Φιλοσά̣ραπις. Philosarapis, which is originally a kind of honorific title (cf. p. 38), is attested as a personal name from the end of the second century AD, hence after the Trajanic occurrences of our toponym. If Philosarapis referred to a person, the scribe would have written λατ( ) Φιλοσα̣ράπιδος, as he writes elsewhere λατ( ) Μάρωνος. Nevertheless, λατ(ομίᾳ) should be restored, in this count, in the dative, judging by λατ(ομίᾳ) τῇ αὐτ̣[ῇ] and λατ(ομίᾳ) Μέσῃ Εἴσιδι in lines 20 and 54. If Φιλοσά̣ραπις is the title of the quarry, one should have Φιλοσαράπιδι (but see λατ(ομίᾳ) Κόχλαξ, lines 63 and 64). I find it difficult, however, to explain the masculine article, instead of the feminine, in Ο.Claud. IV 810.1: δὸς εἰς τὸν Φιλοσά̣ραπ(ιν)? Because Φιλοσάραπις should be feminine as well as masculine, like φιλόπατρις. In fact, it behaves as if it were a theonym.

40

Rome in Egypt’s Eastern Desert

Figure 24. O.Claud. IV 658.1 (detail). © A. Bülow-Jacobsen

There was probably another quarry with the same name when work resumed at the end of the second century. In a letter to the procurator metallorum, the quarry workers report that, not knowing the official name of the quarry in which they work, they have on their own initiative called it “Philoserapis” (Ο.Claud. IV 853.19 [c. 186/7]). Χρησμοσάραπις (Trajan) – στύλῳ Χρη⸌σ⸍μοσεράπιδος: Ο.Claud. IV 657.1–2. – λατομ̣[ίᾳ] Χρη⸌σ⸍σ̣(μο?)σαρα[πιδ-]: Ο.Claud. IV 658.1 (λατο(μίᾳ) Χρη̣σμ̣(ο)σαρ( ) ed.). The infrared photo reveals, under the quarry name, an erased line which I cannot read, and does not help to understand the tangled traces that we would like to read as -μο- (Fig. 24). – δὸς εἰς τὸν Χ̣ρε̣σ̣μ̣ο̣σ̣ά̣ρ̣(απιν): Ο.Claud. IV 811.1. – Χρησμοσαρ(άπιδι): chart for the distribution of water O. Claud. inv. 1538.1 and 4 (Chapter 11). Loewe (1936) does not cite any compound theophoric toponym,82 but we can compare Χρησμοσάραπις to another quarry name likewise made from two nouns, both of which, however, are proprial: Νικοτύχαι. The order of the elements implies that χρησμο- is the modifier, so that this quarry name should be understood as “Sarapis of the oracle/oracles.” Is this neologism a quick way of expressing the concept of Σάραπις χρησμοδότης? Is it an “anecdotal”83 toponym, reminiscent of a prophetic dream sent by Sarapis to some foreman?

82. I thank Claire Le Feuvre and Sophie Minon for providing this citation. 83. On the distinction between memorial and anecdotal toponyms, see Dorion and Poirier 1975, s.v. “anecdotique.”

A survey of place-names in the Egyptian Eastern Desert

41

2. Conclusion on the toponymy of quarries Table 1.4. Taxonomy of the quarry-names in Mons Claudianus Dynastic Αὐγο(ύστη) Νερωνιανή Πλωτίνα Τραιανή Philoc(aesar)?

Theophoric Ἄμμων? Ἄνουβις Ἄπις Ἀπόλλων Ἁρποκράτης Ἀφροδίτη Διόνυσος Ἐπίκωμος Ἐπιφανής? Εὔπλοια Ζεύς Ἥρα Ἴσις (Μέση Ἴσις) Μίθρας Νικοτύχαι Ῥώμη Σελήνη Σέραπις? Σῴζουσα Χρησμοσάραπις

Anthropophoric Βάρβαρος Εὐτύχης Ἱερώνυμος Κάνωπος Λέων Μάρων Μυρισμός Νειλάμμων Philoc( )?

Descriptive Καινὴ λατομία Κόχλαξ Λουτήρ Λουτ(ὴρ?) Ἱερωνύμου Μεγάλη λατομία Μέση Μέση Ἴσις Μάρωνος Mακρά Πορφυρίτης

epithets of devotion (?): Φιλάμμων Φιλοσάραπις (2 quarries)

Most often theophoric, the names of latomiai draw on the same sources as boat names, usually theophoric and allegorical, as highlighted by P. Heilporn in his comment to P.Bingen 77, p. 343 f. (second cent. AD). Three merchant ships in this papyrus have the same names as some of our quarries: Zeus, Aphrodite, and Selene. See also the ship called Ἀντίνοος Φιλοσάραπις Σῴζων in SB XIV 11850 (149). The intention was to put the stone work under the protection of a god whose goodwill was solicited by calling the quarry “devotee of (such-and-such god).” The extraction of monoliths was at the mercy of unexpected weaknesses of the material, undetectable until it was too late, as evidenced by the basins and broken or split columns that were abandoned on site. In the letters in which they announce to the procurator metallorum the good news (ἱλαρὰ φάσις) of the completion of an order, the quarry workers are quick to attribute their success to Sarapis, assisted by the Tyche of Claudianus and the baraka (which they also call Tyche) of the procurator.84 The generic λατομία is only expressed when the specific is an appellative adjective; Λατομία then follows it: Μεγάλη λατομία, Καινὴ λατομία. Unlike the element πόλις always implied in the city name Καινή,85 λατομία is never omitted in these two quarry names. Μέση is an exception: there is only one known example of Μέση λατομία, and this quarry is otherwise mentioned only as Μέση. While Loewe (1936) notes no specific element of theophoric toponyms that is a compound (it is always a pure theonym or a derivative), we see several compounds among the theophoric quarry names: besides the common Philammon or Philosarapis, also known as anthroponyms, we note the polytheophoric Νειλάμμων and Νικοτύχαι—the latter being new in Egyptian documents—and the hapax Χρησμοσάραπις. We cannot, however, totally exclude the possibility that Philammon, Philosarapis, and Nilammon are anthroponyms referring, as do Leon, Hieronymos, or Myrismos, to real people, supervisors with whatever status on the working site. 84. O.Claud. IV 850, 853, and 857. 85. Except in the Anonymous of Ravenna (Caenopoli).

42

Rome in Egypt’s Eastern Desert

Whether drawn from a god or a human, quarry-names follow the same syntactic pattern. The theonym or anthroponym is in the genitive preceded by λατομία (e.g., λατομία Ἀπόλλωνος), but, when it happens to be used alone, instead of staying in the genitive (as in the case of another microtoponym at Mons Claudianus, the well of Cattius, designated as ὕδρευμα Καττίου or simply Καττίου or τὸ Καττίου), it takes the case that would be that of λατομία: this is no longer the name of a god or individual, but that of the quarry. Thus, Λέων, Μάρων, Διόνυσος, Ἀπόλλων, and Ἄνουβις are quarries. They are found in the dative, mixed with common nouns designating places, such as στομωτηρίῳ in the distribution lists of O.Claud. IV 769–787; in the series of delivery orders 804–819, those names, despite the general trend in Greek towards the omission of the article after a preposition,86 are systematically preceded by the article (e.g., δὸς εἰς τὸν Διόνυσον), either because the sequence εἰς (locative) + anthroponym/theonym would have sounded strange or because it is customary, according to Mayser, for the Lokalnamen.87 It is rare that, used after λατομία, the theonym/anthroponym, instead of being a complement to the name in the genitive, is in apposition: the only clear example is in Latin latomia Sozusa, the two Greek examples being ambiguous (Εὔπλοια, Μίθρα). In general, it is unclear who the eponymous persons of the quarries were. In the case of the Latin account in O.Claud. IV 843, where Hieronymi is an entry along the same lines as Cochlax, we can say that the quarry took the name of the architect Hieronymos. But is the quarry “of Hieronymos” the same as the one called Λουτήρ—if indeed Λουτήρ is the short form of Λουτ(ὴρ) Ἱερωνύμου? Other anthroponyms used as quarry names do not overlap the prosopography of the managerial staff. It is possible that these Maron, Myrismos, and Leon were ergodotai (foremen) whose working sites did not have a name—unlike the Nikotychai Quarry, equipped with a sign announcing both its name and that of its foreman Sarapion. The late ostracon Ο.Claud. IV 841 differs from earlier texts in that the generic λατομία, so often omitted otherwise, is systematically employed in front of the specific element. It is also the only document where we note a double specific: λατ(ομίᾳ) Μέσῃ Εἴσειδι, λατ(ομία) Μάρωνος Μακρά.

3. On the fringe of quarries: krepides and kopai In several lists of λατομίαι there are also names of κρηπῖδες, some of which are called after a quarry. Documents such as O.Claud. IV 872 and 880 strongly suggest that these “quays” are platforms for putting blocks on wagons, as can be seen at the bottom of the Pillar-Wadi (Fig. 25). In O.Claud. IV 769, 770, 774, 775, and 776, κρηπῖδ(ι) appears, clearly without further specification, in lists of microtoponyms in the dative which are generally quarries, but not exclusively, since there is also στομωτηρίῳ (steelworks, O.Claud. IV 776.12). It could be what some toponomasticians call “appellative,”88 giving the word the restricted sense of a generic used to designate a specific place, in other words, a common categorizing noun used as a toponym: in this case, “the Quay.” In the charts of water distribution, κρηπῖδι is, however, accompanied by a determinant. In the best preserved of these charts, inv. 1538 (Chapter 11), we find, among the latomiai and other microsites which received rations of water, κρηπῖδι Μεγάλ(ῃ) and κρηπῖδι Ἥρας.

86. Mayser, Grammatik II.2.1, 14. 87. Mayser, Grammatik II.2.1, 17 f. 88. In this they follow the recommendations of the Second United Nations Conference on the Standardization of Geographical Names (see Dorion and Poirier 1975: 55).

A survey of place-names in the Egyptian Eastern Desert

43

Figure 25. Mons Claudianus. Column shafts on the krepis at the bottom of the Pillar-Wadi. © A. Bülow-Jacobsen

The term κοπή appears late in the ostraca of Mons Claudianus: four documents mention it, dated to c. 150 according to stratigraphy and prosopography. A. Bülow-Jacobsen convincingly suggests that it is a calque of caesura, a term that recurs in quarry-marks at Dokimeion.89 The kopai are named after an important figure: the emperor (κοπὴ κυριακή: O.Claud. IV 841.9), the procurator sc. metallorum (ἐπίτροπον… τῆς αὐτοῦ κοπῆς: O.Claud. IV 885.6–7), Epaphroditos (κοπὴ Ἐπαφροδίτου: O.Claud. IV 841.12, 30, O.Claud. inv. 7134), Enkolpios (κοπὴ Ἐνκολπίου: O.Claud. IV 841.26, 41, O.Claud. IV 896.7). We can probably identify the κοπὴ Ἐπαφροδίτου in the Latin mark CEP [= c(aesura) Ep(afroditi)] present at many quarries. The large account O.Claud. IV 841 shows the latomiai as subdivisions of kopai. It is unclear whether kopai corresponded to geographic areas of the metallon, or whether they were merely administrative categories, the meaning of which escapes us. Although it does not appear in the Trajanic ostraca, the system of the kopai must have existed before 150, since Enkolpios is known as procurator metallorum under Trajan (I.Pan 38); concerning Epaphroditos—whose name is more common—he is probably an imperial slave who was conductor metallorum in the period when work in Claudianus and Porphyrites resumed under Hadrian, after the Jewish revolt of 115–118 (I.Pan 21 and 42).

89. O.Claud. IV 841, introduction.

44

Rome in Egypt’s Eastern Desert

V. Forts and fortlets (praesidia) 1. The toponyms on the amphora of the Barbarians (O.Krok. I 87) O.Krok. I 87 is an amphora which the curator of Krokodilo used in year 2 of Hadrian’s reign to copy circulars passing through his hands. Many of them mention a mysterious Parembole,90 as well as several praesidia with names that are not otherwise attested on ostraca in the desert. These sites are probably outside Mons Berenicidis. The toponymic character of Παρεμβολή is not above suspicion, since it could be a common noun meaning “camp” (= Latin castra). Two circulars recount military incidents at two praesidia which depend on this π/Παρεμβολή: Patkoua and Thonis Megale. Παρεμβολή appears each time in a formula in which the author of the circular, the centurion Cassius Victor, says that he attaches the copy of reports sent from these two praesidia and that have come to him εἰς παρεμβολήν. The absence of the article before παρεμβολήν is not a reliable indicator for deciding if this word is a toponym or a common noun. However, two good arguments suggest that παρεμβολήν does not simply mean “camp.” First, Cassius Victor addresses himself to a number of recipients scattered over a wide area that exceeds the limits of the Berenike Desert (since it is assumed that his correspondence will reach several prefects). Second, the language is detailed and codified, as shown by the care with which he states his identification and that of his informants: ἐπάρχοις, (ἑκατοντάρχαις), (δεκαδάρχαις), δουπλικαρείοις, σησκουπλικαρείοις, κουράτορσι πραισιδ[ί]ων κατʼ ὠρεινῆς Κάσσ̣[ειο]ς Βίκτωρ (ἑκατοντάρχης) [σπ]είρης β {δευτέρας} Εἰτουρ̣αί̣ ω ̣ ν̣ ·̣ ἀντείγραφ[ο]ν διπλώματος πεμφθέντος μοι εἰς Π̣αρ̣ ε̣ νβ̣ολ̣ ὴ ̣ ν κτλ. (O.Krok. 87.109–111), “To prefects, centurions, decurions, duplicarii, sesquiplicarii, curators of the praesidia of the desert, Cassius Victor, centurion of cohort II {Second} of the Itureans. Copy of a diploma sent to me at Parembole etc.” If it is a place-name, as I think, Παρεμβολή is what toponymists call an “appellative.”91 In O.Krok. I, pp. 139–41, I suggested identifying this Παρεμβολή as Parembole, which is the first fortified Roman site south of Syene according to the Antonine Itinerary (161.2). We have seen that the two praesidia which depend on the castra of Parembole were called Πατκουα and Θῶνις Μεγάλη. The first is probably an Egyptian word.92 The second hybrid combines an Egyptian noun (Θῶνις = “The Lake, The Pond”) and a Greek appellative adjective. O.Krok. 87.68 shows a third name for a praesidium, Νιτρίαι: one of the recopied circulars emanates from a prefect,93 Cassius Taurus, and introduces a report from the κουράτωρ πραισι[δίου] Νειτρειῶν. The fortlet perhaps owed its name to the neighbouring natron deposits of Laton polis (O.Krok. I, p. 142).

2. Praesidia on the route to Myos Hormos Κροκοδιλώ, Κορκοδιλώ (Al-Muwayh, [25° 56’ 33” N/33° 24’ 04” E]) The praesidium of Krokodilo takes its name from the profile of the rocky hill which overlooks it, when viewed from the northeast. The resemblance so struck a traveller that he represented the hill in the form of a crocodile with a rounded snout in a rock graffito in the area (Fig. 26 and Fig. 27).94 90. Lines 19 and 111. 91. Cf. p. 42. 92. John Rea (per litt.) observed that a reading Παγκουα was not excluded. 93. But probably not the prefect of Berenike, who seems to be at this time Arruntius Agrippinus (O.Krok. I, pp. 137 ff.). 94. It is only after seeing the graffito, after a month of excavation, that I made the connection between the shape of the rock and the curious name of the praesidium.

A survey of place-names in the Egyptian Eastern Desert

Figure 26. The rock of Krokodilo seen from the northeast. © H. Cuvigny

Figure 27. A rock graffito on a nearby cliff, probably inspired by the shape of the hill. © H. Cuvigny

45

46

Rome in Egypt’s Eastern Desert

Κροκοδιλώ seems to fit in the series of place-names in -ώ well known in Egypt, but they are usually late recharacterizations from the Byzantine and Arabic periods, such as Λύκων πόλις becoming Λυκώ or Ἀφροδίτη, Ἀφροδιτώ.95 Here the toponym was created ex nihilo in this form, and, moreover, in the early empire, since Krokodilo was founded at the end of the first century AD, probably under Vespasian. Contemporaries were themselves confused, since the form is spelled twice Κορκοδίλων (O.Krok. I 18.5: εἰς Κορκοδίλων; O.Krok. I 78: κουράτο̣ρ̣ι̣ [πραισιδίου Κορκοδί]λ̣ων). Κροκοδιλώ is actually a case of a popular formation in -ώ, about which the words of P. Chantraine96 are illuminating: “the suffix was used mainly to provide derivatives of nouns. Sometimes, it only serves to characterize the word as feminine: ἀνθρωπώ [= ‘woman’ in Laconian dialect], but it is most common in nicknames like μορφώ [‘beautiful,’ the name of Aphrodite in Sparta] or in words that designate a person or an animal feared or despised: ἀκκώ [= old woman, scarecrow], μιμώ [= female monkey, hag].” This expressive suffix reveals that the toponym belonged to popular speech. Such a whimsical appellation is atypical for a name given by the Romans. Πέρσου97 In the epigraphic proskynemata from the Paneion in the Wadi al-Hammamat, this toponym appears firstly in demotic (prs) at the beginning of the Ptolemaic period, then in Greek under the early JulioClaudian emperors. At that time, Persou is the name of the Wadi al-Hammamat quarries (called “Rohanou Valley” in Pharaonic times), as proven beyond doubt by the proskynemata “in front of the gods of Persou” engraved on the entrance of the small sanctuary of the village situated opposite the Paneion (25° 59’ 25” N/33° 34’ 12” E). Three dated proskynemata mention Persou: Kayser 1993, no. 4 = SB XXII 15642 from year 18 of Tiberius’s reign (AD 32); Kayser 1993, no. 15 = SB XXII 15655, probably from year 10 of the same emperor (lacunose titulature); Kayser 1993, no. 7 = SB XXII 15645, dated year 9 without the emperor’s name, which Fr. Kayser suggests would not be that of Tiberius, but Nero (62), due to the position of the graffito on the doorpost. In the Paneion facing the village, quarry workers carved several Demotic proskynemata, and one in Greek, in which they claimed to be “from Persou and Tamostymis” (σκληρουργὸς ἐκ Πέρσου καὶ Ταμοστύμεως, I.KoKo. 105), which drove me to the hypothesis of a complex toponym consisting of a correlated noun phrase (although this structure is mainly attested in Egypt for the names of kleroi). Πέρσου would have been the area of bekhen quarries (Wadi al-Hammamat) and Ταμόστυμις (assuming the name inflects), the mines of Wadi al-Fawakhir.98 Ταμόστυμις is not attested elsewhere, other than as epiklesis of Isis (Isis Ταμεστομε) in a lost stela from Qusayr mentioned by Adοlphe Reinach and dating back to year 25 of Augustus’s reign (6–5 BC).99 The ostraca from Krokodilo and Maximianon contain many references to a praesidium of Persou, where Athena was honored, and which had a vegetable garden that provisioned these two sites. This praesidium, close to an abundant well, cannot be the village of Wadi al-Hammamat, which does not have a well and which is not a praesidium: it must be near Biʾr Umm Fawakhir, located some 5 km from the Paneion, but it has been completely destroyed. I deduced that the toponym Persou migrated, and I distinguished between Persou I (the village opposite the Paneion cave of Wadi al-Hammamat) and Persou II (the praesidium next to Biʾr Umm Fawakhir, attested at least from the time of Trajan). An 95. J. Gascou, JJP 24 (1994) 14 n. 4; J.-L. Fournet, REG 105 (1992) 236; Fournet 2002: 52 f. 96. Chantraine 1933: 116. 97. Cuvigny (ed.) 2006: I 55; II 281 f., 383. 98. Thissen (1979: 88) sees in this toponym the Egyptian name for galena. 99. Reinach 1910: 43.

A survey of place-names in the Egyptian Eastern Desert

47

ostracon found in Wadi al-Fawakhir and dating from the reign of Caligula or Claudius100 reveals the presence of a Roman military post earlier than the ostraca from Maximianon and Krokodilo, but we cannot tell if the name was already Persou or still Tamostymis (if that name was indeed attached to the Wadi al-Fawakhir). The oldest papyrological reference to Persou predates the ostraca from Maximianon and Krokodilo. This is a receipt on an ostracon belonging to Nikanor’s archive found at Koptos, O.Petrie Mus. 112. It was issued in year 2 of Nero’s reign (55/56) by a κουράτωρ Πέρσου. It is difficult to decide where the soldier was stationed: still in the village next to the Paneion of Wadi al-Hammamat, where the last proskynemata engraved on the door posts of the chapel door date to Nero,101 or already in the Wadi al-Fawakhir? The migration of the name Πέρσου from the Wadi al-Hammamat (Persou I) to Wadi al-Fawakhir (Persou II) may have occurred when the village of Wadi al-Hammamat was abandoned: it is a common fact that a toponym migrates from a disused site to a nearby one, and it is not possible for two adjacent sites to have the same name at the same time.102 Unless, as J.-P. Brun thinks, the name Tamostymis had fallen into disuse and the complex toponym Πέρσου καὶ Ταμοστύμις was simplified from the Ptolemaic period to Πέρσου to signify the combination that might be considered a metallon, formed by the two sites. It is unclear whether Πέρσου, an indeclinable genitive of Πέρσης, is to be understood as the ethnic or as the derived anthroponym. Μαξιμιανόν (Al-Zarqaʾ [26° 00’ 03” N/33° 47’ 15” E]) A proprial adjective derived from the anthroponym Maximus. This name probably commemorates an important figure. One naturally thinks of someone occupying a role comparable to that of the prefect of Egypt Iulius Ursus, who ordered the construction of forts on the Berenike road during Vespasian’s reign. Two candidates come under consideration: C. Magius Maximus, prefect in year 1 of Tiberius’s reign (14/15), and L. Laberius Maximus, prefect in 83 (year 2 of Domitian’s reign). In the first case, Maximianon would have already been the name of the Julio-Claudian military post, traces of which were found under the dump. In the second case, it would mean that the present praesidium was built a few years after the Berenike road was developed, in 79. The first hypothesis is the more attractive, since it is natural that a military post, when rebuilt, would retain its name (cf. for instance the praesidium of Apollonos Hydreuma, which kept the name of the station that predates the praesidium). Σιμίου (Biʾr Sayyala? [26° 07’ 29” N/33° 55’ 50” E]) Simiou is probably an anthropotoponym, a genitive of the Greek anthroponym Σιμίας, which has a variant with an expressive geminate, Σιμμίας.103 It is the only ancient name for a praesidium that we know for the section of road between Maximianon and Myos Hormos. Thus, there are three candidates for identification: Al-Hamraʾ, Biʾr Sayyala, and Dawwi. O.Max. inv. 920, dating from c. 90–125 (according to stratigraphy), clearly shows that the praesidia of Persou, Maximianon, and Simiou directly followed from west to east when the ostracon was written. Dawwi, which is farthest from Maximianon and whose construction date falls within the final phases of Maximianon, can be ruled out.104 Al-Hamraʾ (26° 02’ 18” N/33°53’ 29” E) is today the immediate neighbor of Maximianon in the direction of Myos 100. Guéraud 1942: no. 14 (= SB VI 9017). Cf. Cuvigny (ed.) 2006: I 196. 101. This does not necessarily mean that the soldiers were still stationed in the village: the proskynemata may have been left by travellers. Ostraca found in the village indicate the presence of a mixed population of quarry workers and soldiers, but none are dated (Kayser 1993, nos. 20–60 = SB XXII 15660–15700). 102. I owe this remark to Herbert Verreth. 103. In theory, it could also be the genitive of Σίμιος, name of a Syrian god (Cuvigny [ed.] 2006: 56). 104. Brun, in Cuvigny (ed.) 2006: 135.

48

Rome in Egypt’s Eastern Desert

Hormos, but this fortlet appears to have been built in the second quarter of the second century, which leaves open the most interesting hypothesis, that Simiou could be Biʾr Sayyala, a fortlet in which J.-P. Brun’s excavations revealed a complex evolution that could date back to the Ptolemaic period (although surveys, unfortunately limited, have not revealed Ptolemaic material). Hence the idea that the praesidium would have followed on after the digging of a well, ordered by Simmias, when he was commissioned by Ptolemy III to capture elephants.105 The Roman toponym would have retained the spelling without the geminate.

3. Praesidia on the Berenike road Names of stopovers on the Berenike road have long been known—with some mistakes that ostraca have enabled us to correct—thanks to three written Roman itineraries. The route description by Pliny the Elder, which dates from a period before the construction of the first praesidia in AD 76/77, shows that some of these were positioned around older wells, the names of which were preserved, hence the generic element ὕδρευμα in some of the complex toponyms. We cite them in the order in which they are encountered from Koptos to Berenike. Φοινικών (Al-Laqita [25° 52’ 57” N/33° 07’ 22” E]) Today there are no remains of the praesidium where the roads to Myos Hormos and Berenike meet, but the palm grove from where it got its name still exists. Φοινικών is never preceded by the article. Δίδυμοι (Khasm al-Minayh [25° 45’ 16” N/33° 23’ 40” E]) The official name is Δίδυμοι (so εἰς Διδύμους, ἀπὸ Διδύμων), but some writers have hesitations. Thus, we find the singular in two ostraca of the third century: O.Did. 35, a letter of the curator of Aphrodite to Psenosiris, κου̣[ρά]τορι Διδύμου, and O.Did. 39, a draft of a hypomnema produced by Memnon, κουράτορο(ς) π(ραισιδίου) Διδύμου. An earlier ostracon,106 expertly written, is an order from a certain Psenthotes to provide rations to two passing donkey drivers or camel drivers, addressed “to the contractor of the well at Didymos,” κονδούκτωρι Διδύμου ὑδρεύματ(ος). This Psenthotes must have belonged to a transport company and may not have known the official name of Didymoi, which would have been corrupted, taking its form from Ἀπόλλωνος Ὕδρευμα. Placed under the protection of the Dioscuri (invoked in the proskynemata of letters written there),107 Didymoi belongs to a series of three theotoponyms on the road from Berenike, although δίδυμοι is not a proprial name, which explains that, unlike Διός and Ἀφροδίτη(ς), it is never fixed in the genitive. Ἀφροδίτη(ς) Ὄρους, “Aphrodite of the Desert” (25° 36’ 21” N/33° 37’ 27” E) Two ostraca from Didymoi attest the full name of the praesidium. The letter O.Did. 406, a contract of sorts whereby a husband entrusts to another man the protection of his wife, hired as a prostitute at Didymoi, stipulates that the protector give her back to her husband ἐν πραισειδίῳ Ἀφροδίτης Ὄρους. O.Did. 430 is a letter to a pimp written by Longinus, κουράτωρ Ἀφροδείτης Ὄρο⟨υ⟩ς. This is the only toponym in our corpus to include a locative modifier,108 perhaps to distinguish it from another Aphrodite of which we have 105. Brun, in Cuvigny (ed.) 2006: 133. 106. O.Did. 54, c. AD 96. 107. O.Did. 458; O.Dios inv. 264. 108. The phrase “Berenike of the Trogodytes” (French Bérénice des Trogodytes) has been invented after Pliny, Nat. 2.183, Berenice urbe Trogodytarum, which is not a complex toponym, urbe Trogodytarum being only an explicative gloss in apposition.

A survey of place-names in the Egyptian Eastern Desert

49

no knowledge. In Egypt, Ὄρους is only found in the same position in the toponyms Κερκέσουχα Ὄρους, Ἰσίδιον Ὄρους (villages), and Ἰσιεῖον Ὄρους (monastery). There is some fluctuation in the ending of the theonym, which sometimes remains fixed in the genitive Ἀφροδίτης, and sometimes inflects: we find εἰς Ἀφροδείτης in O.Xer. inv. 1306 (titulus on amphora) and in the post register O.Did. 22.3 and 5, but εἰς Ἀφρ̣οδίδην in O.Did. 39.8–9 and ἰς Ἀφροδείτην in O.Xer. inv. 995 fr.c, l. 5 (a poem about the wells on the Berenike road). In the list of praesidia O.Dios inv. 818, it appears as [Ἀφ]ροδίδη. Κομπασι (Biʾr Daghbagh [25° 23’ 52” N/33° 49’ 08” E]) The reading Compasi of It. Ant., already confirmed by the inscription commemorating the construction of the lacci (cisterns) at Apollonos Hydreuma, Compasi, Berenike, and Myos Hormos,109 is also found on many ostraca from Dios, the direct neighbor of Κομπασι to the south. The linguistic affiliation of the name is unclear; one would expect an Egyptian name, like that of its tutelary deity, the goddess Techosis (Τέχωσις), whose name, attested as an anthroponym, but never as theonym, means “Nubian” (Tȝ-igš).110 This is the only deity with a vernacular name honored in a praesidium in the Eastern Desert; that is the result of the age of the site. Here auriferous quartz veins were exploited from the New Kingdom and still under the Ptolemies (from this period there are remains of ore mills, identified as such through the best preserved mills excavated in 2014 at North Samut).111 When they opened the Koptos to Berenike road to traffic, the Romans took advantage of the well at Daghbagh, which was presumably already named Kompasi, although Pliny only calls it a hydreuma in his list of stages on the road. The well must have been particularly productive, because the ostraca from Dios reveal that Kompasi was a center both for market gardening and for laundry: the inhabitants of Dios used to order fresh vegetables from Kompasi and to send their dirty laundry there. Διός (Abu Qurayya [25° 12’ 52” N/34° 02’ 03” E]) Built, as indicated by its dedication, in AD 115/6, this praesidium probably kept the name of the nearby fortlet it replaced, Biʾr Bayza. The latter may have been abandoned because its well was not satisfactory. The excavations carried out by J.-P. Brun and M. Reddé at Biʾr Bayza showed that this fort dates from an earlier period and is probably contemporary with Didymoi and Aphrodites. Hence the idea that it was already called Dios, which fits nicely in the series. In Latin texts, Διός is not transliterated but translated, as shown in the Antonine Itinerary (Iovis), and in a Latin post register, where the curator presents himself as curator praesidio Ioves (l. Iovis).112 Διός is a fine example of the spontaneous adaptation in daily speech of an institutional toponym. If the title of the commander of the fortlet is κουράτωρ πρεσιδίου Διός (O.Xer. inv. 310), the brevity of the word obviously displeased the Greeks who, to say “go to Dios,” “send to Dios” wrote εἰς τὰ τοῦ Διός and rarely εἰς τὸ τοῦ Διός, “Zeus’s Way,” so to speak.113 Finally, Τατουδιός undergoes univerbation among some writers. Thus we read in a letter: τίνι ἂν δῇς μονομάχοͅ τῶν Tατουδιός, “the messenger of Tatoudios to whom you will hand over...” (O.Dios inv. 1238). Or: κόμισον παρὰ Ἁτρῆν ἱππεὺς Τατουδιός, “receive from Hatres, cavalryman from Tatoudios” (O.Dios inv. 507). 109. ILS 2483 = I.Portes 56. 110. http://www.trismegistos.org/nam/detail.php?record=1427. 111. For the gold mines at Daghbagh see Klemm and Klemm 2013: 161–8 (the authors interpret the mills as ore washeries); however, cf. Redon 2016. 112. O.Dios inv. 922. 113. Ο.Dios inv. 53. According to Mayser, Grammatik II.1.8, the neuter plural article τά used before a personal name means “the house of, the property of.” But the formula can also designate the offices of an official (εἰς τὰ τοῦ βασιλικοῦ γραμματέως) or, in the Byzantine period, a religious building, a monastery: τὰ τοῦ ἁγίου ἄπα Φοιβάμμωνος.

50

Rome in Egypt’s Eastern Desert

Ξηρὸν Πέλαγος (Wadi Jirf, Al-Faysaliyya [24° 55’ 39” N/34° 13’ 50” E]) Although the dedication of the fortlet was destroyed, depriving us of a precise date, Xeron Pelagos was probably built at the same time as Didymoi, Aphrodite, and Dios (Biʾr Bayza) on the orders of the prefect Iulius Ursus, in year 9 of Vespasian’s reign. Yet it does not register in the theotoponymic system of these foundations, but received the beautiful descriptive name “Dry Sea,” also used for an establishment of unknown character in the vicinity of Mons Claudianus.114 The bestowal of two identical names is possible only because the two areas, that of the great imperial quarries and that of the Mons Berenicidis, fell under two different administrations, so that there was no risk of confusion. The recurrence of the oxymoron Ξηρὸν Πέλαγος shows that the large sandy stretches of the Eastern Desert spontaneously reminded the ancients of dried seas. Literary reminiscences may have influenced the choice of the toponymic authority: at the time when the two sites were given this name, that is in the second half of the first century AD, the paradoxical concept of a dry sea appears in several passages in Latin poetry;115 it was only later, especially in the Byzantine period, that it is found in Greek, also in poetic texts. Like Διός, Ξηρὸν Πέλαγος may have been translated, not transliterated by Latin-speakers: the fortlet is called Aristonis, probably a corruption of Aridum,116 in the Antonine Itinerary, which also gives the neighboring site the Latin form Iovis, while the Peutinger Table and the Ravenna Cosmography have the Greek forms Xeron and Dios respectively. The placename, which appears only on ostraca from Dios and Xeron, is often abbreviated Ξηρόν,117 and no known example of Ξηρὸν Πέλαγος comes from within these two forts, where deposits of ostraca are later, as if the second element had been abandoned in the third century. In fact, there is no trace in the three ancient itineraries of the transferred generic, Πέλαγος. Φαλακρόν (Duwayj [24° 44’ 09” N/34° 25’ 40” E]) This is probably the name of the remarkably well-preserved fort of Wadi Duwayj investigated by Michel Reddé in 2010, during our first season in Xeron, but the toponym does not appear in the few ostraca collected on site,118 nor in the letters from Duwayj found at Xeron. In all ostraca where it appears, the toponym (Falacro in It. Ant., Philacon in Tab. Peut.) is a neuter word of the second declension. The earliest occurrence is in O.Krok. I 61.4 (102/3 or 121/2), where a Beduin attack is perhaps mentioned, which would have occurred κατὰ Φαλακρόν. This suggests that the praesidium already existed, though the rarity of the name in ostraca from Dios and Xeron could have made us doubt this. In the latter site, an immediate neighbor of Duwayj, the name of Phalakron appears only in a poem on the wells (O.Xer. inv. 995, beginning of the third century). In O.Dios inv. 818, a list of praesidia, Xeron directly follows Apollonos (it is true that this list also omits Dios between Xeron and Kompasi). The generic element which tacitly relates to Φαλακρόν cannot be πραισίδιον, as a bald fortlet does not make sense; I am inclined to believe that Φαλακρόν is the abbreviation of a complex toponym, like 114. See pp. 74 f. 115. For the early Roman Empire, it is mainly in Latin that I found poetic reminiscences, e.g., Ovid. Met. 2.235: mare contrahitur siccaeque est campus harenae / quod modo pontus erat. Manilius, Astronomica 5.688: congeritur siccum pelagus. Lines 448–49 of book 5 of Sibylline Oracles, an antipagan poem probably composed in Alexandria by a Jew between 80 and 130, use the same image: ἔσται δ’ ὑστατίῳ καιρῷ ξηρός ποτε πόντος, / κοὐκέτι πλωτεύσουσιν ἐς Ἰταλίην τότε νῆες. Other poetic examples can be found in Greek, but in the Byzantine period. 116. Bülow-Jacobsen 2013: 561, n. 3. 117. Ostraca of Xeron: 7 occ. of Ξηρὸν Πέλαγος, 8 of Ξηρόν. Ostraca of Dios: 1 occ. of Ξηρὸν Πέλαγος, 10 of Ξηρόν. 118. A prosopographic overlap allows the date of 216–219. Phalakron was abandoned in the early third century, before the time characterized in Didymoi, Dios, and Xeron by rubbish deposits inside the fort, and by the uncontrolled proliferation of loculi.

A survey of place-names in the Egyptian Eastern Desert

51

Figure 28. The praesidium of Phalakron. In the background, the bald mountain. © M. Reddé

Ξηρὸν Πέλαγος, often abbreviated Ξηρόν, and that the generic element that can be restored is Ὄρος as in Μέλαν Ὄρος, or Ἄκρον, Ἀκρωτήριον. “Bald Mountain” would refer to the pointed mountain that stands out behind the praesidium (Fig. 28). Ἀπόλλωνος Ὕδρευμα (Wadi Jamal [24° 32’ 06” N/34° 44’ 15” E]) The oldest attestation of the site is in Pliny, Nat. 6.102 (mox ad hydreuma Apollinis), but it may be mentioned in an ostracon from the third century BC found at Biʾr Samut (O.Sam. inv. 720). We read τοῖς ἐπὶ τοῦ Ἀπολλωνίου καὶ Προεμβήσει καὶ Πανιείωι in the prescript of this circular addressed to supervisors responsible for several successive stages of the road from Edfu to Berenike. Without being completely sure, we have reason to believe that Proembesis is the ancient name of Biʾr Samut; as for Paneion, the last named stage, this can probably be identified with the station located in front of the Paneion of al-Kanaʾis. The order of the sites mentioned thus runs from the south towards the valley. It is, therefore, tempting to identify this τοῦ Ἀπολλωνίου with Ἀπόλλωνος Ὕδρευμα, admitting the preterition of the generic element, but also an error with the specific, the theonym having been replaced by an anthroponym. In the ostraca of the praesidia, the generic Ὕδρευμα is indifferently employed119 or omitted.120 Καβαλσι? (Abu Ghusun [24° 23’ 13” N/35° 02’ 56” E]) Until recently, the name of the stage situated between Apollonos and Kainon Hydreuma occurred only in the three itineraries of the manuscript tradition, which offer various spellings: Gabaum, Cabalsi, and Cabau. It now appears in a poem of the third century on wells found at Xeron (O.Xer. inv. 995, Br. C.14 119. O.Xer. inv. 257 (post register); O.Xer. inv. 956.5 (soldier’s letter). 120. O.Dios inv. 818 (list of praesidia from Apollonos to Phoinikon); O.Xer. inv. 488.

52

Rome in Egypt’s Eastern Desert

Figure 29. O.Xer. inv. 995, fr. c: Kabalsi? mentioned at the end of line 14. © A. Bülow-Jacobsen

[Fig. 29]).121 Unfortunately, it is compressed at the end of a line and may be abbreviated. Only the first three letters are certain: Καυ. The grapheme b of the Latin transliterations can be explained by the spirantization of the consonant v in Latin. At the time of Pliny, there were no wells between Apollonos Hydreuma and Novum Hydreuma. Despite its oddity, the toponym Kabalsi? should be part of the building campaign ordered by Iulius Ursus in AD 76/7. The traditional identification with the visible remains of Abu Ghusun, where the surface ceramics were dated first to second and fourth to sixth centuries,122 is questionable. Καινὸν Ὕδρευμα (Wadi Abu Qurayya?) This praesidium on the Berenike road, built on the site of a well already mentioned under the name “New Well” by Pliny,123 appears also in the three itineraries of the manuscript tradition. We do not know if this toponym, of which Pliny was aware, dates back to (at least) the Ptolemaic period, or if it refers to the sinking of a new well in the first years of the principate. Kainon Hydreuma is too far south to appear with any frequency on the ostraca from the fortlets we have excavated. It is however mentioned in the poem found in Xeron, which celebrates the water from the various wells on the Berenike road, listed from north to south: Καινὸν Ὕδ(ρευμα) (O.Xer. inv. 995, fr. c.15). According to the stratigraphy, the poem dates from the beginning of the third century and agrees with other sources that make Kainon Hydreuma the last stage on the road before Berenike. The identification of Καινὸν Ὕδρευμα is debated. I share the not uncommon opinion that it is the site called today Wadi Abu Qurayya, a cluster of several buildings situated uphill and downhill. Among them, two fortlets are readily visible on Google Earth at the bottom of the mountain (24° 03’ 45” N/35° 18’ 05” E). Up to now, the most suggestive description is that of Meredith, based on Wilkinson’s unpublished travel diary: “Wilkinson found here separate walled enclosures of various shapes and sizes, one being a normal Roman square but without bastions and another a square but with one rounded end (with bastions), closely resembling the castellum at Semnah (…). Three small forts perched on isolated hills are situated at intervals extending over a mile up a wadi. The last of these overlooks a well beside which are remains of what may be the beginning of a long conduit or aqueduct down to the main enclosures. This small hill fort contains within its wall a high point from which all the other enclosures are visible.”124 According to S. E. Sidebotham, Wadi Abu Qurayya was occupied, from the evidence of the surface ceramics, from the Ptolemaic period to late antiquity.125 The site is at 25 km from Berenike, which nicely fits with the 18 Roman miles indicated by the Antonine Itinerary between the two places. There the Greek Καινὸν Ὕδρευμα is transliterated in Latin Cenon Hidreuma. 121. More details in Chapter 18, pp. 301 f. 122. Sidebotham and Gates-Foster (eds.) 2019: 103 f. 123. mox ad Novum Hydreuma (Nat. 6.102). 124. Meredith 1953: 100–1. 125. Sidebotham 2011: 130, 149, 163.

A survey of place-names in the Egyptian Eastern Desert

53

The question of the identification of Καινὸν Ὕδρευμα has been embroiled by the mention, in Pliny, of a vetus hydreuma which has focussed scholars’ attention, probably because of a mistranslation in the Loeb edition: mox ad Novum Hydreuma a Copto CCXXXVI. Est et aliud hydreuma vetus—Trogodyticum nominatur, ubi praesidium excubat deverticulo duum milium; distat a Novo Hydreumate VII. This passage should be understood as follows: “Then (one arrives) at New Well, at 236 miles from Coptos. There exists also another well, an old one, which is called ‘Trogodytic,’ where a garrison mounts guard, two miles off the main route. It is at a distance of 7 miles from New Well.” In this text, it should be observed first that hydreuma vetus is not a toponym. The toponym which Pliny attributes to this “old well” is Trogodyticum. It appears in no other source. Scholars are thus wrong to speak of a station which would have been called Vetus Hydreuma. Second, why did many scholars identify Wadi Abu Qurayya with “Vetus Hydreuma”? Because, I think, of the mistranslation in the Loeb edition, where it is understood that the garrison there consisted of two thousand men. Five forts were not too much to host this huge troop! But, if “Vetus Hydreuma” was identified with Wadi Abu Qurayya, Novum Hydreuma had to be placed elsewhere. On Meredith’s map, it is situated at Wadi al-Kashir (24° 11’ 03” N/35° 14’ 05” E). Until recently, this was also Sidebotham’s view. However, there is nothing there, except for what looks like a hafir, i.e., an oval levee of gravel meant to retain flood water.126 Therefore, S. Sidebotham now prefers to identify Novum Hydreuma with the tiny fort of Wadi Lahma (or Lahami, 24° 09’ 55” N/35° 21’ 55” E).127 Scholars who identify Wadi Abu Qurayya with the alleged Vetus Hydreuma are well aware that Wadi Abu Qurayya tallies, in terms of distance, with the Cenon Hidreuma of the Antonine Itinerary. It obliges them to imagine either that Pliny’s New Well is another one than the Itinerary’s, or that Novum Hydreuma and “Vetus Hydreuma” are one and the same site.128 As for the Plinian Trogodyticum Hydreuma, it remains a mystery.

4. Praesidia surrounding Mons Claudianus Three names of praesidia, other than those that control the metalla of Claudianus and Tiberiane, appear in the O.Claud. Raïma is by far the most represented, with approximately 90 occurrences; Kampe has 16 references, and Lakkos, only one that is certain. Ῥαϊμα, Ῥαειμα (Abu Zawal? [26° 40’ 18” N/33° 14’ 26” E]) Both spellings are well represented: they are two different ways of rendering the diaeresis. The word normally does not inflect, but in inv. 1801 it exceptionally has an accusative ending (εἰς Ῥαϊμαν). Christian Robin, whom I consulted on this toponym, believes that a Semitic origin (specifically Arabic or Aramaic) is possible. I reproduce the note, dated 28/07/1994, with which he kindly provided me: “The RYM root is well attested in Arabic, see especially (dictionary Kazimirski): raym: supplement; hill; tomb; white gazelle rīm: tomb; burial

126. Sidebotham 2011: 97. Map of this hafir: Sidebotham and Zitterkopf 1995: 44, fig. 6. 127. Sidebotham and Gates-Foster (eds.) 2019: 95. 128. So, Meredith 1953: 100. John Ball, in a confused note, considers that Pliny reverses the names of the two last stages of the road and erroneously calls Novum Hydreuma “Vetus Hydreuma” (Ball 1942: 83). But then where would be, according to Ball, the true Vetus Hydreuma?

54

Rome in Egypt’s Eastern Desert

In Arabic, adding -a(t) gives the singular (as opposed to the collective). The Greek Ῥαϊμα can, therefore, transcribe an Arabic word Rayma(t) whose meaning could be “hill,” “tomb,” etc. In Hebrew and Syriac, the correspondent of the root RYM is RWM, which can give derivative forms with a Y in place of W in Syriac (dictionary Payne Smith): raymoʾ: wild bull, buffalo, unicorn reyomoʾ: support reyomtoʾ: very large stone, barrier Not only may the toponym Ῥαϊμα be of Semitic origin (specifically Arabic or Aramaic), but it has a matching term in Yemen, and a rarer one in Saudi Arabia. In south Yemen, the repertoire of toponyms indicates four Rayma (People’s Democratic Republic of Yemen. Official Standard Names. Prepared by the Defense Mapping Agency Topographic Center, Washington 1976, 165). In North Yemen, a less systematic repertoire gives seven. In Saudi Arabia, I note five Rīm, one ar-Rīmān, one Rīmān and one arRayma. Rayma is a very common toponym in southern Arabia (but rarer in the Arabian desert).” The many letters written by curators of Raïma to correspondents at Mons Claudianus (centurion, curator, oikonomos) show that Raïma was the first way station after Claudianus in the direction of the Nile and, conversely, the last station before Claudianus on the ὁδὸς Κλαυδιανοῦ. These messages are often cover letters or receipts tracking the official letters. Raïma was not only an official relay station of the mail service, but also a stage where caravans and teams drawing wheeled vehicles stopped and drank. The praesidium must have had a well like many fortlets of the desert. It is possible that they had to build a second one nearby: this could be the ⟨h⟩ydreuma about the successful drilling of which Antistius Flaccus, writing from Raïma, tells his colleague Caninius, who is at Mons Claudianus (O.Claud. I 2). They certainly did not wait until the reign of Trajan (date of the letter) to sink a well in Raïma, the name of which is restored with a fair degree of certainty on an ostracon of AD 68.129 This new well near Raïma may be the same as that referenced in the letters of Demetras O.Claud. II 383 and IV 864: Demetras informs his correspondents that a stonecutter sent from Mons Claudianus was not present at the hydreuma, and, secondly, that baskets which should have been sent from Raïma to the hydreuma went off to Claudianus. Several curators’ letters betray the fear of running out of water or are requests for equipment for the maintenance of shadufs (plural, κήλωνες, κηλώνεια: one for each well?).130 It comes as no surprise that Raïma had a vegetable garden (κῆπος), which allowed the curator to send vegetables to the centurion stationed at Claudianus (O.Claud. II 370). One would think that Raïma received its supplies from the valley, when the poreia went towards Claudianus, passing through Raïma. Yet we observe that Raïma’s granaries are sometimes supplied from Claudianus (O.Claud. I 124 and 125: acknowledgments for loads of achyron [chaff] sent from Claudianus to Raïma by the caesarianus Successus; inv. 2188: acknowledgment to Philon for three artabai of bread). Raïma had a granary, and its manager (θησαυροφύλαξ) informed his counterpart in Claudianus that a carpenter collected his rations at Raïma (inv. 555). The only major way station that could correspond to what we know about Raïma from this correspondence is Abu Zawal,131 c. 33 km walk from Claudianus. We owe to R. and D. Klemm the best description of this site, which was exploited as a gold mine under the New Kingdom and under the 129. O.Claud. inv. 8828, found in the “Hydreuma:” εἰς Ῥα[ϊμα]. 130. O. Claud. inv. 2238. For a shaduf in a desert way station, see fig. 13.8 in Sidebotham et al. 2008: 320 (reconstruction of the way station of Wadi Abu Shuwayhat, also called Talʿat al-Zarqaʾ, on the road to Claudianus). 131. Sidebotham 2011: 119–20.

A survey of place-names in the Egyptian Eastern Desert

55

Figure 30. Ostracon found at Abu Zawal. © R. Klemm

Ptolemies.132 According to their observations, Abu Zawal was only a station on the road to Claudianus in Roman times. They published a photo of an ostracon from the dump west of the fortlet, with a translation given to them by the late Georges Nachtergael. Here is the text, which I transcribe from the photo (Fig. 30):133

4

2 l. κόμισαι

Ψεντάησις Ἀμμωνίῳ τῷ ἀδελφῷ χαίρειν. κόμισον ῥακάδιν σκάτεα χυριδίων. ἄσπασε Κασυλλᾶτι. vacat ἐρρῶσθαί σε.

3 l. σκατῶν χυ- ex χρ- corr., l. χοιρ-

l. ἄσπασαι 3–4 l. Κασυλλᾶν 5 σε––

“Psentaesis to Ammonios his brother, greetings. Take delivery of a rag containing pig manure. Greet Kasyllas. Farewell.” The three people mentioned have not been identified with individuals known from the Claudianus ostraca, whence the letter was probably sent. It could indirectly confirm the identification of Abu Zawal 132. Klemm and Klemm 2013: 70–74. 133. I thank Rosemarie Klemm for kindly providing me with the original photo and permission to publish.

56

Rome in Egypt’s Eastern Desert

with Raïma. Not only did Raïma have a vegetable garden, but an ostracon from Mons Claudianus, O.Claud. II 280, a letter from a praesidium where market gardening is practiced (it ends with the information, “vegetables have not grown yet”),134 accompanies the empty return of “the basket you sent us loaded with excrements,135 that you will return to us when you get the opportunity.” The soiled basket could probably only serve for these contents thereafter in the comings and goings between Claudianus and Raïma. On the ostracon from Abu Zawal manure is not transported in a basket, but in a piece of cloth. The diminutive ῥακάδιον (of τὸ ῥάκος, “rag”) is only known from papyrological documentation, and nearly all the occurrences are found on ostraca from the Eastern Desert, where a ῥακάδιον is for packaging up sticks of collyrium, dates, and a tunic.136 The editor of O.Claud. II 280 questions the nature of the fertilizer sent: the ostracon from Abu Zawal perhaps offers a response (and suggests additionally that there were no pigs at Raïma when this letter was written). It remains, nevertheless, curious that this gardener from Raïma is not happy with donkey or camel dung, which must have been locally plentiful: did the pig waste have a special reputation? This is the case at least for the cultivation of certain fruit trees, according to Theophrastus (CP 2.14.2, 3.10.3 [ὑεία κόπρος]) and Pliny (Nat. 17.258–59 [fimum suillum]). But there may be another hypothesis: this was not for fertilizer, but an ingredient for a medicinal preparation or magic potion, as in the letter O.Did. 395, which contains the promise to bring τὸ σκῶρ τῶ χυρογρύλω (τοῦ χοιρογρύλλου), “the dung of hyrax.”137 A morphological argument could be made in favor of this hypothesis: the curious accusative plural σκάτεα calls to mind a remark by Gignac on the non-contracted accusative plural -εα instead of -η neuter nouns ending in -ος (type γένος, gen. γένους): accusative -εα is found only in magical papyri.138 In fact, next to the classical form τὸ σκῶρ, gen. σκατός, Phrynichus (second century AD) indicates a common form τὸ σκάτος, gen. σκάτους.139 It should be observed, however, that the pig is virtually the only domestic animal whose droppings are not widely advocated by ancient pharmacopoeia.140 Nevertheless, it is remarkable that none of the three ostraca from the Eastern Desert mentioning manure employs the usual term, which is κόπρος, both in magical or medical recipes and when it refers to fertilizer.

134. The editor, Jean Bingen, still doubts whether it is Raïma, because the author of the letter also acknowledges receipt of a water amphora: why indeed send water to Raïma? But it is undoubtedly special water. This is certainly not an amphora with water to water the vegetables. 135. τὸ σφυρίδιν ἃ (l. ὃ) ἔπεμψες ἡμῖν ὑπὸ τῶν χεσμάτων. 136. The papyrological occurrences of ῥάκος and of ῥακάδιον are gathered together and discussed by Mascellari 2015: 151–9. 137. In the Eastern Desert, it is the rock hyrax, Procavia capensis, not the bush hyrax Heterohyrax brucei (Yves Lignereux, email 17/07/2016). This small mammal is the size of a rabbit, and it is hard to imagine that its droppings were used to fertilize a vegetable garden. 138. Gignac, Grammar II, 66 f. Robert Daniel, whom I consulted about this remark and would like to thank, cautioned me that Gignac cites only χείλεα in PGM 4.401, which is actually the only example, and that it is interpreted as a poetic reminiscence: in prose passages of magical papyri, we have the regular form χείλη. 139. Τὸ σκάτος· καὶ τοῦτο ἐπ’ εὐθείας τιθέμενον ἀμαθές· γενικῆς γάρ ἐστι πτώσεως, τοῦ σκατός, ἡ δὲ εὐθεῖα τὸ σκώρ. ἁμαρτάνοντες δὲ οἱ πολλοὶ τὴν μὲν ὀρθὴν τὸ σκάτος ποιοῦσιν, τὴν δὲ γενικὴν σὺν τῷ υ τοῦ σκάτους (Eclogae 260). 140. While the use of goat, sheep, horse, cattle, camel, dog, and cat dung is constantly attested by the authors and in medical texts, Pliny is the only source mentioning pig excrement that is reported by Durling 1999: 28. After describing the different ways of preparing wild boar droppings for medication, Pliny (28.138) simply adds that pig manure has properties similar to those of wild boar manure, which gives the impression that it might serve as a substitute (proximam suillo fimo putant vim). Boar dung, which, unlike that of pigs, is frequently cited in the medical literature, was, according to Pliny, used against bruises and injuries due to falls, so that charioteers would commonly use it. For this reason, Nero would put on a show of consuming it.

A survey of place-names in the Egyptian Eastern Desert

57

Καμπή This praesidium141 is mentioned in sixteen ostraca, Trajanic and later, of Mons Claudianus. Its name seems to suggest that it was on a bend. This concept is common in toponymy.142 Once, in a letter, the word is preceded by the article (εἰς τὴν Καμπήν, O.Claud. I 155). In a proskynema in front of the Tyche of Kampe, the word is s-enlarged (τῇ Τύχῃ Καμπῆτος, O.Claud. II 237.5). Two letters addressed by curators of Raïma (presumably Abu Zawal) to curators of Claudianus acknowledge receipt of post sent from Claudianus and confirm that it has been forwarded at once to Kampe (O.Claud. inv. 7027 and 7595). It seems therefore that Kampe was a station beyond Raïma in the direction of the Nile. Hence three identifications are possible: (1) Kampe would be the station of Al-Saqiya on the road to Porphyrites; its name, “The Bend,” would refer to the fact that one would turn off at Kampe towards Claudianus. Al-Saqiya (26° 44’ 04” N/32° 52’ 54” E) is at c. 38 km walk from Abu Zawal. (2) Kampe would be Talʿat al-Zarqaʾ (26° 35’ 09” N/33° 11’ 56” E). This site, which was never excavated, is called today Abu Shuwayhat by the Beduins; it comprises a well and animal-lines, and is situated at the entrance of the mountains, where a network of small valleys allows direct access to Abu Zawal (at least it looks feasible on Google Earth for a pedestrian or a rider); the shortest way between the two sites is c. 11 km. (3) Kampe would be Al-Qurayya (26° 22’ 16” N/33° 01’ 08” E). This praesidium, which can be seen behind the railway from the old Qena-Safaga road, is equipped with animal-lines. It may have been at the junction of the via Claudiana and the routes towards more southern sites, especially the Ophiates. Λάκκος O.Claud. inv. 2283 (Trajanic?) suggests that Lakkos, the Cistern, is a praesidium, probably so called by metonymy, because it did not have a well, only a cistern that was filled thanks to the caravans of animals loaded with skins. The recipient of this wrathful letter, in which he is accused of not having returned all the skins, is κουράτωρ Λάκκου, and the title curator is attested in the Eastern Desert for curatores praesidiorum only. The small praesidium at Mons Claudianus which, since Schweinfurt, has infelicitously been called “The Hydreuma,” despite the fact that it has no well but only a cistern, might be a good candidate. However, this identification is not confirmed by the ostraca found on the spot: the address of destination, when written on amphoras found at “The Hydreuma” is εἰς Κλαυδιανόν.143 O.Claud. inv. 2283 is the only certain evidence of a praesidium called Lakkos. Among the other occurrences in the Claudianus corpus of the noun λάκκος, I am tempted to believe that in O.Claud. IV 714 and 717 (Trajanic) this is also the praesidium because of its proximity to two other toponyms: in these lists, the last items are, in order: λάκκῳ (Λάκκῳ?), Ῥαϊμα, Αἰγύπτῳ, ἄρρωστοι, “to the cistern (C-?), at Raïma, in Egypt, sick people.”

5. Praesidia surrounding Umm Balad A. The field data Four fortified quadrangles exist within a radius of ten kilometers of Umm Balad. The most frequent names of praesidia in O.KaLa. necessarily refer at least to some of them. These four praesidia are:

141. O.Claud. inv. 7038 mentions a curator of Kampe. 142. Gendron 2006: 42 f. 143. O. Claud. inv. 8851, 8875, 8890, 8908, and 8923.

58

Rome in Egypt’s Eastern Desert

Figure 31. Tracks to Porphyrites, according to Maxfield and Peacock 2001: 194. The wells identified are reported. © S. Goddard

A survey of place-names in the Egyptian Eastern Desert

59

– Qattar (27° 05’ 21” N/33° 13’ 44” E) is located directly on the ὁδὸς Πορφυρίτου and 9 km as the crow flies from Umm Balad, c. 11.5 km for a traveller. It is the immediate neighbor of Umm Balad in the direction of Qena/Kaine.144 Like all the way-stations on the hodos Porphyritou (which Umm Balad is not), the praesidium of Qattar is flanked by animal-lines. All traces of the ancient well have disappeared: either it was inside the fort and was destroyed by the construction of modern wells, or it was outside and has silted up. – Badiya (27° 12’ 52” N/33° 20’ 42” E) is at 9 km as the crow flies, 11.20 km drive from Umm Balad, and its immediate neighbor in the direction of Porphyrites.145 Like Umm Balad, it is located a short distance from the large Wadi Belih that the ὁδὸς Πορφυρίτου follows. The animal lines, which extend in front of the fort, show that it was an important stopping point for beasts of burden and draft animals coming or going to Porphyrites (see Fig. 31). From Badiya there were two routes to Porphyrites. Men and beasts could get to the “footpath station” at the foot of the mountain and pass through it by a zigzag trail on the hillside that led to Wadi Maʿmal and the administrative center of the metallon. However, transport wagons that were to be loaded with monoliths at the main loading ramp (Fig. 31) would have followed a track skirting the massif of Jabal Dukkhan. This long detour route was mainly used for the monoliths coming down from Porphyrites. A few meters from the fortlet of Badiya rises a curiously fortified rocky hill. There is no well within the praesidium; the closest known well, Biʾr Badiya, still in use, is 500 m away; it is not possible to know whether it is the ancient well. A nearby well, that is unknown today, may have supplied the occupants of the fortlet. – The “footpath station” (27° 13’ 03” N/33° 18’ 25” E) is a tiny fort 8 km from Umm Balad as the crow flies, on a shortcut to Wadi Maʿmal, where the administrative center of Porphyrites was.146 This winding route147 allowed quicker access to Wadi Maʿmal for human and animal pedestrian traffic148 than that via Wadi Umm Sidri, which wheeled vehicles would have used (Fig. 31). It is unclear whether camels could use it or not.149 The satellite image shows that one could reach this fort from Umm Balad bypassing Badiya, but we did not have time to explore exactly which route the ancients took. I believe, from the image, that this route was about 12 kilometers. – Wadi Belih (27° 14’ 19” N/33° 22’ 55” E, at 13.5 km from Umm Balad as the crow flies).150 This is a small praesidium with an atypical plan, seemingly devoid of an inner well, located in the estuary of Wadi Belih. It was, therefore, at the entrance (for traffic from the Red Sea) of the diagonal connecting Porphyrites and, later, the fort of Abu Shaʿr to Qena/Kaine across the Arabic chain, with Wadi Belih continuing into Wadi al-Atrash. Was the fortlet used to control the entry of the 144. Sidebotham et al. 1991: 582 f. (with plan). 145. Maxfield and Peacock 2001b: 215–37. Well: p. 236 f. 146. Maxfield and Peacock 2001b: 200–2. 147. Photo in Maxfield and Peacock 2001b: 201, fig. 5.13. 148. A. Bülow-Jacobsen, per os. 149. My two informants, to whom I could only show the photo cited n. 147, are of different opinions. L. Nehmé thinks that camels could use that mule track, “provided the path is wide enough, as the photo suggests. There are trails of this type around Petra, on plots of caravan routes” (email of 28 June 2016). But Carlo Bergmann, who travelled across the Sudanese and Egyptian deserts with small groups of camels, is less certain: “Such trails are not really suitable for camels. Firstly, the trails seem to be quite narrow. Camels walk in amble and would be afraid to follow such lanes, especially if these (like the one in the middle) are running almost parallel to a quite steep slope and if the beasts carry heavy loads. Secondly, the trail in the centre seems to pass over very rough gravel which fills the front of the picture. Anyone caring for his camels would have cleared from the track at least a few of the roughest stones. (...) The hoofs of donkeys would not require such clearance” (email of 18 September 2016). 150. Sidebotham et al. 1991: 577 f. (with plan); Maxfield 1996: 17–19.

60

Rome in Egypt’s Eastern Desert route to the Nile? Was it a way-station on the via Hadriana?151 It was one kilometer away from the heavy-loads route to the Porphyrites, which went through Badiya.152

From Wadi Belih one could reach Badiya by walking just over 5 km. Wadi Belih has never been excavated; it has been dated from the surface ceramics to the first and second centuries AD.153 In the following table all the toponyms in the Umm Balad corpus are presented. A site is characterized as a praesidium when under the command of a curator or when the generic praesidium is attached to the specific. Names of metalla have already been studied; those of wells are in the section on hydreumata. Table 1.5. The toponyms mentioned in the ostraca from Umm Balad Topographic Feature

Number of Occurrences1

… of Which the Number of Tituli

Kaine Latomia

metallon + praesidium

102

93

Domitiane

metallon + praesidium

43

31

Porphyrites

metallon + praesidium

36

5

Claudianus

metallon + praesidium

6

0

Prasou

praesidium + well

29

2

Sabelbi

praesidium + well

22

6

Akantha

praesidium

10

0

Kardameton

well

7

0

Arabarches

metallon?

4

1

Germanike Latomia

metallon

1

0

Melan Oros

praesidium

5

1

Berkou2

praesidium

1

0

1. This is precisely the number of ostraca mentioning these names. 2. See below s.n. Σαβελβι.

Μέλαν Ὄρος (Dayr al-Atrash?) Among the praesidia mentioned in the ostraca from Umm Balad, Melan Oros has—apart from Berkou—the fewest references: five in all. So, it was probably the farthest away. Yet its identification seems the soundest. Two of these occurrences clearly demonstrate that Melan Oros was between Umm Balad and the Nile Valley. In O.KaLa. inv. 637, the soldier Marcus Ares Verus, stationed at Melan Oros, promises to send vegetables to the Chief Doctor at Umm Balad “when the steward for the provisions for soldiers comes (sc. from the valley).” In O.KaLa. inv. 275, Antistius, short of bread, asks his correspondent Antonius, who is at Umm Balad, to send him some on the next poreia: we understand that when the supply caravan comes from the valley, Antistius will receive his bread ration and will take a portion that he will resend on the poreia so that it gets to Umm Balad, for Antonius. To reassure Antonius of a quick refund, Antistius added this encouraging information, “they say that the poreia is at Melan Oros” (ἐν 151. Sidebotham et al. 1991: 577. 152. Sidebotham et al. 1991: 577. 153. Sidebotham et al. 1991: 577.

A survey of place-names in the Egyptian Eastern Desert

Figure 32. Dayr al-Atrash. © A. Bülow-Jacobsen

Figure 33. Black hills behind Dayr al-Atrash. © A. Bülow-Jacobsen

61

62

Rome in Egypt’s Eastern Desert

Μέλανι Ὄρει). The find in Umm Balad of an amphora titulus mentioning the destination of Μέλαν Ὄρο[ς] could be explained by a similar arrangement.154 The demand for bread suggests that there is at least one praesidium (the one where Antistius is stationed) between Umm Balad and Melan Oros. As Melan Oros cannot be far away, it is tempting to identify it as Dayr al-Atrash (Fig. 32), Antistius being then stationed at Qattar. Dayr al-Atrash may owe its name “Black Mountain” to the color of the nearby hills (Fig. 33). Μέλαν Ὄρος is once followed by the generic: a soldier is described in a draft of a letter as στρατιώτης ἀπὸ Μέλανος Ὄρους πραισιδίου (O.KaLa. inv. 637). Or should we write Πραισιδίου? It is tempting to identify the praesidium of the Black Mountain as that of the Black Stone Mountain (μέλανος λίθου ὄρος) that Ptolemy, in his list of the Eastern Desert quarries Geogr. 4.5.27,155 inserted between the mountain of porphyry and the mountain of basanites (i.e., graywacke quarries of the Wadi al-Hammamat).156 The editors of the Geography, including Müller, do not provide an identification. Πράσου (Badiya or Qattar?) Indeclinable, the genitive of τὸ πράσον, the leek, which is also the name of a cape on the eastern coast of the African continent, mentioned several times by Ptolemy (τὸ Πράσον ἀκρωτήριον).157 In the case of a cape, τὸ πράσον must refer not to the vegetable, but to a seaweed, posidonia. According to letters sent to Umm Balad, this praesidium had a well, a cistern, and a vegetable garden. As it seems to be an immediate neighbor of Umm Balad, it must be Badiya or Qattar, but the content of the letters does not allow us to be sure. Abundantly attested in the ostraca of Umm Balad, Prasou may be mentioned in O.Claud. inv. 3438 (Trajanic), a fragmentary letter in which Epagathos asks Geta to send him something related to cavalry men; line 4 reads: ἐν Π̣ράσου π̣[. Σαβελβι (Badiya or Qattar?) The spelling of this enigmatic place-name varies: Σαβελβια̣ (O.KaLa. inv. 697), Σεβιλβια (accusative Σεβιλβιαν) in O.KaLa. inv. 208 and 564 (although in this case, the paleography suggests rather Σαβελβιλ), Σαβαλ[ (O.KaLa. inv. 541). I have found only two foreign names with a similarity to Σαβελβι: the Libyan name Ταβαλβις158 and the name of the Mysian Σταβέλβιος in Aristotle’s Economics (1353b). Cornelia Römer suggested to me a comparison with the Arabic sabil. This site, which is mentioned 22 times in ostraca from Umm Balad, cannot be far away. The existence of a curator of Sabelbi guarantees that it is a praesidium. The letter O.KaLa. inv. 266 is particularly informative: its author, Paulinus, asks his correspondent to send three men as reinforcements to clean out the well at Sabelbi so that the wagon will have water when it returns from Porphyrites. It is, therefore, tempting to assume that Sabelbi can be identified with Badiya (for another argument in favor of Badiya, see p. 63). However, Qattar cannot be ruled out: would not the wagon returning from Porphyrites need water not only from Badiya, but also from Qattar? A. Bülow-Jacobsen cites O.KaLa. inv. 422 to support the second hypothesis:159 Hieronymos addresses this letter to Maximus, stationed in Umm Balad, asking 154. O.KaLa. inv. 483. 155. = p. 695 in Müller’s edition. 156. Called by Ptolemy ἡ ὀρεινὴ ῥάχις τοῦ βασανίτου λίθου ὄρους. On this passage of Ptolemy, see pp. 15 f. Remarkably, Mons Claudianus is not in this list (and does not appear at all in the Geography). 157. Cf. also Marcianus, Periplus maris exteri, 1.13.15: ἐν δὲ τῷ τέλει τοῦ κόλπου κεῖται τὸ μέγιστον ἀκρωτήριον, ὃ καλεῖται Πράσον ἄκρον. 158. Masson 1976: 60. 159. “The Ostraca from Umm Balad,” PapCongr. XXVIII, 503–11. Electronic version https://repositori.upf.edu/ handle/10230/41902.

A survey of place-names in the Egyptian Eastern Desert

63

him to send ropes “so that the wagon can hold steady on the descent (τὸ καταβατόν) which is at Sabelbi.” Close to Qattar the track descends gently: for A. Bülow-Jacobsen, this is what is meant with καταβατόν. A wagon carrying a monolith must have stopped at the top of this slope, waiting for the ropes to retain it. In January 2020, the excavation at the neighboring Ghozza allowed some progress on the question.160 Ghozza (i.e., “razzia”) is a Ptolemaic village of gold-miners dating to the third–early second century. Beside the village, a Roman praesidium was built, perhaps under Domitian, and briefly occupied (one ostracon is dated to Domitian’s 10th year). Several tituli picti show that its name was Βερκου, which allowed me to recognize this place-name in one fragmentary ostracon from Umm Balad (O.KaLa. inv. 847), dating to c. 91–100. Although they are few, ostraca from the praesidium of Berkou contain several occurrences of the place-name Sabelbi. Since there is, according to Google Earth, an almost direct route from Ghozza to Qattar, via Biʾr Umm Disi, this is a serious argument for the identification of Qattar with Sabelbi. Ἄκανθα The existence of a curator of Akantha, who is the author of a letter sent to a centurion residing in Umm Balad,161 seems sufficient to consider that this site, probably named after a remarkable acacia, due either to its isolation, or size, had the status of a praesidium. The moderate number of occurrences suggests that Akantha is either relatively far from Umm Balad (but perhaps not as far as Melan Oros), or off the road to Porphyrites. In reality, it cannot be very far: the purpose of the curator’s letter is to complain about a certain Amais, posted at Umm Balad, who refused to send his rations to a certain Mithron stationed at Akantha. Mithron is forced to go to Umm Balad to pick them up, probably with the letter from his curator. Akantha may not be a stopover for the poreia, the supply caravan. Three letters date from the Domitianic-Trajanic phase of Umm Balad, which are requests to the authorities of Umm Balad (a centurion, the curator, and the architect Hieronymos)162 to order the garrison at Sabelbi to send a cavalryman to escort a camel going to Akantha, so that it will not be going alone. Akantha seems, thus, to be a remote site accessed via Sabelbi. If Sabelbi is to be identified with Badiya, then Akantha could be either the footpath station or the fortlet of Wadi Belih. Two of these requests come from Turranius, who I think is curator at Prasou.163 On the other hand, I do not see which praesidium Akantha could be if one identifies Sabelbi with Qattar, unless Akantha is the same site as the well of Akanthion (once called Ἄκανθα) mentioned in the ostraca from Mons Claudianus and the eponym of a road.164 Is there a road linking Claudianus to the region of Umm Balad? The possibility of such a route exists: after crossing the sand-sea that spreads NW of Claudianus, one enters the mountains, then bears southwest towards Umm Shejilat (26° 56’ 36” N/33° 14’ 53” E); 1.5 km northeast of this little quarry, one follows up a north–south wadi towards Qattar. There Meredith’s map has a well called Biʾr Umm Disi, invisible on the satellite photo. However, I do not think that this discreet well could be the remnants of the Akantha of O.Claud. and O.KaLa. In the O.Claud., the Akantha provides water for Mons Claudianus (p. 66): Biʾr Umm Disi is too far away for that. On the other hand, the Akantha in the O.KaLa. is under the command of a curator: it must have been a fortified well, in other words a praesidium, which would have left traces. Akantha still existed in the third century: it is mentioned in several of the ten letters of the small archive of the centurion Caninius Dionysios, who seems to have had a short-lived stay in the praesidium of Umm Balad at the time. 160. 26° 52’ 09” N/33° 06’ 32” E. Excavations of the MAFDO, under the direction of Thomas Faucher (CNRS–IRAMAT). 161. O.KaLa. inv. 549. 162. O.KaLa. inv. 783; 785; 811. The last two letters are written by Turranius, who we have reason to believe is curator of Prasou. 163. On Turranius’s letters, see Chapter 33. 164. See below p. 66, s.v. Ἀκάνθιον, Ἄκανθα, “The Acacia.”

64

Rome in Egypt’s Eastern Desert

Ἄκανθα inflects. Unlike for the well named τὸ Ἀκάνθιον in the O.Claud., the use of the article in front of Ἄκανθα is variable and appears to depend on individuals or perhaps the date.165 Turranius, author of many letters during the first phase of Umm Balad, uses the article, which is absent in three letters (of two different writers) from the third-century archive of Caninius Dionysios.

6. Conclusion on the toponymy of the praesidia Among the Greek and Latin names of praesidia, some are declinable, others are fixed in the genitive; these are underlined in the table below: Table 1.6. Names of the praesidia Noun

Adjective

appellative

Ἄκανθα, Δίδυμοι, Καμπή, Κροκοδιλώ, Νιτρίαι, Πράσου, Ξηρόν, Φαλακρόν Φοινικών

2 appellatives (noun + adjective)

Ξηρὸν Πέλαγος, Μέλαν Ὄρος, Θῶνις Μεγάλη

proprial

Ἀφροδείτης/Ἀφροδείτη, Διός, Ἀπόλλωνος, Σιμίου, Πέρσου

Μαξιμιανόν

It appears from this table that the names of praesidia which remain in the genitive are proprial names (theonyms and anthroponyms), Πράσου being the exception. In the case of Ἀπόλλωνος, the fossilization of the genitive comes from the abbreviation of Ἀπόλλωνος ῞Υδρευμα, but this explanation does not account for Διός and Ἀφροδίτης, where it is known that the praesidium was not founded on the site of an old well whose compound name would have been retained. These genitives are, therefore, understood as complements to the term πραισίδιον, even if this noun, unlike ὕδρευμα, is practically never used as a generic constituent in a complex toponym. In three cases, the names of praesidia are unique adjectives. Μαξιμιανόν, a proprial adjective, agrees with the implied generic πραισίδιον; in the case of Ξηρόν, the noun understood is not πραισίδιον but Πέλαγος, which is also not always omitted and which has the distinction of being a generic element “transferred,” i.e., referring to another topographic object that it names. Φαλακρόν also cannot refer to a πραισίδιον, but probably to a noun meaning “mountain”; so this is probably another example of transferred generics, but always omitted, at least in the documentation that we have.

VI. Wells (hydreumata) Misled by the text of Pliny and some complex toponyms which include the generic Ὕδρευμα, modern scholars have long identified the fortlets of the Eastern Desert as hydreumata, from where, for example, comes the unfortunate name “the Hydreuma” for the small fort of Mons Claudianus, although it has no well, as it was long believed that hydreuma meant any water point, whether it was a well or a tank. This is not so: the ostraca of the Eastern Desert have shown that this term, peculiar to Greek spoken in Egypt, only meant a well. In the desert, a water-tank is ὑδρεία or δεξαμενή and, in Roman times, λάκκος. The hydreumata listed below are isolated wells, not integrated in any praesidium. In the ostraca from Umm Balad, a single well (which has not been identified in the field) seems to correspond to this definition: Καρδάμητον is designated as τὸ Καρδαμήτου ὕδρευμα in O.KaLa. inv. 747. This name is found in four ostraca. It is a hybrid, made from the Greek κάρδαμον, “cress,” and the Latin suffix 165. Five occurrences without the article, and three with it.

A survey of place-names in the Egyptian Eastern Desert

65

Figure 34. The lintel of the room of cisterns at Mons Claudianus. © A. Bülow-Jacobsen

-etum, which serves to form collective nouns.166 Καρδάμητον, therefore, means “watercress bed,” a feature that is difficult to imagine in the desert. Unlike τὸ Ἀκάνθιον, Καρδάμητον is never used with the article, perhaps because the speakers did not see this strange hybrid as a common noun: ἐν Καρδαμήτῳ (O.KaLa. inv. 307), εἰς Καρδάμητον (O.KaLa. inv. 850). All other hydreumata addressed in this section are mentioned in ostraca from Mons Claudianus, with one exception, which we can no longer locate. One of them was perhaps the well of Wadi Umm Diqal,167 located in the middle of a quadrangle, which is 3.4 km from the praesidium of the Wadi Umm Husayn, when one uses the way that passes in front of the “Hydreuma.” The exception is the Hydreuma Traianon Dakikon, which was sunk at the time of Trajan, a few meters west of the praesidium of the Wadi Umm Husayn. This honorary toponym (the only one in our corpus) is known from two inscriptions: ὕδρευμα εὐτυχέστατον Τραιανὸν Δακικόν fons felicissimus Traianus Dacicus (altar I.Pan 37, AD 108/9) (ὕδρευμα εὐτυχέστατον) Τραιανὸν Δακικόν fons abundans aquae felicis Traianus Dacicus (lintel in the room of cisterns, SEG XLII 1574) These inscriptions were probably engraved for the inauguration of the well by the prefect Sulpicius Similis, who came in person.168 The altar-inscription, on the podium of the Serapeum, has long been known, but has now been destroyed. The other text is engraved on the lintel of the door of the cistern room, and was discovered during the 1991 season inside the praesidium (Fig. 34). Jean Bingen, the editor of the lintel inscription, considers Τραιανὸν Δακικόν/Traianus Dacicus as the name of the well, to the exclusion of the rest (ὕδρευμα εὐτυχέστατον, fons felicissimus, fons abundans aquae felicis) which he calls a “qualification.”169 The structure of this toponym, however, is not unusual: generic (ὕδρευμα/ fons), followed by an accumulation of specifics with a variant in Latin. For lack of space, the stonecutter has omitted ὕδρευμα εὐτυχέστατον and placed in the center as a common factor fons abundans aquae felicis, while the last two specifics expressing loyalty to the emperor were arranged on both sides in Latin and Greek. It is possible that this cumbersome toponym, which was not in common use, refers, through metonymy, to the whole settlement of Wadi Umm Husayn. This would explain why neither of the inscriptions was found next to the well proper. In the third century BC, another well-station of the Eastern Desert combines the idea of good luck with a dynastic name: τὸ Ἀρσινόης Εὔκαιρον ὕδρευμα, mentioned in several circulars found in Biʾr Samut. 166. I owe this explanation to the perspicacity of Jean-Louis Perpillou. 167. Peacock and Maxfield 1997: 151–54. 168. H. Eristov and H. Cuvigny, “Le faune et le préfet. Une chambre peinte au Mons Claudianus,” BIFAO 121 (2021): 183–254. 169. Bingen and Jensen 1992: 16.

66

Rome in Egypt’s Eastern Desert

Τὸ Ἀκάνθιον, Ἄκανθα, “the Acacia” Several Trajanic ostraca from Mons Claudianus mention a well, where trains of camels went to fetch water, normally called τὸ Ἀκάνθιον (always with the article), and once, in a Greek letter written by a Latin-speaker, Ἄκανθα (no article and not declined: εἰς Ἄκανθα, O.Claud. II 362). The money account O.Claud. inv. 3819 ends with the order: ἔρχου ἰς τὸ Ἀκάνθιον̣ τὸ ὕδρευμα. This is the only instance where that toponym explicitely refers to a well. A road was called after this well, according to two passes addressed to stationarii/epiteretai [ὁδοῦ] Ἀκανθίου (O.Claud. I 77 and 81); ὁδοῦ is restored each time, but this seems to be a necessary restoration. It must be a secondary path leading from Claudianus to this well. The well of Dioskoureia The letter O.Claud. inv. 490 concerns a case of a diverted waterskin, in which “the men at the well of the Dioskoureia” (οἱ ἐκ τοῦ ὑδρεύματος ⸌τῶν⸍ Διοσκορίων) are involved. According to this ostracon, water was transported from this well to water the horses accompanying the poreia, i.e., the caravan periodically provisioning the metallon. The toponym is also attested in the great Trajanic water-chart,170 where it appears among the names of sites and quarries receiving water rations: latomiai, loading ramps (κρηπῖδες) of latomiai, steelworks (στομωτήριον), and a watchtower (σκόπελος). The names of the supplied sites are in the dative, and so is Διοσκορίοις, which probably refers to the hydreuma of the Dioskoureia in the letter, but as water is sent there, the well must be under construction. A further probable instance of the toponym appears in O.Claud. inv. 7955.

Figure 35. O.Claud. inv. 7955. © A. Bülow-Jacobsen 170. O.Claud. inv. 1538.6, published in Chapter 11. The reading ] Δ̣ιο̣σ̣κ(ορίοις) in O.Claud. IV 695.3 seems doubtful given the infrared photo.

A survey of place-names in the Egyptian Eastern Desert O.Claud. inv. 7955 (Fig. 35) Well—S1 C-W 4

8 × 6.5 cm

67 Trajan Nile silt clay

Upper right corner of a letter, perhaps unfinished, written by an anonymous curator who broke off when he saw an error in the prescript, for which we do not have a similar formula. Should we restore ἐν (τοῖς) Δι]ο̣σκορείοις? But what came before? ↓

]ς κουράτωρ vac. -- Δι]ο̣σκορείοις χαίρειν -- ξ]ύλα δ̅ ἀνθ̣’ οὗ vac. ] vac.

Καττίου, the well of Cattius Mentioned in the Trajanic corpus, but also in later ostraca, Καττίου is either the name of a well, or rather perhaps that of a sector of a metallon where there was also a well. Besides the entry ὑδρεύματι Καττίου in O.Claud. IV 700.10 (Trajanic), an account of workers broken down by assignments, and O.Claud. inv. 3114, where Kattiou is a place whence camels are returning laden with water, one notes also the phrase τὰ Καττίου (with a variant τὸ Καττίου) or τὰ Κ̣α̣τ̣τ̣ίου [μέ?]ρη (O.Claud. IV 760, but, there, I consider the dotted letters as uncertain, as well as the restoration). Kattiou is also a place that provides extraction tools. Occurrences of Καττίου are: O.Claud. IV 697.8 (Trajanic): an account of workers broken down according to place or kind of work (often in the dative), including Καττίου. O.Claud. IV 746 (Trajanic): a note appointing a certain Leonides εἰς τὸ Καττίου (implying ὕδρευμα?). O.Claud. IV 632.1–2 (Trajanic): εἰ̣ς̣ τὰ̣ | Καττίου is tempting (see O.Claud. IV 757), although there seem to be marks after τά (Fig. 36). O.Claud. inv. 2853.1 (Trajanic): a water-chart of the same series as inv. 1538. The first section, which concerns blacksmiths’ rations distributed to various stations, includes [the loading ramp (κρηπίς)], the Mese quarry, steelworks (στομωτηρίῳ), and Καττίου. O.Claud. inv. 3114 (Trajanic): a report on the movements of water-transport camels; 76 camels, loaded with water, came from Kattiou. O.Claud. IV 797 (Trajanic): a note about the sending of tools εἰς Καττίου. O.Claud. inv. 3342 (Trajanic): an acknowledgment for a skin; the ticket, written in charcoal, was prewritten in ink with the place of origin of the skin (ἀπὸ Καττίου) and the date. O.Claud. IV 757 belongs to a series of daily activity reports (Antonine); on that day, Epeiph 1, workers cleaned the trench of the column which was being extracted in the sector of Kattiou: τοῦ εἰς τὰ Καττίου στύλ(ου). O.Claud. IV 760, mentioned above, is the only other example of Καττίου during the Antonine period. There is no indication at this time that the well was still working. The topographic feature which Kattiou refers to is not clear. Used on its own, could Καττίου perhaps be considered as a name of a λατομία or as the abbreviation of the phrase ὕδρευμα Καττίου? Does τὰ Καττίου mean “Cattius’s sector” or perhaps, for short, “the sector (of the well) of Cattius”? Because the names of the other wells are not drawn from an anthroponym, I suggest that Καττίου was the name of an area, before being that of the well drilled there.

68

Rome in Egypt’s Eastern Desert

Figure 36. O.Claud. IV 632.1–2. © A. Bülow-Jacobsen

Σαλαεις This name, possibly Semitic and indeclinable (εἰς Σαλαεις, ἀπὸ Σαλαεις) is attested only in Trajanic reports on the movements of the camels entitled ἀπόλογοι ὑδροφορίας.171 At Salaeis, there was a well where camels were sent to fetch water.

Conclusion on well-names None of the names of these peripheral wells is a complex toponym with the postpositioned generic ῞Υδρευμα (cf. Ἀπόλλωνος ῞Υδρευμα). They are indicated either by means of a phrase where the toponym, in the genitive, is a complement to ὕδρευμα (τὸ Καρδαμήτου ὕδρευμα, τὸ ὕδρευμα τῶν Διοσκορείων) or by appending ὕδρευμα to a toponym, the use of articles following the rules of apposition (τὸ Ἀκάνθιον τὸ ὕδρευμα). If they are used without ὕδρευμα, the well names are declinable, except the one that is taken from an anthroponym, Καττίου.

VII. Roads In the road-names of the Eastern Desert, the generic constituent ὁδός, contrary to πόλις, κώμη, and ὕδρευμα, precedes the specific. The specific, except the particular case of the via Hadriana, is either the name of the terminal of a road in the genitive (usually its destination from a reference point situated in the valley),172 or a proprial adjective derived from that name. In the first case, ὁδός is followed by the name in the genitive of the metallon or seaport. The second case conforms to Latin usage, where the use of an adjective is preferred when forming a complex toponym which contains a generic.173 The road from Kaine to Mons Claudianus is called ὁδὸς Κλαυδιανοῦ in three passes issued by the centurion Antoninus, but ὁδὸς Κλαυδιανή only in one pass issued by the centurion Accius Optatus. The road from Kaine to Porphyrites is called ὁδὸς Πορφυρίτου (no adjectival variant). Unlike Claudianus, the Greek noun Πορφυρίτης normally cannot be used directly as an adjective: the corresponding adjective is πορφυριτικός. The phrase ὁδὸς πορφυριτική, though grammatically correct, is not attested, but we find the adjectives πορφυριτικός and Κλαυδιανός in a toponymic phrase in 214/5: τοῖς Πορφυρειτικοῖς καὶ Κλαυδιανοῖς μετάλλοις.174

171. O.Claud. inv. 1287, 1288, 1306, 1378, 1530, 1801, and 3322. 172. In the case of the minor road to Akanthion, this reference point is Claudianus. 173. Valat 2008: 249. That criterion allows a more precise date for I.Mylasa 214 (=IK 35), for which the editor does not offer a date and which McCabe, in the Searchable Greek Inscriptions, dated second to first century BC: in line 12 there is a mention of a Τροβαλισσικὴ ὁδός. The use of the adjective suggests that this inscription is in any case later than the Roman takeover of Caria that occurred at the end of the Republic. Another indication of the Roman influence is provided in line 1 by the toponym Ὀμβιανὸν πεδίον, the proprial adjective being provided with a Latin suffix. 174. P.Oxy. XLV 3243.14.

A survey of place-names in the Egyptian Eastern Desert

69

The road from Koptos to Myos Hormos is sometimes called ὁδὸς Μυσόρμου, sometimes ὁδὸς Μυσορμιτική: the use of a derivative adjective, although in this case with a Greek suffix, perhaps betrays the Latin influence. The road to Berenike is always written as ὁδὸς Βερενίκης, with the sole exception of Βερνικησία ὁδός in O.Dios inv. 457 (219), where the order of the terms is unusual for a road-name. The suffix -ήσιος serves to transpose into Greek the situative Latin suffix -ensis. This suggests that, in Latin, the road was called via Berenicensis. In O.Claud. I 77 and 81 [ὁδὸς] Ἀκανθίου is probably a small road, from Claudianus, to one of the wells on which it depended for its water supply, the Akanthion. The Latin name of via Hadriana was formed in the Medieval period using the only evidence of this toponym, which appears only in Greek in the dedication of this road, inscribed on a pedestal erected in Antinoou polis (I.Pan 80.8): ὁδὸς Καινὴ Ἁδριανή. The inscription shows that, in the mental representation of the developers of the via Hadriana, the orientation of the road was reversed compared to other roads of the Eastern Desert, since its destination was not Berenike, but Antinoou polis: ὁδὸν Καινὴν Ἁδριανὴν ἀπὸ Βερενίκης εἰς Ἀντινόου.

VIII. Ports of the Red Sea 1. Berenike, Berenicis, and Mons Berenicidis The southernmost area of the Egyptian Eastern Desert took the name Desert of Berenike in about 4 BC, once the Romans had revived the Ptolemaic port of Berenike and built a new road from Koptos to that port. The term “Desert of Berenike” is abundantly attested. In Greek, it is Ὄρος Βερενίκης. In Latin, one encounters the expression in inscriptions mentioning the prefect of Berenike and we see that the name of the city was then, with one exception, suffixed by -is. We find once praefectus praesidiorum et Montis Beronices, but three times praefectus Montis Berenicidis or simply praefectus Berenicidis. There is also the ablative Berenicide in the large inscription from Koptos commemorating the construction of water-tanks by the army (ILS 2483 = I.Portes 56).175 However, the port of Βερενίκη on the Red Sea was never called Βερενικίς in Greek. There is a parallel phenomenon with Βερενίκη in Cyrenaica, which is called Berenicis in the Pharsalia of Lucan (9.524) and in the Punica of Silius Italicus (3.249).176 Editors justify this discrepancy considering that Berenicis is suffixed following the model of choronyms such as Ἀργολίς, Μεγαρίς, and Περσίς, where the generic γῆ or χώρα is elided.177 The suffix -is would mean that the referent is not the city alone, but the city and its chora. I admit that it would provide a very satisfactory solution in the case of the title praefectus Berenicidis. This is probably the reason why F. De Romanis interpreted the toponym Berenicis as a choronym in Latin inscriptions, which leads him to consider that the inscription ILS 2483 refers to the construction of a well in Berenicis (i.e., the region of Berenike), and that this well is Καινὸν Ὕδρευμα, which is 25 km above Berenike.178 However, we should abandon the idea that, in the case of Βερενίκη/Βερενικίς, the suffix -ίς denotes a choronym. In the Peutinger Table and Ravenna Cosmography, Berenicide clearly designates 175. Below are the last lines of the inscription: per eosdem qui supra scripti sunt, lacci aedificati et dedicati sunt: Apollonos Hỵdreuma VII K(alendas) Ianuarias, Compasi K(alendis) Augustis, Berenicide XVIII K(alendas) Ianuar(ias), Myos Hormi Id[ib]us Ianuar(iis) castram aedificaverunt et refecerunt. 176. In the case of these two Latin poets, the use of the suffixed form is explained perhaps only by the needs of the meter. 177. Chantraine 1933: 339. None of the examples cited by Chantraine is derived from a town or village name. The RE correctly points out that Berenicis in the two Latin poems is the city itself, not its surroundings (RE III.1, 282 [8]). 178. De Romanis 1996: 175, n. 23. This interpretation is further vitiated by interpreting laccus as “well” instead of tank.

70

Rome in Egypt’s Eastern Desert

the end point of the road, not a region, and corresponds to Berenike in the Antonine Itinerary, which is more faithful to the Greek usage. Two toponyms of the Arsinoïte nome, Βερενικὶς Θεσμοφόρου and Βερενικὶς Αἰγιαλοῦ, are examples of the suffix referring to simple villages, and not to districts. Therefore, it appears that, when -ίς is added to a dynastic anthroponym, it is to characterize it as a toponym, not to designate a region. The example of the two villages shows that when a settlement was called after a Queen Berenike, the naming authorities had the option of keeping the name unchanged, or of adding a suffix to emphasize that this was a toponym.179 In the case of the port of Berenike, there was a degree of fluctuation in Latin, but not in Greek. Frédérique Biville, whom I consulted on this Latin initiative to add a Greek suffix to a Greek base where the Greek model never appears to have a suffix, replied: “It is not surprising, I think, that there were Latin-speakers who felt the need to over-characterize the name Berenike: we also see that there is what could be called ‘the Greek of the Romans,’ Greek forms and words created by the Romans, which are only documented in a Greco-Roman context.”180 In the Greek and Latin names of the Desert of Berenike, the generic ὄρος/mons does not designate a mountain as in classical Greek or Latin, or as in the phrase Mons Claudianus, but it is a calque of the Egyptian ḏw.181 Unaware of this Egyptian peculiarity, the editors of the latest edition of Ptolemy’s Geography return to the old idea that Berenicidis mons could be an alternative name for “Mons Smaragdus.”182 The Romans showed a double conservatism, first by reproducing, through the Greek, an Egyptian word, and secondly by maintaining the dynastic name of the port of Berenike (which is nothing exceptional). But they were highly innovative in renaming the region after this port so far out, while the desert was named during the Ptolemaic period by reference to Koptos, as shown by a Greek inscription from 130 BC, where it is called τὸ κατὰ Κόπτον ὄρος, literally “the desert adjacent to Koptos.”183 This expression is also the calque of the ancient Egyptian name ḏw Gbtyw. One can question this desire to put Berenike in the limelight at the expense of Koptos, while Roman Koptos was a thriving city with a monumental center, whereas Berenike, devoid of an agricultural hinterland, was at the end of the earth—so much so that the governor of the Desert of Berenike was sometimes called, for short, the prefect of Berenike, while its offices were at Koptos and he cannot often have come to Berenike. The idea behind the name was to suggest that Berenike, rather than being on the borders of the empire, was now at the center of a zone of Roman influence that extended well beyond it. There was perhaps some truth in this view, when we consider that, at least under Antoninus Pius, but perhaps even before, the Romans had a military base in the remote Farasan Islands.

179. It is the same for Ἀρσινόη/Ἀρσινοίς, Κλεοπάτρα/Κλεοπατρίς, and Φιλωτέρα/Φιλωτερίς (for the latter name, see pp. 72 f.). 180. Message from 15 April 2013. 181. When one thinks about it, Greek and Latin did not have a noun to refer to this geographical feature, other than ἐρημία, solitudo, which are perhaps too suggestive. 182. Klaudios Ptolemaios, Handbuch der Geographie, ed. A.  Stückelberger and G.  Grasshoff I (Basel 2006) 425, n.  123. H. Verreth (pers. comm.) observes that this confusion must go back to the Barrington Atlas 2000, pl. 80, F4 (Berenicidis Mons = Smaragdos Oros). This edition of the Geography also wrongly distinguishes between Berenike in 4.5.15 and Berenike Trogodytika (with reference to Calderini, Diz. Geogr. II 40 and K. Sethe, Berenike [5] in RE III.1, 280 ff.). 183. I.Pan 86. The parallel expression τὸ κατὰ Συήνην ὄρος attested in OGIS 168.11 and 14 = I.ThSy. 244.40 and 54 (AD 115 ) does not designate a vast desert region such as the Desert of Koptos or Berenike, but only the quarry area of Syene; it is a descriptive gloss rather than a toponym and, if we consider it as a toponym, it could be the name of a metallon.

A survey of place-names in the Egyptian Eastern Desert

71

2. Origin of the toponym Myos Hormos Myos Hormos184 is a Ptolemaic foundation, but its picturesque name contrasts with the series of dynastic names conferred on Ptolemaic port foundations on the Red Sea. Agatharchides of Knidos (floruit c. 150 BC) attributes an alternative name to it, Ἀφροδίτης Ὅρμος, which, according to the version of his treatise on the Red Sea transmitted by Photius, had replaced, in his time, the original toponym: ἐφεξῆς δὲ λιμὴν μέγας ἐκδέχεται, ὃς πρότερον μὲν Μυὸς ἐκαλεῖτο ὅρμος, ἔπειτα δὲ Ἀφροδίτης ὠνομάσθη, “immediately after, there comes a major port formerly known as the anchorage of the Mouse and which was called afterwards the anchorage of Aphrodite” (Phot., Bibl. 250.81). If we are to believe Agatharchides, this port offers a case of competition between a spontaneous, popular naming and one that has a more formal flavor and would have temporarily replaced the first. Μυὸς Ὅρμος was definitively imposed during the imperial period. The alternative name Ἀφροδίτης, of which one no longer hears in Roman times, remains an enigma: on what occasion and why would the toponym have been changed? In most of its occurrences in the ostraca from Krokodilo, the complex toponym Μυὸς Ὅρμος has become a compound, Μύσορμος.185 This is probably a back-formation from an unattested *Μυσορμίτης,186 from which is derived the adjective Μυσορμιτικός, employed as a specific in the road name ὁδὸς Μυσορμιτική.187 In versions of the treatise passed down by Diodorus Siculus (which mentions only the theotoponym) and Strabo, we also learn that the entrance to the anchorage is curved: … κεῖται λιμὴν σκολιὸν ἔχων τὸν εἴσπλουν, ἐπώνυμος Ἀφροδίτης (Diod. 3.39.1); εἶτα Μυὸς Ὅρμον ὃν καὶ Ἀφροδίτης Ὅρμον καλεῖσθαι, λιμένα μέγαν, τὸν εἴσπλουν ἔχοντα σκολιόν (Strabo 16.4.5). Note that the word translated as anchorage, ὅρμος, can be used figuratively to mean “refuge.” Scholars have been intrigued by the name Μυὸς Ὅρμος. In the nineteenth century the Egyptologist H. Brugsch, followed more recently by F. De Romanis, tried to explain it through the corruption of the Egyptian toponym mstj contained in the list of foreign peoples of Tutmosis III at Karnak. Not being an Egyptologist, I will not dwell on this hypothesis, which implies that Myos Hormos would have been a very old foundation, for which there is no archaeological evidence. Other researchers have proposed that in this toponym, the Greek word μῦς (genitive μυός), “mouse,” took on one of its other meanings, “mussel,” which would make more sense for a coastal town. The map of the port published by the British mission of David Peacock and Lucy Blue suggested another hypothesis to me, which I have already presented in the article “Myos Hormos” in The Encyclopaedia of Ancient History: Myos Hormos had the distinction of having been built in an inland lagoon, now silted up. This lagoon was connected to the sea through a narrow opening where the coral-reef was interrupted because of the fresh water of the wadi floods that poured into the lagoon. The ships had to go through a curved channel, a feature described by Agatharchides of Knidos in such a way as to suggest that it was a striking sight for the eyewitness. I think the image behind the name Myos Hormos is that of

184. Today Qusayr al-Qadim. 185. Seven occurrences of Μυὸς Ὅρμος against nineteen of Μύσορμος/Μυσορμιτική. Some manuscripts of the Geography of Ptolemy present the reading Μισηρμος, Μισορμος. 186. But cf. Ὡρμίτῳ in P.Berenike II 129, comm. ad 22 (“It is attractive to suppose that the writer meant ὁρμίτῃ, which should be capitalized and connected to Myos Hormos.” 187. I thank Claire Le Feuvre for helping me with this question, and I refer to her paper, Le Feuvre 2018, esp. 197, where she remarks that in Greek, as well as in Latin, “the addition of a derivational morpheme triggers the univerbation of the underlying syntagm.”

72

Rome in Egypt’s Eastern Desert

Figure 37. Maximianon (Al-Zarqaʾ), Myos Hormos, Biʾr Karim (from Meredith 1958). © Rights reserved

boats sneaking like mice into a small hole to find refuge. Some types of light ancient ships were, indeed, named after the mouse, such as myoparon or musculus.188

3. Philotera, Philoteris, and the Philoterion Ancient writers mention a port on the Red Sea, named after a sister of Ptolemy Philadelphos and founded by Satyros, an explorer of the Trogodytike for capturing elephants.189 Stephanus Byzantius (Ethnica, ed. Meineke, p. 666) reports that besides Φιλωτέρα, the suffixed form Φιλωτερίς is known, which was retained by Pomponius Mela, Chorogr. 3.80.190 Pliny, Nat. 6.168, cites Philoterias191 as an alternative name for an oppidum parvum otherwise known as Aenum. Ptolemy places Φιλωτέρας λιμήν directly south of Myos Hormos (Geogr. 4.5.14), whereas for Artemidorus and Pliny, Philotera is at the north end of the Gulf of Suez. Ostraca found in Myos Hormos and, to a lesser extent, at Maximianon and Krokodilo, show that the memory of the Ptolemaic princess remained significant in the area of Myos Hormos. Ostraca from Myos Hormos mention both a port called Philoteris (a boat went there from Myos Hormos) and a Philoterion (preceded by the definite article): are the two mixed up? Or was the Philoterion the temple of which scarce remains were found at Qusayr, a few kilometers south of Myos Hormos? Or is it the temple erected at Biʾr Karim, where one of the wells was still supplying water to Qusayr in the nineteenth century?192 These hypotheses are discussed by W. Van Rengen, who also uses as evidence the letters on ostraca found at Maximianon, including proskynemata to the goddess Philotera.193 Where do these letters come 188. The late David Peacock liked this idea and wrote back: “I think your suggestion is a good one. The entrance is narrow and divers have seen a reef in the middle. Compared with the other Red Sea ports it must have been a pain to get into—and there is the wreck at the mouth as proof!” (email of 27 November 2010). 189. Artemidorus ap. Strabo 16.4.5. 190. On this passage of Pomponius Mela, see Cohen 2006: 312. 191. Mox oppidum parvum est Aenum—alii pro hoc Philoterias scribunt. But Philoterias (H. Verreth pointed out to me that it must be an accusative plural) is a conjecture by Mayhoff: all manuscripts give the final -ria or –rias, while the beginning of the name is more or less corrupted. On Philotera see Desanges 2008: 53. 192. M. Prickett, in Whitcomb and Johnson 1979: 271. Biʾr Karim: 25° 55’ 53” N/34° 03’ 27” E. 193. Van Rengen 2018.

A survey of place-names in the Egyptian Eastern Desert

73

from? Considering that these proskynemata to Philotera matched those to Athena contained in the letters from Persou (Biʾr Umm Fawakhir), we considered at the outset that the letters citing Philotera were written in the praesidium that came after Maximianon in the direction of Myos Hormos, presumably the Simiou of our ostraca, if we accept the Ptolemaic origin of this toponym (it would derive from the name of the explorer Simmias).194 Although the aedes of that praesidium could have been called a Philoterion (thus, the aedes of Xeron, a fortlet where epistolary proskynemata address Athena, is called “the Athenadion” in a report on a nighttime incident), it is unthinkable that a military chapel enjoyed such fame that it was spoken of at Myos Hormos. The Philoterion was a larger feature; the name of the temple may have been extended by metonymy, to the place where it was. To evidence of Philoterion in the Myos Hormos ostraca we can add a letter found in Maximianon, which also contains one of these proskynemata before Philotera.195 It is, unfortunately, very fragmentary, but it suggests that letters arriving at Maximianon with proskynemata before Philotera may not have been written in a praesidium on the road, but at Philoterion, wherever it might be. Biʾr Karim would not be a bad candidate: located south of the hodos Mysormitike, it is the same distance from Maximianon as Biʾr Sayyala (Simiou according to us), and the topography allows one to reach Maximianon without having to go round the mountain (Fig. 37).196 The surface ceramics date from the Roman period,197 but there are traces of Ptolemaic occupation.198 The toponym Philotera/Philoteris does not appear in the ostraca of the praesidia except perhaps in O.Krok. I 46.6, a fragment of a liber litterarum which reads -ελ]η̣λυθέναι ἰς Φιλ[. Finally, it should be noted that the name Philotera, which is not common in Roman Egypt, is carried by a young woman from the circle of Philokles, a food supplier and pimp who operated in the time of Trajan in the praesidia of the northern part of the Desert of Berenike.

IX. Unidentified topographic features Ἵππος O.Krok. I 120 is an activity report dated Pachon 7 and written in the first person by K---s, a signifer, who is said to be going on reconnaissance with the cavalryman Marinus ἕ̣ω̣ς τ̣ο̣ῦ Ἵπ̣π̣ο̣υ̣. The infrared photograph made since the publication failed to provide a more accurate reading, but neither did it challenge it (Fig. 38): it remains possible and even quite probable. Since these daily notes are usually prepared by the curator praesidii, and since signifer is the only rank we know of for this non-commissioned officer, K---s is probably the curator himself, who took with him one of the cavalrymen of his garrison. Perhaps τοῦ ἵππου must be taken merely as a common noun (a cavalryman was attacked between two praesidia, the corpse of his horse left behind became a temporary landmark). However, we can read in an ostracon from Maximianon, in a lacunose context, ἐμὲ ὄντα ἰς ἵππου πετρ̣[, “me being at Hippou,” or, if we restore πέτρ[αν, “at Hippou Petra,” “the horse rock” (Fig. 39).

194. A. Bülow-Jacobsen, in Cuvigny (ed.) 2006: I 56. 195. O.Max. inv. 1149. We read line 7: ἀλλὰ ἐν τῷ Φιλωτε[ρίῳ]. 196. I make my argument using the map of Meredith, having never been to Biʾr Karim myself. 197. Whitcomb and Johnson 1982: 292. 198. See Van Rengen 2018: §16. The site of Biʾr (Wadi) Karim is described, with a good satellite image, in Klemm and Klemm 2013: 148–51.

74

Rome in Egypt’s Eastern Desert

Figure 38. O.Krok. I 120. © A. Bülow-Jacobsen

Figure 39. O.Max. inv. 1099.7. © A. Bülow-Jacobsen

Κάνωπος Address of the destination of a letter found in Didymoi (εἰς Κάνοπον, O.Did. 370.1). Ξηρὸν Πέλαγος Homonym of a praesidium on the Berenike road, this site, probably near Mons Claudianus, is mentioned in three ostraca from the Trajanic period. One of them, O.Claud. I 141, gives the impression that it is a quarry. The author of the letter says that he passed Xeron Pelagos, where he met the centurion Crispus and another man who told him: κατασπῶμεν τὸν λουτῆρα, “we brought the tub down.” The reading of the editio princeps has been corrected, cf. BL XI 294 f., where the verb has, however, been misunderstood. It is not about knocking down, i.e., starting the exploitation of the Louter quarry in Mons Claudianus, but bringing down a sink or tub from the quarry at Xeron Pelagos; the meaning of κατασπᾶν and of the verbal derivative κατάσπασις is ascertained through several ostraca from Umm Balad (e.g., P.Worp 50.11–12).

A survey of place-names in the Egyptian Eastern Desert

75

A. Bülow-Jacobsen has suggested that this Xeron Pelagos could be the small quarry at Wadi Fatira al-Bayda (26° 44’ 01” N/33° 19’ 25” E), to which J. Harrell drew our attention.199 As the crow flies, the site is 18.30 km southwest of Claudianus, and 10.5 km northeast of Abu Zawal (Raïma?). There is no fortlet there, but cellae, some of which form a row. Further west along the same wadi are other traces of Roman mining, where S. E. Sidebotham noted the inscription of the possible procurator Diadumenus.200 Ὀστρεών This toponym (“place of oysters/shells”) appears only in O.Krok. I 47 (AD 109), in the report of a skirmish where Phoinikon is also mentioned (l. 14: ἀπὸ Ὀστρεῶνος; l. 22: ἐ]ν Ὀστρεῶνι). The site was therefore not far away, but the damaged context does not give us any more information. τὸ Συκου, τὸ Συκα There are only two occurrences of this microtoponym variously spelled in two ostraca in the series of ἀπόλογοι ὑδροφορίας (O.Claud. inv. 1530 and 2470 [Trajan]). Each time, only one camel loaded with 4 waterskins leaves for that destination (εἰς τὸ Συκου, εἰς τὸ Συκα). Is the toponym derived from τὸ σῦκον “fig”? It does not seem abbreviated (in this series, the abbreviations are always signalized with a graphic mark). Two indeterminate toponyms result from misreadings and are ghost-names: Σιαροι In O.Max. inv. 639.12–13 (Cuvigny (ed.) 2006: I 57 = SB XXVIII 17083) εἰς Σια̣ρ̣ους̣ → εἰς Σιμί̣ου (interpunction). Σμιλία O.Claud. IV 841.66. πρὸ̅ σμειλίων → πρὸ ϛ̅ μειλίων, “before 6 miles.” This is an indication of distance. Distances were estimated in the desert in Roman miles.201

X. General conclusion 1. Place and preterition of the generic element Generics to be considered in the analysis of the toponyms of the Eastern Desert are: ὁδός, λατομία, μέταλλον, ὅρμος, ὄρος, πραισίδιον, and ὕδρευμα. They do not all behave in the same way in the formation of names: ὁδός is necessary, while elision is almost the rule for μέταλλον and πραισίδιον except in the title of the curatores; some generics (λατομία, ὅρμος, ὕδρευμα) may, like πόλις or κώμη, be placed after the specific to form a complex toponym (which I have chosen to highlight using the uppercase). Ὁδός and ὄρος normally precede the specific element which, except in the case of the via Hadriana, is a city name in the genitive. Ὁδός is never omitted. 199. Email of 20 March 2017. 200. Sidebotham 1996: 190–2 and pl. XIX. A. Bülow-Jacobsen and I have vainly sought this inscription Friday, January 25, 2011. 201. See O.Did. 44 and the letter of Nemesous O.Did. 400, which shows that even a procuress referred to this unit of measure.

76

Rome in Egypt’s Eastern Desert

The toponym Ὄρος Βερενίκης and its Latin equivalent Mons Berenicidis/Berenikes have the particularity of being shortened in two ways: as part of the title of the territorial prefect, it can be reduced to one or the other of its two components: ἔπαρχος Ὄρους Βερενίκης, ἔπαρχος Ὄρους, ἔπαρχος Βερενίκης. The latter formula, which is ambiguous, is the least common; it may be early: of the four examples, all epigraphic, two are dated from the time of Augustus and Tiberius, respectively.202 Such metonymies are characteristic of administrative denominations (cf. the ancient city-states, and today Quebec or Mexico). When Ὄρος/Mons is used alone to mean “Desert (sc. of Berenike),” it can be considered as an “appellative” in the narrow sense fit for toponyms.203 In toponyms, elision of the specific is normally encountered only in a local and familiar context. It is, therefore, remarkable that the term ἔπαρχος Ὄρους was used in official contexts and even outside the desert (the petition P.Turner 34; ἐπίτροπος Ὄρους in the dedication from Koptos I.Portes 86). Usually omitted, except in the title of curators, μέταλλον and πραισίδιον204 are not included in the toponym with two exceptions.205 They always precede the specific element: κουράτωρ πραισιδίου Μαξιμιανοῦ, κουράτωρ μετάλλου Κλαυδιανοῦ, εἰς πραισίδιον Μαξιμιανόν.206 Λατομία and ὕδρευμα may be placed after the specific. A complex toponym thus formed is sometimes abbreviated. This is the case with Ἀπόλλωνος/Ἀπόλλωνος Ὕδρευμα, but it never happens when the specific component is an adjective (Καινὴ Λατομία and Καινὸν Ὕδρευμα are never abbreviated to Καινή and Καινόν). When a complex toponym is abbreviated, it is the second element that goes (Ὄρος Βερενίκης → Ὄρος,207 Ἀπόλλωνος Ὕδρευμα → Ἀπόλλωνος, Ἀφροδίτη(ς) Ὄρους → Ἀφροδίτη(ς), Ξηρὸν Πέλαγος → Ξηρόν). Unlike Ἀπόλλωνος Ὕδρευμα (or toponyms with πόλις), Μυὸς Ὅρμος is never abbreviated to Μυός. However, it often becomes a compound, Μύσορμος (see supra p. 71). When Ἀπόλλωνος Ὕδρευμα is abbreviated, Ἀπόλλωνος remains in the genitive, which is not the case for λατομία Ἀπόλλωνος at Mons Claudianus: when the generic is elided, the specific theophoric (as well as the anthropophoric) inflects. The rule is, therefore, different for λατομίαι and πραισίδια. Indeed, if we compare λατομία Ἀπόλλωνος with the similar phrase πραισίδιον Ἀφροδίτης, we see that Ἀφροδίτης remains (mostly) in the genitive when πραισίδιον is omitted.

2. Use of the article Analysis of the use of the article with toponyms is complicated by the scarcity of examples and by the flexibility of use. Even when the rule is that a toponym takes the article, it tends to be omitted after a preposition (this is not unique to toponyms, but all nouns),208 especially in documents where brevity is favored, namely lists and accounts.

202. I.Pan 51 (AD 11), ILS 2698 (Tiberius), I.Pan 68 (76/7), I.Memnon 14 (s.d.). 203. See p. 42. 204. No more than σταθμός in the Ptolemaic era: this is the noun which refers to desert way stations in the ostraca from Biʾr Samut, which, in the third century BC, is one of these σταθμοί on the road from Edfu to Berenike. The only appellative used then as a generic element in a complex toponym is ῞Υδρευμα. 205. ἀπὸ Μέλανος Ὄρους πραισιδίου (O.KaLa. inv. 637); κουράτωρ Κλαυδιανοῦ μετάλλου (O.Claud. II 371). 206. SB XXVIII 17096.5–6. 207. But by exception (see above) the ellipse may affect the generic in the phrase ἔπαρχος Βερενίκης. 208. Mayser, Grammatik II.2.1, 14.

A survey of place-names in the Egyptian Eastern Desert

77

For obvious reasons, most of our topographic features are not included in the pages that Mayser209 dedicated to the use of the article for toponyms (countries and islands, towns and villages, mountains, rivers, sanctuaries, urban neighbourhoods, squares, etc., which rank in the category Lokalnamen). Let us revisit the observations of Mayser: Table 1.7. Use of the article before place-names Countries, Regions

– countries: use ad hoc; generally with Ἀσία or Αἴγυπτος, exceptionally with Αἰθιοπία – Egyptian nomes: always the article

Towns, Villages

do not take the article. Exceptions: – in the case of the repetition of a toponym (anaphoric article) – when the toponym is a generic appellative, it may take the article – in an abbreviated phrase, a complex toponym sometimes takes the article – foreign toponyms sometimes take the article

Mountains, Rivers

take the article, which can be omitted after a preposition

Microtoponyms (urban areas, temples, squares)

take the article, also after a preposition (except in a concise style)

What do we observe in the Eastern Desert? The article is used more easily before toponyms that are or contain common nouns. The impression emerges that the article is stronger in the Ptolemaic ostraca from Biʾr Samut where, even after a preposition, αἱ Πύλαι and τὸ Σαπαρ always take the article, and Ῥάμνος almost always, which suggests that these places were sites of minor importance. In the Roman period, the names of praesidia behave like names of cities or villages, and do not normally take the article, even when they are common nouns; though there is an exception, in the case of εἰς τὴν Καμπήν and especially Ἄκανθα in ostraca from Umm Balad: out of eight examples, four take the article. The names of latomiai can take the article even if they are proper nouns (εἰς τὸν Διόνυσον). Usage is flexible for Πορφυρίτης and Κλαυδιανόν,210 perhaps because these metalla are treated as regions. The name of the small metallon of Καινὴ Λατομία almost only appears as the destination address on amphoric tituli (about 70 examples); we note only one example of εἰς τὴν Καινὴν Λατομίαν (O.KaLa. inv. 435). Δομιτιανή (16 examples) is never preceded by the article.

3. The toponymic program of the Romans in the Eastern Desert The Romans have profoundly marked the Eastern Desert. They excavated the mountains from where they extracted huge monoliths for monuments built by the emperors. They appropriated space by equipping it with roads lined with fortlets in order to make the desert passable for travellers, and to implant an effective communication system, keeping the Beduins in check. Nevertheless, their toponyms are simple, when compared with the warlike names that the pharaohs of the Middle Kingdom gave their menenou (strongholds) in Nubia: “Who subjugates Setyou,” “Who pushes Medjayou,” “Who secures foreign countries,” “Who kills the desert dwellers.”211 In Roman times, they relied more on the virtues of diplomacy than on the performative magic of words. 209. Mayser, Grammatik II.2.1, 13–18. 210. Ἐν τῷ Κλαυδιανῷ: O.Claud. inv. 7294; 7484; P.Claud. inv. 32. Also mentions of the Τύχη τοῦ Κλαυδιανοῦ. 211. Somaglino 2017. The antecedent of “who” is more likely the Pharaoh than the fortress itself.

78

Rome in Egypt’s Eastern Desert

Neither did the Romans target a symbolic appropriation of space by replacing the existing names: they retained the old names, even the dynastic Ptolemaic ones, and even local cults that deified Ptolemaic princesses. Most of the names are Greek or Latin, but some belong to other languages: - Kabalsi?, praesidium on the road to Berenike; - Kompasi, praesidium on the road to Berenike, a former gold mine dating back to the Pharaonic era; - Patkoua, praesidium, perhaps in Lower Nubia; - Thonis Megale, praesidium, perhaps in Lower Nubia (Thonis is an Egyptian word); - Raïma, praesidium on the road to Claudianus; - Sabelbi, praesidium on the road to Porphyrites; - Salaeis, a well close to Claudianus; - Senskis, presumably a district of Smaragdos;212 - Tamostymis (Egyptian), mine or quarry in the area of Wadi al-Hammamat/Wadi al-Fawakhir. One can add to this list the names of three metalla named after the material which was extracted and whose names belong to the language of populations who traditionally carried out the exploitation: Μαργαρίτης is a loan-word, borrowed from Pahlavi marvārīt, “pearl.” This word is attested for the first time in Greek by Theophrastus, who was on the lookout for discoveries made by the explorers of Alexander the Great. But the main pearl fisheries were in the Persian Gulf. Σμάραγδος is an oriental word, already known by Herodotus, that applies to different varieties of green stones. Τοπάζιον/Βαζιον is the name of topaz and of St. John’s Island where it was mined. Among these three toponyms, it is the only one that could have local origins, if we follow Pliny, who thinks that it is, according to Juba, a word borrowed from the language of the Trogodytes (Nat. 37.109).213 Exotic names mentioned in the ostraca, some of which are Semitic, raise an insoluble problem. It is not possible to decide if they belong to a Beduin substrate predating the arrival of the Romans, who would have chosen to keep the names, or if they were bestowed by Roman officers of Eastern origin. In favor of the former hypothesis, we should look at the recurring microtoponym τὸ Σαπαρ, probably of Arabic origin, which refers, in ostraca from Biʾr Samut (third century BC), to a small site without a well, located in the vicinity of the Ptolemaic fort. Such toponyms probably betray the intervention of Beduin guides in the exploration of the desert by the Romans. The Romans were moderate in their use of imperial eponymy, which they reserved for a few metalla and latomiai. Indeed, it was after a discovery phase of the physical geography of the Eastern Desert that they bestowed imperial names. The first metalla discovered were named after the material extracted from them, and the Romans gave thanks in Greek to the local Egyptian deity (Min) for the discovery of minerals. Apart from dynastic names, the toponyms they invented are apolitical, innocuous, and sometimes even rather bland (cf. the series of Kaine/Kainon).

212. See supra, n. 38. 213. J. Desanges doubts this interpretation. He believes that Pliny misunderstood its source, and that Juba had only written that the island takes its name from the Greek verb τοπάζειν (Desanges 2008: 62).

2 Ulpius Himeros, imperial procurator (I.Pan 53) I propose here two corrections to the inscription I.Pan 53, which comes from Wadi Samna; they have been included in the text given below.1 The first edition,2 established on the basis of three non-joining fragments (A, B, C), is owed to L. A. Tregenza and A. H. M. Jones (Tregenza 1950); it was accompanied by facsimiles and two photographs of fragments A and B, each at a slightly different scale.3 The text was reprinted in capital letters by A. Merlin in AE 1952, 249 (without dotting letters) and again by A. Bernand in Pan du désert (I.Pan 53, without illustration). None of these authors gives dimensions for the fragments. The three fragments, handed over by Tregenza to the Greco-Roman Museum in Alexandria, are kept there under the inventory number 25876.4 Since the making of the photos published by Tregenza, they have been encased in a plaster plaque, and only the area with letters remains visible. Fragment C is clearly foreign to the main document (see below): unlike my predecessors, I shall use only fragments A and B in the establishment of the text.

1. The text in my version is reprinted in AE 2001, 2026. 2. It is given as a kind of restored facsimile rather than as a typeset text (Tregenza 1950: 87). 3. This might awaken doubts about the positioning of fragments A and B with respect to one another as proposed by Jones and Tregenza, which I have in the end kept: the height of the letters and spacing of the lines are identical in the two fragments. Actually, if lines 2–3 and 3–4 are consecutive, the sequence iusso legati in lines 4–5 presents us with an aporia. 4. Fr. Kayser informed me that they were located in a storeroom which, at the time when he was working in the Museum on his corpus of the Latin inscriptions of Egypt, had become inaccessible because of a collapse. It seems that the inscription had not been examined since Tregenza (Jones did not see the original). I owe to J.-Y. Empereur, then the director of the Centre d’études alexandrines, and to Mervat Seif ed-Din, curator in the Greco-Roman Museum, its retrieval. Madame Dorreya Saïd, director of the Museum, authorized me to consult the stone and permitted A. Lecler, the IFAO photographer, to re-photograph it. I thank them all for their help. I also thank Fr. Kayser for his comments on a first draft of the original article, as well as J. Bingen, who provided me with a copy of Tregenza 1950, which was unavailable in French public libraries. Thanks also to Karen van Opstal and Raymond Bird, who gave me the opportunity to revisit the Samna dossier and who really provided the impulse for this chapter.

81

82

Rome in Egypt’s Eastern Desert

The ancient name of the establishment of Wadi Samna, Ophiates, is known from I.Pan 51.11 (ἐν τῶι Ὀφιάτηι).5 I. Pan 51 was originally situated in the interior of a small sanctuary located in an area of ancient granodiorite quarries, along with a workers’ village (a nearby rectangular enclosed camp with dispersed huts); it is the dedication of the sanctuary itself, a Paneion, by Publius Iuventius Agathopous, freedman of the prefect of Berenike Publius Iuventius Rufus; Agathopous enjoyed at this early period (we are in AD 11) the baroque title of ἐπίτροπος καὶ προνοητὴς καὶ εὐεργέτης πάντων τῶν μετάλλων τῆς Αἰγύπτου, “procurator, administrator, and benefactor of all the mines and quarries of Egypt.” The village and quarries are located in a tributary wadi that flows into Wadi Samna about 2 km from them; at this junction there is a small fort that Tregenza called a “Roman castellum,” a term to which we would

Figure 40. Sites in Wadi Samna (drawing by Tregenza, after BIFAO 96 [1996] 91). 5. Contrary to what I wrote in the first version of this chapter, we should not suppose Ὀφιάτης is a shortened form of a toponym composed of the element ὄρος, and we must exclude the possibility of a Latin place-name of the form *Mons Ophiates (cf. Chapter 1, pp. 14–16). Rather, we are dealing with the name of a stone, here used as a toponym, and this stone is indeed the ophite of Pliny (Nat. 36.55) and Lucan (Phars. 9.714: paruis tinctus maculis Thebenus ophites). Pliny, after mentioning the stones discovered in Egypt in the reigns of Augustus and Tiberius and named for them (marmor Augusteum, Tibereum), felt immediately the need to explain in detail how they are visually distinguished from ophite; the three materials thus must have resembled one another. The Augusteum has not been identified, but Clayton Fant thinks that the Tibereum can be identified with the granodiorite of Barud, of which fragments in opus sectile have been identified in the Domus Tiberiana (Fant 1993: 150, n. 43); the discovery of the ancient name of Barud, Tiberiane, which is attested in the ostraca of Claudianus and Barud, supports this hypothesis. As it happens, Barud is not far from Samna. Pliny does not say anything about the source of ophite; in Nat. 36.56, he refers to a “Memphite” rock which is useful against snakebite, but it it is not clear to me if he identifies this with the ophite which he describes earlier.

Ulpius Himeros, imperial procurator (I.Pan 53)

83

Figure 41. I.Pan 53 (AE 2001, 2036) © A. Lecler (IFAO)

prefer praesidium, the word used in the ostraca from the Eastern Desert; this praesidium controlled the access to the quarry zone.6 I.Pan 53, on the other hand, was found by Tregenza in February 1950 in what he called “Site 6”7 or the “gold-crushing site”8 of Wadi Samna; according to his summary and scaleless map (Tregenza 1950: 86, reproduced here as Fig. 40), and the map of the region drawn up by Meredith,9 this site for gold extraction was about ten kilometers northwest of the praesidium. There would be less than 10 miles (16 km) between “Site 6” and the quarries (Tregenza 1951: 50). The mention of the prefect of Egypt L. Munatius Felix allows us to date the inscription approximately: the earliest attestation of this individual is from 17 April 150, the latest in February or March 154; his latest known predecessor, M. Petronius Honoratus, is last attested on 11 November 148, and our document cannot be later than 29 August 154, the date of the first attestation of his successor, M. Sempronius Liberalis. A (left fragment): 24 × 24 cm; B: 13.3 × 22.5 cm; height of the letters: 2 to 2.3 cm. B preserves a section of the upper edge, which is straight, and of the side edge, which is irregular. 6. I visited this praesidium and the quarry sector in 1991; on the other hand, we had to pass up without seeing them the other sites mentioned by Tregenza, including that where I.Pan 53 was found. 7. And not “Site 5,” as it is given in Tregenza 1950: 85 (cf. the correction, Tregenza 1951: 49, n. 1). 8. Tregenza 1951: 49. 9. Meredith 1956: 120 (map reproduced in I.Pan, pl. 49).

84

Rome in Egypt’s Eastern Desert

Green schist (described by Tregenza 1950: 86, as “green mudstone”). For working purposes, I have omitted almost all of the abbreviations proposed in the lacunae by earlier editors (Fig. 41).

4

[Imperator Caesar Titus Aelius Hadri]anus Antóninụ[s Augustus Pius s]up Munatió Felic[e praefecto Aegypti,] sup Ulpió Himeró p̣[rocuratore Augus]ti vac. iussó legati Arrius Iuli[anus grade coh(ortis)] ỊII Ituraeor(um) praesidio vac. Fýnicon[e (– – – – – – –)

1. Imperator written in full is uncommon in titulature, but not unparalleled. ]anus is the only element that suggests that the emperor’s name was in the nominative, and that he should therefore be the subject of a verb (unless it is a statue base, which is hardly imaginable here). There is, to be sure, another nominative in line 5. But the size of the lacuna between these two fragments may have been underestimated, in which case there would be room for a verb between Pius and sup in line 2. Sometimes the verb of which the emperor’s name is the subject is simply understood: cf., e.g., ILS 92 and 93, as well as ILS 2280, 2282, 2286, and AE 1969-1970, 443, references which I owe to Fr. Kayser. When one takes account of the position of the letters in the different lines in relationship to one another, one is led to put forward two types of possible restorations for the entirety of this inscription, one long (a), with few abbreviations, the other shorter (b): a.

[Imperator Caesar Titus Aelius Hadri]anus Antoninu[s Augustus Pius  s]up Munatio Felic[e praefecto Aegypti,] sup Ulpio Himero p[rocuratore Augus]ti iusso legati Arrius Iuli[anus ]II Ituraeor(um)

b.

[Imp Caes T Aelius Hadri]anus Antoninu[s Augustus Pius s]up (or) Antoninu[s Aug Pius  s]up Munatio Felic[e praef Aeg] sup Ulpio Himero p[roc Augus]ti iusso legati Arrius Iuli[anus 7 coh] III Ituraeor

2. Antónin[s. ANTONINVs, Merlin , Antonin[us Bernand. On the original, the bottom of u is visible. 4. Tregenza, Merlin, and Bernand restore p[raef. Mon]ti. E. L. Wheeler (2000: 290) has proposed p[raep(osito) Mon]ti. I dot the p of p[rocurator: only the hasta is visible, with its bottom turned to the left; the lapicide is not consistent in the form of his p: that in Ulpio has an oblique foot, the hasta of that in sup (l. 3) is simply straight. The t of ]ti was already dotted by Bernand: actually, its horizontal stroke (very straight in this hand) is barely visible on Tregenza’s photo, but clear enough

Ulpius Himeros, imperial procurator (I.Pan 53)

85

on the original. Augusti full written out is not usual: normally one has Aug( ); cf. for an example, however, procuratori Augusti in AE 1988, 739. 5. legati. The a is eroded but readily recognizable on the stone. Arrius Iuli[anus. In Egypt, I have found this combination only in O.Dios inv. 152 (second cent.), a list of twelve names, the purpose of which is unknown and which combines Greek, GrecoEgyptian, and Latin names. Ἄρρις Ἰουλιανός is the only one on the list to have the duo nomina. Iuli[anus rank coh(ortis)] ỊII Ituraeor(um). ⸍ỊII ituraeor on the stone. So far as the rank is concerned, previous editions propose, for reasons of space, centurio represented by the symbol 7; this solution seems indeed unavoidable in the “short” version. As to the unit, coh(ortis) ỊII is the restoration adopted by previous editors; it supposes that the horizontal stroke at the top of the first vertical stroke of the number III is parasitic. Jones justifies his preference by claiming that the usual abbreviation of cohors is coh( ), and that to his knowledge there was no example of the abbreviation cohort( ) (apud Tregenza 1950: 89). But such do in fact exist, for example, ILS 5549 and 9158. We cannot, therefore, exclude the possibility of restoring cohor]t(is) II. Historically, both cohorts are possible. If we are dealing with the third cohort, the parasitic stroke could be explained by the fact that the lapicide had begun to write tertia out in full before changing his mind. The apex above the group III (or tII) is probably the abbreviation stroke of coh( ) or of cohort( ), rather than the numeral stroke, which should have been horizontal; there is a similar abbreviation stroke after Ituraeor( ) (since it is carved just after r, it cannot be an accent on o). 6. Fýnicon[e. Tregenza and Merlin read Eynicon[, and Bernand Eunicon[. Under this word, what remains of the inscribed space does not show any trace of carving; the maximum height of this vacat is 2.7 cm, which is more than the average interline spaces (respectively 2.3, 2.4, 2.4, and 2.5 cm), which, to be sure, tend to become wider: line 6 could thus be the last line, but this is uncertain.

Commentary Ulpius Himeros, procurator Augusti In line 4, Jones apud Tregenza 1950: 88, hesitatingly proposed p[raef(ectus) Mon]ti, a restoration that Bernand rightly thought doubtful, but which all the same earned Ulpius Himeros a place in a subsequent list of the prefects of Berenike:10 instead of the dative monti we expect the genitive montis, which is commonly used in the title of the prefects of Berenike. Moreover Ulpius Himeros, with his imperial gentilicium and his Greek cognomen, is out of place in the prosopography of the prefects of Berenike.11 The restoration p[rocurator instead of p[raefectus is confirmed by the mention of Ulpius Himeros with the title of procurator in two Greek ostraca of Mons Claudianus: O.Claud. IV 886 is a letter addressed to Athenodoros, described as ταβελλα]ρείῳ Εἱμέρου ἐπιτρόπ[ου (understand ταβουλαρίῳ),12 “accountantagent of the procurator Himeros”; O.Claud. IV 861 is the draft of a collective letter of the quarry workers to 10. Devijver 1975: 464, n. 79. An updated list of the prefects of Berenike is given in this volume, Chapter 3, pp. 105–9. 11. In the Latin onomastic system, a Greek cognomen betrays servile or non-citizen origins on the part of an individual or of his ancestors (cf. the classic article of Kajanto 1968). Only three prefects of Berenike have a non-Latin cognomen, but they belong to the end of the second-third century (nos. 21, 24, 27 of the prosopography, cf. preceding note). 12. In O.Claud., the occupational title ταβελλάριος is used either to refer to genuine messengers or in place of ταβουλάριος, which is itself attested only once in the corpus of O.Claud. (inv. 7362); ILS 4395 offers an epigraphic example of this confusion. Athenodoros is called tabellarios in O.Claud. inv. 7396 and βοηθός in O.Claud. inv. 7272 (with no mention of his superior).

86

Rome in Egypt’s Eastern Desert

Οὔλ(πιος) Εἵμερος himself (his title is lost in the lacuna but is not doubtful, as several other ostraca from Mons Claudianus are collective letters of the workforce addressed to procuratores of various periods). The date of the two ostraca mentioned agrees perfectly with the inscription of Wadi Samna: Athenodoros receives another letter dated to the 16th year of Antoninus, or 152/3 (O.Claud. inv. 7726). The ostraca do not give any further details of the title of Himeros. Most likely he was procurator metallorum (ἐπίτροπος τῶν μετάλλων).13 Like his predecessor M. Ulpius Chresimοs, attested as procurator metallorum in some inscriptions of Porphyrites (I.Pan 21, 117–19) and of Mons Claudianus (I.Pan 42), he administered several areas of quarrying, in this case Mons Claudianus and Ophiates; it is certain that Porphyrites and the small metallon of Tiberiane (Barud)14 also belonged to his area of responsibility: at this period, the imperial labor force (the familia) employed in this sector belonged to the numerus of Porphyrites, which was itself subdivided into arithmoi of Claudianus and of Tiberiane.15 The case of Ulpius Himeros is thus to be added to the controverted dossier of the territorial competence of the mining procurators;16 it should be remembered as well that the competence of a much earlier predecessor of Himeros, P. Iuventius Agathopous (supra), extended to all of the metalla of Egypt. It remains to consider the question of the legal status of our procurator: was he a freedman or equestrian procurator? We now know that the title procurator Augusti, although it is much commoner for equestrian procurators (including the prefects of Berenike) than for freedman procurators, was not deliberately reserved for equestrians; the explanation was given by F. Millar, who argued that the reason for the scarcity of freedmen procurators called procurator Augusti is that it was wished to avoid the useless repetition of Augusti in a title that already included the expression Augusti libertus.17 If Ulpius Himeros was a freedman of Trajan, it was all the easier to call him procurator Augusti in that his status as Augusti libertus was not mentioned, an omission paralleled elsewhere.18 The difficulty of making him a freedman of Trajan would instead derive from the date at which we see him in office.19 Normally, imperial freedmen obtained a procuratorship around 40–45 years of age, under the reign of the successor of their patron:20 even if we admit that he was manumitted at the minimum legal age (30) in 117, in the last year of Trajan’s life, Himeros would have been a man of 65 in 152! Even so, that remains possible: he could have been manumitted before the age of 30, although that is less common among the slaves of emperors (especially administrative personnel) than among slaves of private individuals.21 Moreover, we know a certain number of imperial freedmen who lived 15 to 40 years after their patron.22 The difficulty could also be overcome by supposing that Ulpius Himeros was the son of a freedman of Trajan.23 13. One can think also of ἐπίτροπος Ὄρους, but this title is securely attested only in one inscription of the third century, I. Portes 86 (219, cf. Chapter 6). 14. His subordinate the tabularius Athenodoros was responsible for at least Claudianus and Tiberiane. 15. On this classification of the imperial workforce, see O.Claud. III, pp. 36–40. What is not known is if the quarrymen of Samna also belonged to the numerus of Porphyrites, or to another. 16. Recently reviewed by Andreau 1989: 97–99. 17. JRS 53 (1963) 196; cf. recently Bruun 1990: 279, n. 38. 18. Cf. the cases cited by Chantraine 1975: 614. 19. This objection was brought to my attention by P. Le Roux. 20. See, for example, for the procuratores aurariarum of Dacia, Nœske 1977: 296–97. 21. Weaver 1972: 98–100. 24 % of imperial freedmen whose age at death is known were manumitted before the age of 30. 22. Weaver 1972: 30–31. According to him (p. 30), the last freedman of Trajan attested is M. Ulpius Capito, named in CIL VI 10234 (= ILS 7213), the foundation of a funerary college dated to 153; but this example is not entirely germane, as M. Ulpius Capito was dead at the moment of this foundation, which was established by his widow, and we do not know how much time had elapsed since his death. 23. This solution was suggested to me by G. Di Vita-Évrard. I do not know an explicit case of the son of a freedman occupying an administrative post in the imperial familia.

Ulpius Himeros, imperial procurator (I.Pan 53)

87

Other arguments support our hypothesis. Even if the notion of “servile names” is often criticized, it cannot be denied that Ἵμερος/Himeros was particularly common among slaves and freedmen at Rome.24 In addition, rather few Roman equestrians had Greek names (the procuratores in Aegypto with Greek cognomina enumerated in the Supplément to the Carrières of Pflaum are in fact, as Chr. Bruun rightly observed, manumitted procurators).25 It also seems that Ulpius Himeros was active at a period when the procuratorship of the metalla was still a function entrusted to freedmen: in the Danubian provinces, the procurators of the mines were freedmen until the end of the second or the beginning of the third century;26 after that, they were recruited from the equestrian order.27 Finally, the fact that another letter addressed to Athenodoros refers to Himeros without giving his name, simply referring to him as ὁ κράτιστος ἐπίτροπος (O.Claud. IV 878) in no way signifies that he was an equestrian: at this point, we are around year 16 of Antoninus, and it is only under Marcus Aurelius that vir egregius, of which κράτιστος is a translation, becomes the honorific title of equestrian procurators. Before that date, κράτιστος was simply a term of respect, used also for the prefect of Egypt, the epistrategos,28 the strategos of a nome,29 or for imperial freedmen such as the procurator usiacus.30

The praesidium of Phoinikon In line 6, the name of the praesidium was previously read Eynicon. But in the other e’s of the inscription, the upper bar is a particularly short horizontal stroke, while here we are dealing with a rather long oblique stroke, similar to that of the f of Felic[e. Tregenza may have been deceived by the slight return to the right of the lower tip of the hasta; certainly, the f of Felic[e curves instead towards the left, but so does the e; moreover, some f’s in Latin cursive have the bottom of the vertical curved toward the right (there is a fine example in O.Max. inv. 254, lines 3 and 12, BIFAO 94 [1994] 41); this type of f was to appear frequently in Christian texts.31 The reading Fynicon[ (note the y; the word reproduces a vulgar spelling in Greek, φυνικ- in place of φοινικ-) has moreover the advantage of offering a meaning (the personal name Εὔνικος could no doubt have served as a toponym, but not in the plural); numerous Egyptian toponyms are based on palm groves. We naturally expect that the praesidium mentioned in the inscription will have been its place of origin. In that case this would have been the ruined buildings of the “Site 6” that Tregenza describes in the following terms: “Its circumference is a large rectangular embankment of dug-up wadi-gravel, now partly washed away; and the large number of crushing-stones there (many of them built into the walls of the rooms) show that it was an important gold-washing site in antiquity” (1950: 85–86). As he remarked later (1951: 52), this praesidium must be later than the mining activity (which he was not able to date) attested in this place, because the millstones were reused in the masonry (1951: 52); the importance of this site is confirmed by the discovery that he made there in 1951 of three further fragments of Latin inscriptions, cut in the same type of stone as I.Pan 53, a green clayey rock, but in a more careful style of lettering, to judge by the facsimiles (Tregenza 1951: 47; reproduced in I.Pan, pl. 53); the “fragment 2” 24. Solin 1982: III 1240–41: at Rome, of 23 individuals with this name, 9 are of uncertain status, 14 slaves or freedmen. 25. Pflaum 1982: 137; Bruun 1990: 276–77. 26. At Ampelum in Dacia, the last freedman procurator was in office, it appears, under Marcus Aurelius (CIL III 1622 = ILS 1532, reproduced and discussed by Nœske 1977: 348). 27. Dušanič 1977: 92. 28. Bruun 1990: 272–73. 29. P.Oxy. XLVI 3313.4n. 30. Bruun 1990: 278. 31. Cagnat 1898: 15.

88

Rome in Egypt’s Eastern Desert

gives the end of the titulature of Antoninus (= I.Pan 54). It thus seems that there were praesidia not far from one another in Wadi Samna, one, which is still readily visible, controlling the diorite quarries that were in use since AD 11, the other, now levelled, built on an old mining site and founded, restored, or reoccupied, with some upgrade, under Antoninus; this second praesidium, farther from the quarries, was nonetheless under the responsibility of the procurator metallorum, probably because the track that it controlled was used for the transport to the Nile of the products extracted in the quarries of Ophiates and neighboring mines. The distance between the two praesidia is “less than 10 miles,”32 or about 15 km; this is in fact the average distance between the praesidia on the caravan route from Koptos to Myos Hormos;33 moreover, it is impossible to be sure that all of the praesidia along this track were in use at the same time, and the same is true for those of Wadi Samna. The name of the destroyed praesidium should have been Phoinikon (Φοινικών, “Palm grove”), which is also the name of another praesidium, identified today with the small oasis of Al-Laqita, located 34 km from Koptos, which marks the point at which the tracks to Berenike and Myos Hormos diverge. This station was called, according to the sources, Pœniconon (Itin. Ant. 172.1), Fœnicionis (Not. Dign. Or. 31.49), Phenice (Tab. Peut.), or Phinice (Anon. Rav. 2.7), but in the ostraca of the principate found on the road of Myos Hormos, it is always Φοινικών (genitive Φοινικῶνος). It is not impossible that two stations of the Eastern Desert, belonging to different systems, bore the same name: cf. the place-name Xeron Pelagos, which refers to a quarry near Mons Claudianus and also to a praesidium on the road to Berenike (see Chapter 1, p. 50). But it is also possible that the inscription mentioned a series of works carried out in several praesidia, including at Phoinikon = Al-Laqita (praesidio Fynicone might then be the subject of an ablative absolute).

Fragment C The three fragments were found at some distance from one another. Tregenza did not publish a photo of fragment C, but only a small facsimile (Tregenza 1950: 85); he observed that the characters were of smaller size, on average about one centimeter high, that the grooves did not present the interlaced strokes as in the other fragments (Tregenza 1950: 87), and that the stone is thicker and squared off at the top (it seems that it is the upper right corner of an inscription). From all of this he concluded that this was a fragment from the base of the inscription, with the signature of the lapicide (Tregenza 1950: 89). But in 1951 he seems not to have believed any longer that C belonged to the same inscription as A and B (he speaks only of “the two fragments found last year.”34) In my opinion, this fragment actually has nothing to do with the two others: along with the differences noted by Tregenza, we can add that the surface of C is perfectly smoothed, which is not the case with A and B; and, above all, the hand is not the same, as one can see by comparing the respective forms of m. Width 12.5 × height 10.5 cm. Average height of the letters: 1.5 cm. Line written 0.8 cm from the upper edge; below, vacat of 8 cm in height. – – – – – – cons]ummauiṭ 1. The editors of AE 1954, 85 propose ala Thrac]um Maure[tana, but reading a r in place of an i is impossible. 32. Tregenza 1951: 50. 33. Murray 1925: 146. 34. Tregenza 1951: 49.

Ulpius Himeros, imperial procurator (I.Pan 53)

89

Appendix The epigraphical corpus of Samna has recently been augmented by two small pieces found in 1994 by a geologist in the Paneion of the quarry zone (Fig. 42).35 Unlike the Augustan inscriptions of the Paneion (I.Pan. 51 and 52), they are not in diorite but in the same fine-grained schist as the fragments of “Site 6”; although the carving has been carried out by incision with a punch and not with a hammered chisel, the work is careful and elegant: the letters, some of which are attractively serifed, are of regular height (between 1.1 and 1.3 cm) and rubricated. The interline space is 1.55 cm. The two fragments belong to the right side of the inscribed space: in fragment B this is marked by a distinctly incised vertical line, which left space at its right for a margin; with fragment A, it is the absence of any trace of τ after the π of Αἰ]γυπ that seems to me to indicate that a line-break occurred here.

Figure 42. Fragments found in 1994 at Samna © J.-Fr. Gout (IFAO)

Fragment A 7 × 6.3 cm; maximum width of the inscribed area, 5 cm; thickness c. 0.5 cm. (– – –) ] ε̣π̣ι̣ vac.([) Αἰ]γυπ ([) ]ς̣ ([) – – – 1–2.

Perhaps ] ἐπὶ | [τοῦ δεῖνος ἐπάρχου Αἰ]γύπ|[του

35. Vidi. They have been turned over to the Antiquities Service and registered with the material from Al-Zarqaʾ (Maximianon) under the inventory number 1457.

90

Rome in Egypt’s Eastern Desert

Fragment B 10.5 × 9 cm; maximum thickness 0.8 cm. – – – ] vac. ]α̣το̣ ]α̅ – – – 1.

αὐτοκρ]άτο|[ρος?

2.

Above α is a horizontal stroke that ends in a serif; this was probably part of a date (the day of the month). It is noteworthy that the two dated inscriptions found in the zone of the quarries and of the Paneion are dated respectively to Pauni 1 of year 40 of Augustus and to Pauni 1 of an unknown regnal year of Tiberius. The possibility has been raised that there was a festival of Pan on that day (Tregenza 1951: 46).36 However hazardous the suggestion may be, we must mention the possibility of restoring Παυνι] α̅ here.

36. There is no observance on Pauni 1 in the table of festivals drawn up by Perpillou-Thomas 1993.

3 Sulpicius Serenus, procurator Augusti, and the titulature of the prefects of Berenike Adam Bülow-Jacobsen and Hélène Cuvigny with an Appendix by Maël Crépy In 2006 the excavations at Dios1 on the road from Coptos to Berenike brought to light an ostracon which encourages a closer study of the place of the prefecture of Berenike in the equestrian cursus. O.Dios inv. 90 US 3525 Fig. 43

14 × 12 cm

reign of Hadrian sherd of alluvial clay

The ostracon was found in a deep layer of the rubbish deposit outside the praesidium. Thanks to a Latin dedication2 we know that the fort and the well were constructed in the 19th regnal year of Trajan (115/6). The ostracon is a report, in the form of a hypomnema, on a mission along with a demand for human and material means. The writer, an epimeletes of wells, has been ordered by a procurator Augusti, who receives the report, to examine a well of which the yield was failing. Letters on ostraca from the Eastern Desert are normally sent from a neighboring station to addressees who are at the place where the letter is found. Are we to believe that this ostracon has been sent from 1. Dios, today called Abu Qurayya, is better known by the name Iovis by which it is called in the Antonine Itinerary. For a short report on the first campaign, see BIFAO 106 (2006) 409–12. We thank the reading committee of Chiron for remarks and corrections. We also thank colleagues who have permitted us to improve our understanding of this ostracon and its implications by answering our questions and by inviting us to present it in seminars: Ségolène Demougin, Jean Gascou, Rudolf Haensch, and especially Patrick Le Roux, who kindly read through a draft of this paper. 2. Chapter 31 (AE 2010, 1751).

91

92

Rome in Egypt’s Eastern Desert

another site and was delivered to the procurator at a time when he was on a tour of inspection and found himself at Dios? This does not appear the most likely reconstruction, all the more so since, when addressing someone of elevated rank, one would write on papyrus, even in the desert. Consequently, our letter must be a draft or a copy written on site and kept as a record, and the well in question would thus be that of Dios. This is not a unique case, and the Roman installations in the desert have yielded other letters on ostraca addressed to important people, but written and found on site. One might even venture that any letter found in this region, written to an official person who was not a resident in the desert, is not an original document. The most relevant for our purpose are the following: – O.Krok. 14 (109), a letter from the curator of Krokodilo to the prefect of the desert in order to obtain a replacement of a metal piece in a grain-mill at the praesidium; – Two letters, copies of which are written on the same sherd, addressed by the vice-curator of Mons Claudianus, one to the prefect, Vibius Alexandros, the other to the procurator Tertullus;3 – O.Did. 38 (post 185), a letter from the curator of Didymoi to the prefect of the desert and of the ala Herculiana about the progress of work on the well; – O.Did. 40 (= P.Bingen 108, c. 219, Chapter 6), a report submitted to a procurator Augusti by the curator of Didymoi, who has changed three pots on a saqiya. The functions of the author and the addressee of the present ostracon are discussed below (comm. ad 1 and 3). From line 10 on, there are traces of a line in the left margin downwards.

Figure 43. O.Dios inv. 90. © A. Bülow-Jacobsen 3. Chapter 5 (SB XXVIII 16941).

Sulpicius Serenus, procurator Augusti

4

8

12

93

[Σερουίῳ] Σουλπικίῳ Σερήνῳ ἐπιτρόπῳ Σεβαστοῦ. vacat [παρὰ --]ᾶτος ἐπιμελητοῦ ὑδρευμάτων. [γινώσκε]ιν σε θέλω, κύριε, κατὰ τὴν συν[ταγήν σου] κατεληλυθέναι ἰς τὸ ὕδρευμα [καὶ ὀρωρ]υχέναι ὀλίγον τι καὶ ὑπηντη[κέναι] μοι ⟦καὶ⟧ ὄρος σκληρὸν ὅστις ἀπο[τέτ]ρεφι καὶ τὰ σιδήρια. γείνωσκε [οὖ]ν̣, κύριε, μόγις τὸ τρίτον διδὸ{ι}ν α̣ὐ̣τ̣ὸ οὗ ἐδίδι ὕδατος· διὸ παρακαλῶ σε, ἠάν σου τῇ Τύχῃ δοκῇ, ἐκπέμψαι μοι χαλκέα καὶ στομώματος {μνᾶς} ἐλάσματος μνᾶς δεκαπέντε καὶ ἀνθρώπους δ̅ τοὺς δυναμένους ἐκ τῆς φαμελίας, ὧν τὰ ὀνόματα· Ἰοῦστος, Κρητικός, Αλαφθ̣αν, Γλαύκον 4 illegible lines

1 σεβαστου 6 l. ὅ τι 6–7 l. ἀποτέτριφε 7 l. γίνωσκε 8–9 διδο{ι̣}ν-|[]ο̣υ ed. pr. σου 14 l. φαμιλίας 15 Αλαφθ̣αν: α̣φ̣α̣ν ed. pr. ‖ l. Γλαύκων

10 ἐάν

“To [Servius] Sulpicius Serenus, procurator Augusti, [from …]as, overseer in charge of the wells. I want you to know, Sir, that [on your order, I] descended into the well. I dug a little and came upon hard rock that even blunted the points. Be informed that it gives less than a third of the water it used to give. For this reason, I ask you, if it pleases your Tyche, to send me a blacksmith and four minae of hammered-out steel and four strong men from the familia, namely Iustus, Kretikos, Alaphtha, Glaukon …” 1.

The praenomen is supplemented from I.Memnon 20.1 (see commentary 3 below).

3–4.

κατὰ τὴν συν[ταγήν σου] κατεληλυθέναι ⟨με⟩ ἰς τὸ ὕδρευμα? One may hesitate between “I came down (sc. from Berenike) to the well” or “I descended into the well”. The use of this verb permits both interpretations. We have preferred the second since it leads logically to [καὶ ὀρωρ]υχέναι. Even the use of εἰς can be justified because ὕδρευμα sometimes has the narrow sense of “water-table” (cf. the expression hydreuma quaeri … hoc cum esset inventum in the dedications of the three praesidia on the road to Berenike, AE 2001, 2039, 2047, and 2051). At the time when this ostracon was written and unlike in Pliny’s time, hydreuma no longer had a geographical meaning. Since Vespasian the stations along the road in the desert of Berenike were no longer just wells, but forts. If the epimeletes had wanted to indicate where he was at work, he would have employed the place-name (Dios), or would at least have written praesidium, not hydreuma.

6.

ὄρος σκληρόν. The sense “bedrock” is not recorded in the dictionaries, but it will be natural to those who have worked on archaeological sites in Egypt, where the workers use the word gebel about the virgin soil under the archaeological layers (in fact, the correct word is gaballa, as we are informed by Nessim Henein). The terms ὄρος, in Egyptian Greek, and gebel cover almost

94

Rome in Egypt’s Eastern Desert the same meanings, since both may mean a mountain as well as the desert, not necessarily mountainous, that encloses the Nile valley. This use of ὄρος in not a hapax, but is also found in I.ThSy. 237.6–7, to which Jean Gascou has drawn our attention: τὸ μέρος τοῦτο τοῦ τείχο[υ]ς τὸ ἐπὶ τὸ ὄρος where the editor translates “cette partie du rampart touchant la colline.” According to Gascou it is rather that “this part of the wall” was built directly on the bedrock, which posed a technical problem.

8.

διδό{υ}ν, which would be a more acceptable misspelling, is not allowed by paleography.

9.

ἐδίδι. On this form of imperfect of δίδωμι, see Mandilaras, Verb, § 88.

11–12. στομώματος μνᾶς ἐλάσματος μνᾶς δεκαπέντε. στόμωμα is well attested in the papyri and often reckoned in minae (see, e.g., P.Cair.Zen. IV 59782a, passim), but there is no parallel to stomoma being associated as here with ἔλασμα. This word, derived from ἐλαύνειν in its metaphorical sense of “beating metal” (LSJ s.v. III), designates sheet-metal, metal hammered to make a plate, and is mostly used about gold, silver, or lead, see Blümner 1887: 230. There are two possibilities: (1) The writer forgot the number of minae of stomoma and also the conjunction καί and then also the name of the metal to define ἐλάσματος (of which this is the first occurrence in the papyri). So many omissions can only be explained if the writer was copying from a model and forgot a line. Or (2), the writer forgot to delete the first μνᾶς after στομώματος, and στόμωμα defines the metal of which the ἔλασμα is made. This is the hypothesis that we have preferred. The stomoma, steel to strengthen the cutting edge of a tool, was perhaps available in several forms. Having begun to write (or dictate) as it is often found “X minae of stomoma,” the epimeletes thought better of it, and specified that he wanted stomoma in plates. For further explanations about these techniques, see commentary 2, below. 12–14. καὶ ἀνθρώπους δ̅ τοὺς δυναμένους ἐκ τῆς φαμελίας. This formula is rather like one that we find in O.Claud. II 375.7–8, which is also a demand for manpower from the imperial workers: καλῶς ποιήσις πέμψον ἡμεῖν τοὺς δύο φαμιλιαρίους τοὺς δυναμένους ἀντλῖ⟨ν⟩, “please send us the two men from the familia who are capable of drawing water.” But in the ostracon from Dios there is no definite article in front of ἀνθρώπους δ̅, which makes the one in front of δυναμένους unexpected, and there is no infinitive dependent on δυναμένους. Unless we suppose that this was simply forgotten, δύνασθαι (sc. τῷ σώματι) is here used absolutely “to be strong” (ThGrL, s.v., 1703D). The well-diggers must attack the hard rock at a depth of at least 20 meters, and this hard work must further be carried out partly in water – we can imagine that the water was constantly rising and bailing was necessary all the time. So the epimeletes wants to make sure that he is not sent claustrophobic weaklings, and he demands men whom he knows and appreciates. We have tried to render the definite article by “namely.” Literally the epimeletes demands four men, and then specifies exactly whom he means by their physical qualities, status, and names. Note that there is no question of employing soldiers for the job. 15.

Αλαφθ̣αν. A Semitic personal name, common in Palmyra. It transcribes ḥlptʾ, “successor, deputy” (Yon 2018: 43, 54 f.). Until now, known Greek transcriptions were spelled Αλαφαθα(ς).

Sulpicius Serenus, procurator Augusti

95

Commentary 1. The epimeletes of wells The name of this function is new. The title epimeletes is itself very rare in the Eastern Desert, and the only occurrences are: – O.Claud. inv. 1538+2921.9 (c. 110):4 In this list of those entitled to receive water there is an ἐπιμελητὴς Πανείου, who has been interpreted as a caretaker of the sanctuary of Pan. – O.Claud. inv. 3069.5: A fragment of a list, similar to the above, but in a different hand. Among those entitled to water there is, just after the τηρητὴς στάβλων, an ἐπιμελητὴς κτηνῶν. This function is not otherwise attested on Mons Claudianus, and we are tempted to understand it as the equivalent of a κτηνοιατρός (veterinarian), inasmuch as, in other lists, the τηρητὴς σταβλῶν is immediately followed by the κτηνοιατρός. Note, however, that in these lists, the τηρητὴς σταβλῶν and the κτηνοιατρός receive the same water-ration, while in inv. 3069, the first receives ½ keramion while the epimeletes is entitled to ⅓ only. Is it a different man than the usual veterinarian, with a lower status? – O.Krok. II 235 (reign of Trajan or early Hadrian) some sort of account of an inn-keeper. After a section called ἀλλαντίων ἀνήλωμα (expense on sausages) follows one called ἀνήλωμα κεραμίων listing people, each with a number of keramia. Among these people, to whom amphoras (sc. of wine) have been sold (?), there are “the horsemen,” Philokles,5 and the epimeletai (οἱ ἐπιμεληταί) without any further indication of their function. The horsemen have received eight amphoras, the epimeletai two. Our epimeletes is clearly responsible for the maintenance and repair of the wells in the praesidia, but he has nothing to do with the upper-class people who, also with this title, were made responsible for surveying the construction and maintenance of buildings. The concepts of epimeleia and cura are found in all kinds of contexts and juridical relations to describe ad hoc functions. Here, on the other hand, the epimeletes is an employee of the maintenance services of the prefecture of Berenike. We do not know his status. Was he a civilian or military technician at the service of the procurator, or a member of the familia? In any case, he was not a desk-worker, but a practical man, and he seems to have descended into the well himself in order to make a diagnosis of the necessary work and an estimate of the manpower and materials needed.

2. Stomoma and elasma Since iron is softer than most stones, the quarrymen and stone-masons used steel-tipped tools, i.e., a small piece of steel that had been welded to the business end of the tool.6 The word στόμωμα designates this piece of steel7 that was welded to the σιδήρια as well as the special steel of which it is made (LSJ, s.v. II “hardened iron, steel … hard edge or point welded into a blade or shaft, or steel for this purpose”). 4. See Chapter 11 (on epimeletes, cf. pp. 206 f.). 5. About Philokles, greengrocer and procurer, who did business with the garrisons in the desert of Berenike, see Cuvigny (ed.) 2006: 376–83. His correspondence and texts concerning him are now published in O.Did. 376–399 and O.Krok. II 152–235. 6. Agricultural tools might also be steel-tipped, e.g., δρέπανα χορτοκόπα, χορτοτόμα, καλαμοτόμα (P.Cair.Zen. IV 59782a viii.125–131). 7. At Mons Claudianus the stomomata that were welded onto the points of the quarryworkers had the form of a pyramid. See photo in Peacock and Maxfield 1997: 192, fig. 6.9. More recently see O.Claud. IV, pp. 257–59 and Bülow-Jacobsen 2016.

96

Rome in Egypt’s Eastern Desert

The Mons Claudianus ostraca concerning the quarry-work abundantly illustrate this practice, and an appendix in O.Claud. IV deals with the matter in some detail. An abstract of this may suffice here. When working on granite, the stonemason can use his steel-pointed tool (normally called σιδήριον in Mons Claudianus) for about an hour. When the tip of the tool becomes blunted, he passes it to the blacksmith (χαλκεύς) in the little forge that is almost always found next to the quarry, and here the tip is re-forged at red-hot temperature and regains its pyramidal shape. When this operation has been repeated a certain number of times, the steel at the tip (στόμωμα) is worn away and has to be replaced. This operation, called στόμωσις, consists of welding fresh stomoma onto the siderion. This is more complicated than the simple reforging, for the object has to be heated to white heat in order to allow fusion (forge-welding). At Mons Claudianus the stomosis was carried out by a specialized blacksmith, called στομωτής, at a special forge, known as the στομωτήριον, where fusion-temperatures could be obtained. In the present case, the stomosis would be done by the blacksmith, who would also take care of the intermediate re-forging. In O.Claud. IV 826 it is also, exceptionally, the blacksmith who takes care of the stomosis. By its weight, the quantity of stomoma demanded by the epimeletes seems rather large, if we compare it with P.Cair.Zen. IV 59782a, a list of iron tools with separate specification of the weight of the stomoma of each tool. In this document, a λατομίς, a stonemason’s pointed chisel which must be similar to the σιδήρια in our texts, has ¼ mina of stomoma. If we calculate on this basis, the 15 minae of stomoma that are demanded would be enough for 60 sideria. We do not know the frequency with which steel-tips would normally be replaced, and O.Claud. IV 833 allows only the conclusion that it was not necessary every day. Consequently, the epimeletes expects the work to last a certain time. The insistence of the epimeletes that the steel must be in plates (ἔλασμα) can perhaps be explained since, in this form, the steel would already be hammered and purified. The blacksmith would thus not have to purify the steel, which at Mons Claudianus is the job of the σφυροκόπος, on top of the reforging and welding of new tips. On the numerical ratio of smiths to stonemasons, see O.Claud. IV, introduction to the chapter “Lists of specialist-workers assigned to tasks.”

3. The addressee: Sulpicius Serenus The addressee of this hypomnema is not unknown. There is a high probability that he is the same as Sulpicius Cn. f. Quirina Serenus who, during the reign of Hadrian, set up an inscription of thanks to Jupiter for a victory over the nomads of the Eastern Desert, the Agriophagi (I.Pan 87). Serenus does not tell in which capacity he had participated in this action. The inscription is roughly inscribed on a coarse block of schistous sandstone from Wadi al-Hammamat, and was bought in Luxor.8 Sulpicius Serenus has left other epigraphical evidence in the region. On the colossus of Memnon there is his cursus, precisely dated to the sixth year of Hadrian’s reign.9 His cognomen is obliterated, and he has another praenomen than his father:

8. The inscription thus certainly comes from Upper Egypt, but not necessarily Luxor. It might just as well have been found in a praesidium in the Eastern Desert, like I.Kanaïs 59bis, which was bought in Luxor in 1901 from the antiquarian Abd elMagid (G. Poethke, per litt.), but which has been shown to originate from the praesidium Didymoi (Chapter 30 = I.Did. 7). Bought in Luxor and subsequently given to the Graeco-Roman Museum of Alexandria, I.Pan. 87 was first published by Botti in Bulletin de la Société archéologique d’Alexandrie 1902. It is quite possible that there were illicit excavations in the Eastern Desert during these years. 9. And not year 7 as read by all the epigraphists. Until now, everyone has been fooled by the undulating horizontal line of the (ἔτους) sign which, in contact with the following digamma (ϛ) creates the impression of a reversed zeta (ζ).

Sulpicius Serenus, procurator Augusti

97

I.Memnon 20 (Thebes, year 6 of Hadrian = 121/2):

4

8

Σέρουιος Σουλπίκιος [Σερῆνος] ἔπαρχος σπείρης [- - χιλίαρχος] λεγεῶνος κ̅β̅, [ἔπαρχος εἴλης Οὐο-] κουντίων, ἐ̣π̣ [ c. 12 ] νεωκόρος τοῦ με[γάλου] Σαράπιδος, τῶν [ἐν Μουσείῳ] σειτουμένων ἀτελῶ[ν, ἤκουσα] Μέμνονος ὥρας [ -] (ἔτους) ϛ̅ Ἁδριανοῦ ([)

2 σπείρης, χ̣ε̣ι̣[λίαρχος] Bernand

4 κουντίω[ν - - Bernand

9 (ἔτους) ζ alii

2.

σπείρης [- - χιλίαρχος]. We return to the restoration of Ritterling: it is unlikely that the name of the cohort was omitted while those of the other units are written, and such an omission would be more probable at the very beginning of the empire (P. Le Roux, per os). If, as seems probable, it was a cohort of the Egyptian army, Serenus would have made his whole career in this province and was probably born there (Pflaum 1960: 244 f.).

4.

ἐ̣π̣ [. The lower part of the epsilon is clearly seen on the photograph. It is a round letter, and the other possibilities are sigma or omicron. The following traces are not against a pi. The traces of the following letters are inconclusive.

The third militia equestris accomplished by Serenus is the prefecture of the ala Vocontiorum. This unit was stationed in Koptos at the head of the roads from Myos Hormos and Berenike until c. 180.10 There is agreement that it was as praefectus alae Vocontiorum that Serenus led a punitive expedition against the Agriophagi. Between his three militiae and the honors in Alexandria, he mentions another function of which only the bottoms of the first letters remain. The traces are not against the restoration ἐ̣π̣ί̣τ[̣ ροπος as suggested by the ostracon. In fact, it was a normal cursus at this time to obtain a procuratorship after three militiae. But ἔ̣π̣α̣ρ̣[χος would suit the remaining traces just as well. Whether the title that follows his prefecture of the ala Vocontiorum in I.Memnon 20 began with ἐπίτροπος or ἔπαρχος, the ostracon tells us that he was at some point procurator. But which procuratorship is involved?11 The ostracon does not tell us. Two other ostraca from the Desert of Berenike have also preserved letters in the form of hypomnema addressed to an ἐπίτροπος Σεβαστοῦ: (1) O.Did. 27 (discarded around 140–150)12 is a fragment of a copy of received letters. To a circular addressed to curatores praesidiorum in order to put them on guard against the “Barbarians” (i.e., no10. The last attestation of the ala Vocontiorum in Egypt is from 179. In 183 the unit is in Palmyra and has been replaced in Koptos by the ala Herculiana, which is attested for the first time in Koptos in 185. 11. Certainly not the procuratorship ad dioicesin Alexandreae which Pflaum 1960, no. 104bis, followed by Devijver 1975, no. 107, attributes to Sulpicius Serenus. This error goes back to Premerstein, Klio 3 (1903) 32 and is based on a Latin papyrus which was later corrected (cf. Ch.L.A. III 203 and the introduction there). 12. But the ostracon may be earlier than this date, which is based on an analysis of the stratigraphy of Didymoi.

98

Rome in Egypt’s Eastern Desert

mads) has been added a copy of an anonymous report addressed to an ἐπίτροπος Σεβαστοῦ whose name is lost. This report, of which only two truncated lines remain, relates a specific incident, no doubt concerning the Barbarians who are mentioned in the circular (it began “the 16th of the current month”). The author of the circular, whose name is in a lacuna, was perhaps the procurator Augusti himself. (2) O.Did. 40 (= P.Bingen 108, cf. Chapter 6) is later and has already been mentioned. It is a report of a mission addressed by the curator of Didymoi to an ἐπίτροπος Σεβαστοῦ called Valerius Apolinaris. This man appears in an inscription from Koptos, I.Portes 86, dated 219, where his title is ἐπίτροπος Ὄρους, procurator of the Desert (sc. of Berenike). Because of this title it becomes possible that Sulpicius Serenus was also ἐπίτροπος Ὄρους, a procuratorship that would thus have been created some time before 219. The ostracon from Dios casts doubt on the hypothesis that was put forward in P.Bingen 108, namely that the prefecture of Berenike had been replaced by a procuratorship between 216 (last attestation of an ἔπαρχος Ὄρους) and 219 (first – and only – attestation of an ἐπίτροπος Ὄρους in an inscription where one would expect the usual ἔπαρχος Ὄρους). If Sulpicius Serenus is really ἐπίτροπος Ὄρους, this procuratorship already existed in the first half of the second century, even if the ἔπαρχος Ὄρους is more prominent in the epigraphical and papyrological documentation. There are two, or perhaps even three, possible explanations: (1) from the second century on, an equestrian procurator, concerning whom our documentation would be very sparse, participated in the administration of this region in addition to the prefect of the desert; (2) the prefecture of Berenike, normally considered a militia,13 actually was a procuratorship; (3) the epimeletes, by ignorance or flattery, used an incorrect title.14 Let us examine the first possibility. The ostraca from Krokodilo, which date from the reign of Trajan and the beginning of that of Hadrian, and are thus a little earlier than the interventions of Sulpicius Serenus in the Eastern Desert, do not give the impression that the Desert of Berenike was administrated at the same time by a praefectus and a procurator. The prefect (whose title is ἔπαρχος Ὄρους Βερενίκης, very often abbreviated for the same person to ἔπαρχος Ὄρους or just ἔπαρχος) is present everywhere in the correspondence with the curatores praesidiorum. In their private letters, the inhabitants of the praesidia, both civilians and military, often speak of the eparchos in matters of discipline, relief, or litigation, but they never mention any epitropos. There are, however, two documents that present ambiguous evidence: (1) In the postal journal O.Krok. I 26.11–12, which dates to the prefecture of Artorius Priscillus, 109, we read: [ c. 10 [ c. 10

Ἰο]ύλις Κέλερ μετὰ ἐπιτρόπου ἐ]πάρχου καὶ ἐξῆλθε Ἰούλιος

The formulas of the postal journals would be in favor of supplementing, before Ἰο]ύλις Κέλερ, the preposition ἀπό followed by the name of a praesidium in the immediate neighbourhood of Kro13. Lesquier 1918: 428, approved by Pflaum 1960: 186, connects this to praefecturae gentium or civitatum. Equally Thomas 1982: 161. 14. This hypothesis, which has been advanced by P. Le Roux, is preferred by W. Eck, who suggests that the ostracon was thrown away because of this error. The fact that the ostracon is in all probability a draft or a private copy (see the introduction above) is, in our view, sufficient to explain its presence in the dung-heap. Even if the honorific epithets were sometimes treated with a certain nonchalance, we find it difficult to believe that this was also the case with titles of functions, all the more so since there is epigraphic confirmation by I.Portes 86, which mentions an ἐπίτροπος Ὄρους (Chapter 6).

Sulpicius Serenus, procurator Augusti

99

kodilo. The horsemen whose comings and goings are registered in these journals sometimes transport official letters or parcels, sometimes they accompany beasts of burden or people. Letters, parcels, persons, or animals are regularly preceded by the preposition μετά: μετὰ καμήλων (O.Krok. 30.48), μετὰ πορίας (O.Krok. 27.7), μετὰ [ὀ]ψαρίων (O.Krok. 1.29), μετὰ Δομιττίου δουπλικαρίου (O.Krok. 3.11), μετὰ διπλώματο(ς) Ἀρτωρίου Πρισκίλλου ἐπάρχου (O.Krok. 30.44). Here we can but quote the commentary on O.Krok. 26.11: “What kind of procurator? A procurator of the prefect? Or has Celer arrived from Phoinikon with a procurator and some letters from the prefect?” The absence of a name, or at least a definite article before ἐπιτρόπου, as if it could be just any epitropos, is disturbing, especially if the man in question is an imperial civil servant. (2) Among the authors of the circulars copied in O.Krok. 87, a liber litterarum allatarum, there are two candidates for identification with the prefect of the desert in office. A circular of 9 March 118 is sent only to the curatores of the Myos Hormos road by Arruntius Agrippinus, though only the end of his title Ὄρους is preserved. On 26 March 118 the same Arruntius Agrippinus sends another circular, again to the curatores of the Myos Hormos road only, but he leaves out his title in the prescript. Such an omission is normal in the epistolary style of equestrian officials.15 But between these two letters,16 a circular (the date is lost) has arrived at Krokodilo addressed to [prefects, centurions, decurions,] duplicarii, sesquiplicarii, and curators of praesidia in the desert, from [4–7]ειος Ταῦρος ἔπα̣[ρχος? - -]. It is highly likely that we meet the same person, with the same title of ἔπαρχος Ὄρους, in two undated ostraca. Both are copies of official correspondence, written in the same idiosyncratic hand as O.Krok. 87: – O.Krok. 88.9–11: ἐξ̣ [ἐγκε]|λεύσεως Κασσίου Τα̣[ύρου] | ἐπάρχου Ὄρου⟨ς⟩ – O.Krok. 91.4–5: Κασσείῳ  ἐπάρχῳ Ὄρους [παρὰ - - - ]| πρα⟨ι⟩σε̣ι̣[δίου] Κ[ορ]κοδι̣λὼ χα(ίρειν) [.17 O.Krok. 87 gives the impression that in Hadrian’s second year the prefect of the desert was not the only one to have authority over the curatores praesidiorum and whose title ended in Ὄρους. If Cassius Taurus was ἔπαρχος Ὄρους, could his contemporary Arruntius Agrippinus have been ἐπίτροπος Ὄρους? But we have already mentioned the possibility that Arruntius Agrippinus simply belonged to the officium of the ἔπαρχος Ὄρους,18 just like Minucius Marcellus who, without giving his title, writes to the curator of Krokodilo relaying an order from the prefect Cassius Taurinus.19 R. Haensch has put another interesting suggestion to us, namely that Marcellus and Agrippinus were both adiutores of the prefect of Berenike. In this case, the prescript of the circular of 9 March (O.Krok. 87.1–2) could easily be restored as follows: Ἀρούντει{ει}[ος Ἀγριππῖνος βοηθὸς ἐπάρχου]| Ὄρους κτλ. 15. Not indicating your rank was a way of stressing that you were a well-known person. 16. The letters collected in O.Krok. 87 were clearly copied in order of arrival. 17. It is worrying that around the same time the prefect of the desert was a certain Lucius Cassius Taurinus, who is attested in two texts of uncertain date, both from Krokodilo (O.Krok. 60 and 65), and who is also curam agens in the dedication of the fort Dios, dated 115/6 (AE 2010, 1751). We are close to thinking that the cognomen Taurus, instead of Taurinus, is an error of this scribe, since the form Ταῦρος is found in three ostraca written by the same hand. In this case Cassius Taurinus would have been prefect of the desert also after the death of Trajan. 18. O.Krok. I, p. 138. 19. O.Krok. 65.1–4: ἐξ ἐγκελεύσεως Κασίου Ταυρίνου{νου} ἐπάρχου vac. Μενίκις Μαρκέλλος Οὐ⟨α⟩λερίῳ κουράτωρι πρα⟨ι⟩σιδίου Κορκοδιλὼ χαίρειν.

100

Rome in Egypt’s Eastern Desert

Let us take a closer look at the hypothesis that the prefect of the Desert and an equestrian procurator had divided between them the administration of the Desert of Berenike. One would take care of the military government, the other of the imperial revenues, such as, for example, the mines and quarries of the Desert of Berenike20 or the farmed taxes, such as the customs duties from the Red Sea trade, or the transactions in the praesidia (it is known, for example, that the tax on prostitution was farmed to a conductor).21 But this hypothetical division of responsibilities is contradicted by the evidence, because the documents give us a view of the activities of the prefect. O.Dios inv. 90 is not different in style from other letters addressed to prefects of the desert of Berenike (O.Krok. 14, O.Did. 87, both already mentioned). On the other hand, O.Did. 27 is a copy of a report addressed to an ἐπίτροπος Σεβαστοῦ and concerns military matters. Finally, documents earlier than the reign of Trajan show that the exploitation of the metalla and fiscal questions were also within the competence of the prefect of Berenike (see below, the prosopographical table of the prefects of Berenike, nos. 1 and 6). Of course, towards the end of the second century at Mons Claudianus, we find both prefect and procurators, for example, in the letters22 addressed the same day by a vice-curator of Claudianus to the prefect23 Vibius Alexandros and to the procurator Tertullus. Contrary to the ostraca from the desert of Berenike, those from Mons Claudianus from the same period quite often mention an epitropos, in fact more often than an eparchos. But this procurator is an imperial freedman, as suggested by the formula of address in hypomnemata sent to him: ἐπιτρόπῳ τοῦ κυρίου Καίσαρος.24 This procurator is probably procurator metallorum. We do not believe that the ἐπίτροποι Σεβαστοῦ of the ostraca from the desert of Berenike, or the ἐπίτροπος Ὄρους of I.Portes 86, are procurators of the metalla in the desert of Berenike, as Lesquier (1918: 240) supposed concerning the procurator of I.Portes 86. In the first place, the procuratores metallorum known from Egypt under the early empire are freedmen, and second, the metalla do not seem to have been the principal activity in the desert of Berenike during the Roman period, and they are not the reason why the praesidia were built. Let us then look at the second hypothesis, that the prefecture of Berenike was a procuratorship. This was already the opinion of Domaszewski, but it was discarded by Lesquier.25 Although there is no document where Sulpicius Serenus has the title of prefect of the Desert or the like, he has the profile 20. The metalla of the desert of Berenike are not well known, because, unlike the quarry-sites in the north of the Εastern Desert (Mons Claudianus, Porphyrites, Umm Balad), they have not yielded any ostraca, apart from a small group published by Fr. Kayser, ZPE 98 [1993] 111–56). Besides, the ostraca from the praesidia on the road to Myos Hormos and Berenike are totally silent on the subject of exploitation of the mines. 21. Chapter 23. 22. Chapter 5. 23. The ostraca of Claudianus from this period mention several prefects, often as addressees of letters which do not, in their style, differ from those addressed to procuratores Caesaris. The extent and jurisdiction of their prefecture is never mentioned. There is only one prosopographical parallel to the prefects mentioned in the ostraca from the desert of Berenike, namely no. 23 in the prosopography below (Chapter 15, pp. 240 f.). 24. O.Claud. IV 853–55, addressed to the ἐπίτροπος τοῦ κυρίου Καίσαρος Probus c. 185. In Egypt the formula ἐπίτροπος τοῦ κυρίου Καίσαρος (var. τοῦ κυρίου ἡμῶν in P.Fam.Tebt. 15 [AD 114/5]) was reserved for the procuratores usiaci, who were imperial freedmen (Bruun 1990: 277–9). Note that among these people Ulpius Paian in P.Prag. II 132 (second cent.) is very likely to be the same as Paean in ILS 1569 (see now F. Beutler, ZPE 160 [2007] 232–4). The O.Claud. show that the formula was also applied to procurators of the metalla, which was probably the jurisdiction of Probus, who must have been an imperial freedman like his predecessor M. Ulpius Chresimos (I.Pan 21 and 42). On the other hand, we do not know the scope of authority of Μαριανὸς ἐπιτροπεύσας τοῦ κυρίου Καίσαρος in SB VIII 9904 (154). We know of only one example in Latin of the formula procurator Caesaris applied to an eques (IK XVII.1, 3044), but is this title securely established? It is heavily restored and only ]s (hed.) vac. n(ostri)[ can be read. As far as we know, this has not elicited comment from the epigraphists, except, in passing, Pflaum (RHDFE 1968, 381). 25. Domaszewski, without further discussion, places the prefects of Berenike among the sexagenarian procurators in Egypt (1908: 165; Lesquier 1918: 427).

Sulpicius Serenus, procurator Augusti

101

of a prefect of the Desert of Berenike, or, more precisely, his cursus resembles two types of careers of prefects of Berenike: like Artorius Priscillus, he might have obtained this place after three statutory militiae, but he had also been prefect of the ala that was stationed at Koptos. This was also the case of several prefects of the desert attested between 176/7 and the first half of the third century, whose title was ἔπαρχος Ὄρους καὶ εἴλης (+ name of the ala).26 So we now know that one could address Sulpicius Serenus as procurator Augusti. The subject of the letter (the work on the well) is no different from that of another letter, addressed to a prefect of the Desert and of the ala Herculiana (O.Did. 38, post 185). Could it be the status of the letter-writer that decided the form of address? An epimeletes, probably a civilian, perhaps a member of the familia, would call the prefect of the desert procurator, while a curator praesidii would address him by the more military title of prefect? No: in the report of the curator of Didymoi to Valerius Apolinaris (Chapter 6 [c. 219]) and in a petition of private soldiers (prosopography no. 18 [ante 183]), the prefect of the Desert is adressed as ἐπιτρόπῳ Σεβαστοῦ. Let us not forget that in P.Turner 34, Vettius Gallianus, prefect of the Desert and deputy-epistrategos, in an insertion, is styled ἐπιτρόπων μέγιστε. One could also invoke the honorific epithet κράτιστος qualifying ἔπαρχος in the case of Claudius Lucilianus (P.Basel 2, 190). Vir egregius, rendered κράτιστος in Greek, became the technical epithet for equestrian procurators from the reign of Marcus Aurelius.27 The very title of praefectus is not in itself against the idea that the prefecture of Berenike was a procuratorship.28 A number of procurators carried the title of prefect, to which sometimes that of procurator was added. Cf. the procurators governors of Sardinia, who were first called praefecti (provinciae) Sardiniae, but subsequently added to this title procurator, before or after praefectus Sardiniae, sometimes but not invariably, connecting the two titles by et. After Claudius and Nero, the prefects of the navy were integrated into the procuratorial career, and it then happens (but rarely) that the commanders let their title of procurator appear next to that of praefectus. Examples are Iulius Vehilius Gratus Iulianus, styled proc. Aug. et praef. classis Po[ntic]ae,29 and C. Maenius Agrippa L. Tusidius Campester, who was proc. Aug. praef. classis Brittannicae (CIL XI 5632 = ILS 2735). The hypothesis that the title procurator Montis was used, before the third century, or at least that the praefectura Montis was of a procuratorial nature, brings us back to a reading problem in O.Krok. 41.47. This is a collection of circulars addressed by the very same Artorius Priscillus to the curators of the praesidia on the road to Myos Hormos. In the prescripts of these circulars, Artorius Priscillus sometimes just gives his name, sometimes calls himself ἔπαρχος Ὄρους or ἔπαρχος Ὄρους Βερενίκης. In line 47 the title is further elaborated, but could not be read: Ἀρτώρις Πρισκίλλο(ς) vac. ἔπαρχος Ὄρους vac. κ̣α̣ὶ̣ vac. εους Βερενίκης κτλ. 26. Prosopography nos. 17–21 and 25. Sulpicius Serenus could not have become procurator before he had accomplished the prefecture of the ala Vocontiorum. Consequently, we think that, unlike the later prefects of Berenike who have the combined titles, he became prefect of the desert after being praefectus alae. Unfortunately, we do not know the cursus of any of these people. 27. Hirschfeld 1913: 652. In Egypt this epithet was used from the time of Nero for the prefect of Egypt, and from Trajan’s time it was extended also to equestrian procurators and even to freedmen. It was thus just a sign of respect and not an indicator of official rank (Rangtitel) in the hierarchy of equestrian civil servants. It appears, nevertheless, always to have civilian, procuratorial connotations, and is, for example, never used in the papyri of a prefect of an ala or a cohort. Note that in 109 the prefect of Berenike, Artorius Priscillus, is sometimes styled κράτιστος (O.Krok. 47.42, 57; 49.8; 64.5). 28. I come back to this question in Chapter 6. 29. CIL VI 31856 = ILS 1327. One notes that in this cursus, the title of procurator Augusti is added to that of prefect of the Black Sea fleet, which was a centenary procuratorship, but not to those of prefect of the pretorian fleet of Misenum or of Ravenna, both of them ducenary procuratorships.

102

Rome in Egypt’s Eastern Desert

On this line there is the following commentary: “At the end, Ὄ̣ρ̣ους is not impossible and seems to be preceded by a suspended letter. I am tempted to restitute ἐπ̣ι̣τ̣ρ̣ο̣π̣( ). Apart from the ε the letters cannot be identified, since they are almost totally washed out.” Having re-examined the original I think that the following reading is compatible with the traces: ἔπαρχος Ὄρο vac. καὶ vac. ἐπ̣[ίτ]ρ̣ο̣π̣ο̣ς̣ Ὄ̣ρους. The pleonastic expression is suspect,30 and we must accept that it is not just a scribal confusion, but if the reading is accepted, we must conclude that, at this time, the titulature of the prefect of Berenike had a variant that underlined its procuratorial character, but which was little used. This also holds good for the titles of the prefects of non-pretorian fleets.31 Another possibility is that Artorius Priscillus cumulated two functions which, in other circumstances (O.Krok. 87?) would have been discharged by two different men. But if so, we should expect to find the procuratorship mentioned in his epigraphical cursus. In fact, Artorius Priscillus does not list the title of procurator Montis Berenicidis when he presents it in an inscription from Puteoli (CIL VI 32929 = ILS 2700, Prosopography no. 9). The inscriptions and the ostraca let us follow a territorial prefecture for two centuries. The oldest prefects attested are military officers in service. In the case of Severius Severus (no. 7), who is appointed directly as prefect of the desert after some municipal offices, the prefecture of the Desert must be considered as a first militia.32 His career resembles that of Pinarius Natta (no. 2) who also originated in the municipal elite, and who discharged the office of prefect after, or at the same time as he was legionary tribune, which at the time (under Tiberius) counted as the first militia. These prefects of the early empire are thus people with little or no military experience. They accumulate the prefecture of Berenike along with their first militia (nos. 1 and probably 2), or they are municipal gentry for whom the prefecture of Berenike offers a shortcut to entry into the equestrian ranks. In fact, we know from Strabo (17.1.4) that the Eastern Desert was a secure and almost uninhabited region at the beginning of the Roman occupation. In 109, Artorius Priscillus (no. 9) is the first prefect of Berenike who can be seen to have a high level of competence for the job. He had much military experience, since he had accomplished his three militiae. He did not command any military unit in Egypt, and the prefecture of Berenike is probably his first procuratorial post. This change in the profile of the prefects of Berenike is probably not without connection to the mounting danger from the nomads in the Eastern Desert which is clearly attested by the ostraca, of which the earliest date from 108, 109, and 118,33 but it can also be explained by a development in the procuratorial system. Sulpicius Serenus, under Hadrian, is the first in a series of prefects who are sometimes called ἐπίτροπος by their correspondents. This title refers to their rank in the hierarchy of equestrian careersystem, but the official term for their function remains ἔπαρχος.34

30. One would rather expect ἔπαρχος καὶ ἐπίτροπος Ὄρους Βερνίκης. 31. P.Mil.Vogl. IV 230 (133) mentions a certain Valentinus, whose title is ἔπαρχος καὶ ἐπίτροπος Σεβαστοῦ. Unfortunately the context is broken and the mention of the Prosopite nome suggests that we are far from Upper Egypt, but to the best of our knowledge, this is the only praefectus et procurator mentioned clearly and unambiguously in Egyptian documentation. 32. We owe this observation to P. Le Roux. The inscription is not dated. There is an example that another territorial prefecture, the praefectura orae maritimae, counts as a first militia (Devijver 1972: 185). It is worth noting that the definition of the prefecture of Severius Severus is particularly precise, since his title includes the element praesidiorum, which is unique. If it is correct to assume that the praesidia are a Vespasianic creation (Cuvigny [ed.] 2006: 356 f.), the fact that they are mentioned provides a terminus post quem. The mention praesidiorum would be explained by the novelty of mass construction of praesidia. The profiles of later prefects also favor a date for Severius Severus no later than the Flavian period. 33. O.Krok. I 6, 47, 51, and 87. 34. There is an analogy here to the epistrategoi. They are sometimes in petitions appealed to as ἐπιτρόπων μέγιστε, but the term ἐπίτροπος is avoided when referring to their function, although they are procuratores (Thomas 1982: 47–50; there is however, an exception, to which David Thomas kindly drew my attention, P.Rain.Cent. 68.24 [c. 235]).

Sulpicius Serenus, procurator Augusti

103

So, the idea that the prefecture of Berenike had become a procuratorship35 appears to us the most economical explanation of the career of Artorius Priscillus and of the title ἐπίτροπος Σεβαστοῦ as given to Sulpicius Serenus. This post was thus no longer given to an equestrian military officer at the beginning of his career, but to a tribus militiis functus.36 But although the posting became a procuratorship, it kept its military character, first, because the administration of the empty expanse of the desert could only be entrusted to the army, and second, to ensure that the garrisons, constituted of soldiers from many different units, would be under a single command. It is legitimate, then, to draw a parallel between the prefect of Berenike, a procurator who governed a district and had direct authority over the garrisons, and the procurators who were at the same time praepositi of a region or of one or more vexillations.37 During the third century appears the title praepositus limitis, which defines the jurisdiction of some of these procurators of districts.38 To this group belong also the praepositi praetenturae of P.Dura 64 (221) and of P.Euphr. 3 and 4 (252–256). The first, Aurelius Rufinus, has the title procurator Aug(ustorum) n(ostrorum duorum) praep(ositus) praet(enturae); the second, Iulius Proculus, is ἔπαρχος πραιπόσιτος πρετεντούρης, and, according to Jean Gascou, his function is no different from that of the ἐπίτροπος ἐν Ἀππαδάνᾳ in P.Euphr. 1.14 and 2.16.39 If this equivalence is correct,40 the description of this posting shows the same balance between ἔπαρχος and ἐπίτροπος as we see in the case of the prefects of Berenike. For unknown reasons, from 176/7,41 there is a series of prefects of Berenike who cumulate, along with their territorial prefecture, also the prefecture of the ala in Koptos. But is this cumulation, which is rather the opposite of what may be seen in other provinces,42 compatible with the hypothesis that the prefect of Berenike was a procurator? Admittedly, being prefect of an ala is a third-militia-posting and a prerequisite for advancement in the equestrian hierarchy to a procuratorship. If the prefecture of Berenike was still a procuratorship, we have to suppose that the prefecture of the ala in Koptos was not a third-militia-posting, but was reserved for the prefects of the desert. This particular status of the ala in Koptos was perhaps connected to the fact that in the combined titles (ἔπαρχος Ὄρους καὶ εἴλης) the prefecture of the desert always comes first, contrary to what may be observed at the beginning of the empire, when a prefecture of a territory, a tribe, or of civitates was conferred on a prefect of a cohors or an ala, or on a legionary tribune.43 35. Presumably under Claudius, see Chapter 6. 36. Others were appointed to a territorial praefectura directly after their tria militia. But were these prefectures really procuratorships, or were they on the borderline between civil and military, which prepared for a procuratorship without, properly speaking, being procuratorial postings? See for example AE 1926, 20 (ripa Danuvi), AE 1972, 573 (ora Pontica maritima), ILS 2709 (ripa fluminis Euphratis). In AE 2000, 466, the title is even procurator ripae. 37. Y. Le Bohec (1991: 319–22) has investigated the juxtaposition of procurator and praepositus and made a list of them. 38. E.g., Lucretio Marcello v(iro) e(gregio) proc(uratore) Aug(ustorum) n(ostrorum duorum), praeposito limitis Tripolitanae (Bu Njem, 248: AE 1985, 849). 39. Gascou 1999: 67 f. Gascou proposes to explain praetentura as “a military road or a fortified communication line.” It reminds us of the two roads with their praesidia, which formed the principal axes of the Desert of Berenike. 40. It is accepted as highly plausible by Eich 2005: 133 f. 41. Prosopography no. 17. This date is that of the inscription (I.Did. 3) commemorating the restoration of the praesidium of Didymoi, where a break in the stratigraphic continuity suggests that it was abandoned between c. 150 and this date. This interruption of occupation could have been caused by the “Antonine plague.” The reconstruction of Didymoi is perhaps related to the reassertion of authority after the usurpation of Avidius Cassius, and to the visit of Marcus Aurelius to Alexandria during the winter 175/6. 42. Leveau 1973: 181 asks the question: “How would it be possible at the same time to have a military command and be the prefect of a tribe?” and he observes (p. 182): “the only cases where an officer is at the same time prefect of a tribe and in command of a military unit are the oldest in each regional series.” 43. See the examples collected by Leveau 1973: 181 f.

104

Rome in Egypt’s Eastern Desert

The last stage in the history of the prefecture of Berenike comes in 219 (see Chapter 6),44 when the procuratorial character of the office is seen in that the function is named ἐπίτροπος Ὄρους, while ἐπίτροπος (Σεβαστοῦ) was, until that time, just a form of address.

44. One cannot, however, exclude that the ignotus no. 21, whose title would normally place him at the time of nos. 17–25, was in fact later. The gentilicium Aurelius suggest a date after 212, and the ostracon comes from a filling-layer which, according to J.-P. Brun, was formed around 240. If the ostracon is contemporary with the formation of this layer, it would bear witness to the durability of this double prefecture and of the double title that follows, but it may also be a residual ostracon.

I.Portes 67 CIL IX 3083 = ILS 2699

6. L. Antistius Asiaticus

7. D. Severius Severus

O.Krok. (passim) January– May 109

1st–2nd cent.

90

unspecified

no

unspecified

unspecified

unspecified

probably early case of a praef. Ber. chosen from civic magistrates, not experienced army officers. Was aedilis and IVvir i.d. before his praefectura, and scriba quaestorius afterwards

praefectus praesidiorum et Montis Beronices

ἔπαρχος, ἔπαρχος Ὄρους

had the Koptos Tariff engraved by order of the prefect of Egypt

had been earlier aedilis and IIvir quinquennalis

metallarches of all metalla in Egypt

is also archimetallarches of all metalla in Egypt. His freedman is procurator metallorum

Remarks

ἔπαρχος Ὄρους Βερενίκης

praefectus Montis Bernicidis

praefectus Bernicidis

ἔπαρχος Βερενίκης

praefectus Ber(e)nicidis

tribunus legionis III

unspecified

unspecified

ἔπαρχος Βερενίκης

tribunus legionis III

unspecified

Title as Desert Governor

Military Command Exercised Before or at the Same Time as the Desert Governorate

1. Ast (2020). I thank Rodney Ast for information concerning this new prefect, as well as prefect no. 13bis.

8. Cosconius - -tulus

AE 2001, 2039

5. M. Trebonius Valens

Sulmo (Latium)

72

I.Memnon 4

4. L. Iunius Calvinus 76/77

51

two dedications1

3. P. Iulius Rufus

Tiberius

18

I.KoKo. 41 CIL X 1129 = ILS 2698 (cursus)

11

Date AD

I.Pan 51

Documents

2. L.Pinarius Natta

1. P. Iuventius Rufus

Name

Table 3.1. The prefects of Berenike: prosopography

Sulpicius Serenus, procurator Augusti 105

Documents

Cassius - -

Cassius Ta[

O.Krok. 91

O.Krok. 88 c. 118

c. 118

March 118

O.Krok. 87

12. - -sius Taurus

unspecified

unspecified

unspecified

perhaps the same --] Ὄρους as M. Arruntius Agrippinus, praefectus cohortis in Dacia

March 118

O.Krok. 87

ἔπαρχος Ὄρου(ς)

ἔπαρχος Ὄρου(ς)

ἔπ[

simply ἔπαρχος

11. Arruntius Agrippinus?

ἔπαρχος Ὄρους

praefectus Montis

unspecified

unspecified

generally ἔπαρχος Ὄρους Βερενίκης, once ἔπαρχος Ὄρους κ̣α̣ὶ̣ ἐπ̣[ίτ]ρ̣ο̣π̣ο̣ς̣ Ὄ̣ρους Βερενίκης. Mentioned in several circulars as ὁ κράτιστος ἔπαρχος Ἀρτώριος Πρισκίλλος

would he be procurator Montis while Cassius Taurus is praefectus Montis? Or does he belong to Cassius Taurus’s officium?

is praef. Ber. after 3 militiae equestres; afterwards is flamen divi Augusti, patronus coloniae, then epistrategos of Thebaid

praefectus Montis Berenicidis

O.Krok. 65

115/116

unspecified

was praefectus alae I Pannoniorum before being prefect of Berenike

Remarks

Title as Desert Governor

unspecified

I.Dios inv. 262

July– November 109

Date AD

O.Krok. 60

10. L. Cassius Taurinus

O.Krok. (passim)

9. M. Artorius Priscillus ILS 2700 (Puteoli, cursus)

Name

Military Command Exercised Before or at the Same Time as the Desert Governorate

106 Rome in Egypt’s Eastern Desert

ἔπαρχος Ὄρους Βερενείκης καὶ εἴλης Ἡρακλιανῆς praefectus alae Herculianae

post 183

SEG XLVIII 1977

19. Aemilius Celer

ἐπίτροπος Σεβαστοῦ καὶ ἔπαρχος ἴλης Βουκουντίων praefectus alae Vocontiorum

18. Licinnius Licinnianus

τῷ κρατίστῳ ἐπάρχῳ ἡμῶν

ante 183

I.Did. 3

17. [- -]

161

ὁ κράτισ[τος --]|[---]κις Σεουῆρος

ἐπίτροπος Σεβαστοῦ

ἐπίτροπος Σεβαστοῦ

O.Dios inv. 1460

O.Xer. inv. 257 and 618

16. Volussius Vindicianus

mid-2nd cent.

unspecified

unspecified

perhaps the same as Sulpicius Severus, addressee in a petition of two Dacian soldiers found in Umm Balad (O.KaLa. inv. 37, reign of Antoninus according to stratigraphy)

responsible for material and imperial workforce (familia)

ἐπίτροπος Σεβαστοῦ

[ἔπαρχος] Ὄρους καὶ εἴλης --

O.Dios inv. 712

15. - -cius Severus

discarded c. 140-150

123

commemorates his victory over Agriophagoi

Remarks

unspecified

ἐπ[

Title as Desert Governor

prefect of the ala [- -]

O.Did. 27

14. [- -]

unspecified

praefectus alae Vocontiorum

Hadrian unspecified (according to stratigraphy)

Hadrian

121/122

Date AD

176/177

Unpublished dedication found in Berenike (inv. 119002)

13bis. Claudius Cerealis

O.Dios inv. 90

I.Pan 87

Sulpicius Gn. f. Serenus

[- -] Sulpicius Serenus

I.Memnon 20 (cursus)

Documents

13. Servius Sulpicius [- -]

Name

Military Command Exercised Before or at the Same Time as the Desert Governorate

Sulpicius Serenus, procurator Augusti 107

Documents

end 2nd cent.

I.Did. 4

22. - -anius - -

217/218

lacuna

unspecified

O.Xer. inv. 620

216

praefectus alae Herculianae

P.Turner 34

25. Vettius Gallianus

beg. third cent.

lacuna

[ἐπάρχῳ] Ὄρους καὶ εἴλης Ἡρακλι[ανῆς]

ἐπιτρόπων [μέγισ]τε

κράτιστος ἔπαρχος Ὄρους

ἔπ[αρ]χος [Ὄρο]υς

title in lacuna, but known as ἔπαρχος in O.Claud. IV 849–850

- - Ὄρο]υς καὶ [

ἐπάρχῳ Ὄρους καὶ εἴλης Ἡρα[

[ἐπάρχου Ὄρ]ους καὶ εἴλης Ἡ[ρακλιανῆς]

praefectus alae Herculianae praefectus alae Herculianae

κράτιστος ἔπαρχος

Title as Desert Governor

unspecified

O.Xer. inv. 888

I.Did. 5

24. Deiotaros

180s–beg. 3rd cent.

post 183, discarded after 235, perhaps beg. third cent.

Ο.Did. 38

21. - -nias

23. Antonius Flavianus? O.Dios inv. 514

Commodus

190

Date AD

I.Did. 9

20. Claudius Lucilianus P.Bas. 2

Name

Military Command Exercised Before or at the Same Time as the Desert Governorate

simultaneously epistrategos sc. of Thebaïd: διαδεχόμενος καὶ τὴν ἐπιστρατηγίαν

addressee of a report from the curator of Dios

Remarks

108 Rome in Egypt’s Eastern Desert

O.Xer. inv. 187

28. ] Iulianus

I.Memnon 14 (cursus) dedication of the praesidium at Biʾr al-Hammamat (Cuvigny [ed.] 2006: 93)

29. Caesellius Q. f.

30. --bo or --bus

Undated prefects of Berenike

O.Xer. inv. 237

s.d.

s.d.

third cent. (found in the fort)

post 212

Mesore

O.Xer. inv. 731

27. Salvius Menandros

219 (year 2, Pachon 10+)

O.Dios inv. 457

praefectus alae Gallicae

--

unspecified

unspecified

219 (year unspecified 2, Pauni 3)

unspecified

unspecified

O.Xer. inv. 601

[219]

I.Portes 87

unspecified

beg. third cent. according to stratigraphy

219

I.Portes 86 (Koptos)

26. Valerius Apolinaris

P.Bingen 108 = O.Did. 40

Date AD

Documents

Name

Military Command Exercised Before or at the Same Time as the Desert Governorate

praefectus item Ber[ ]

ἐ̣π̣[

τῷ κρατίστῳ ἐπάρχῳ ἡμῶν

ἔπαρχος Ὄρου⟨ς⟩

unspecified

ὁ κράτιστος ἡμῶν ἔπαρχος

ἐπίτροπος Σεβαστοῦ

[---Ὄρ]ους

ἐπίτροπος Ὄρους

Title as Desert Governor

one cannot decide if the two praefecturae were exercised simultaneously or successively

mistakenly called Claudius Apolinaris

Remarks

Sulpicius Serenus, procurator Augusti 109

110

Rome in Egypt’s Eastern Desert

Appendix The geomorphological context of the well at Dios/Iovis Maël Crépy45

The study of topographic46 and geological47 maps, completed by observations from satellite images, allows us to give a preliminary overview of the geomorphological context of the praesidium of Dios; this in turn gives a better understanding of the text of O.Dios inv. 90. The combination of this textual source with the geomorphological data allows us to advance hypotheses of interpretation concerning the supply and functioning of the well at Dios. The translation of the text established by Adam Bülow-Jacobsen and Hélène Cuvigny shows that, during the time it was used, this well was subject to at least one phase of maintenance, that this maintenance was not a simple cleaning, and that the deepening or enlargement of the bottom section of the well was obstructed by the hardness of the local rock. The study of the local geology gives a better understanding of the origin of these difficulties.

Functioning of the well at Dios The well is situated in the flood plain of Wadi Qurayya, of which the upstream catchment is rather small (less than 30 km2), which means that the affluence of water at the time of rain or flash floods is quite limited. The topographic surface consists of quaternary, alluvial, or windblown deposits that lie directly on a basement of the Late Proterozoic rock. This basement is heterogeneous, which causes a very complex geology around Dios, where no less than seven different formations are present (see Table 3.2). Although the maps are not very detailed, we see that the well is on, or quite close to, a zone where three of these formations are in contact (in bold on Table 1: gd, mvb, and osm). This might be the reason why the well was dug in this sector, since the geological contact-zones may be especially favorable to the concentration, the conservation, and the mounting to the surface of water from the water-carrying layers. In this context one may imagine various types of functioning and supply of the well: - Exploitation of the alluvial water-table, which the geological contact limited. This favored the formation of important water reserves at times of rain, while the impermeable underlying rock prevented absorption into lower layers. The functioning of this kind of well is largely dependent on seasonal rain and provides an irregular source of water during the year. The operation described in O.Dios inv. 90 would thus be the result of a wish to enlarge this reserve and augment its volume. - Exploitation of a water-bearing layer or a fault of the underlying bedrock. According to the type of water-table, the refilling of the well will be largely dependent on the amount of rain during the time of use of the well (as for wells in the alluvial water-table), or, on the contrary, it could depend on water reserves from earlier, more humid, periods. The variation in the water reserve is thus not seasonal, but depends on a progressive depletion of the accessible water reserves. The operation will thus have consisted in a deepening or a widening of the bottom section of the well into the bedrock, in order to enlarge the draining surface and thus counter a lowering of the well’s capacity. 45. Chercheur associé à l’UMR5133 – Archéorient, CDD Recherche à l’UMR5189 – HiSoMA, ERC-project Desert Networks directed by Bérangère Redon. 46. Sheet NG36 G3a Habal Al-Yutaymah 1/50 000 by the Egyptian General Survey Authority and by Finnida, 1989. 47. Z. S. El Alfy and M. O. Mostafa (2009), Geologic Map of Wadi al-Barramiyah Quadrangle, Egypt, EGSMA, BGF, Ministry of Petroleum and Mineral Resources – The Egyptian Mineral Resources Authority.

Sulpicius Serenus, procurator Augusti

111

Table 3.2. Geological formations of the Late Proterozoic in the vicinity of Dios.1 In bold: three formations that are in contact in the immediate vicinity of Dios. Code gk1

Group

Rocks

Post-tectonic alkaline granitoids Pink-gray coarse-grained biotite hornblende granite (porphyritic in part) with abundant amphibolite, metavolcanic xenoliths and porphyritic felsite sheets

gd

Syn-late tectonic calc-alkaline granitoids

Gray-off white coarse to medium-grained biotite hornblende granodiorite

gb

Mafic intrusion of syn-late tectonics

Dark gray melanocratic medium- to very coarse-grained hornblende pyroxene olivine gabbro with minor amount of layered anorthosite and pyroxenite

mvb

Schist

vi-a

Calc-alkaline volcanic

Chlorite, amphibole, sericite schist Predominantly acidic to intermediate aphyric lavas and some pyroclastics interdigitating with volcaniclastics sediments; sporadic propylite alteration, feldspar-phyric andesites

vb

Calc-alkaline volcanic

Aphanitic basalt locally pillowed with hyaloclastite breccia and microgabbro

osm

Ophiolite

Tectonic melange comprising serpentine, talc, and magnesite, pillow basaltic lava and rare microgabbro and microdiorite

1. The formations gk1, gb, vi-a, and vb are indicated even if they do not directly concern the upper layers of the well of Dios, since, in the absence of data from local geological drilling, we cannot know if the well traversed these layers deeper down.

The small size of the upstream catchment,48 the thin quaternary sedimentary layer,49 the local geology,50 and the way in which the author of O.Dios inv. 90 describes the variation in volume,51 and his difficulties with digging,52 would point to a well of the second type, which exploited the water in a fault in the bedrock. The operation therefore appears to be an attempt to solve the problem of a progressive, longer-term, reduction of the yield of the accessible resources and a drawdown of the water-table.53

The maintenance operation The study of the geology and the local lithology also lets us understand why the author of the ostracon describes such difficulties in the digging, and why he demands four strong men of the familia, a smith, and 15 minae of hammered steel. The geological map lets us set up a list of the rocks in the region of the well, and gives us an idea of their respective hardness on the Mohs-scale.54 The hardness of steel is estimated at 5.5 to 7 depending on its treatment and quality. When we compare the hardness of the 48. Which considerably limits the potential refilling of the well when it rains. 49. Which limits the potential water-carrying capacity and does not protect against evaporation. 50. The bedrock, characterized by schistosity (mvb) or fracturation (osm), does not provide a good floor for a water-table in the wadi alluvium, since they are open to absorption into the lower layers. 51. The lowering of volume does not seem to correspond to a short-term, seasonal variation, but rather to a reduction observed over a longer period (one year or more). 52. Digging in the quaternary deposits only would not present the difficulties that the author of O.Dios inv. 90 mentions. 53. Either because more water was used than flowed in, or because the water table became lower over a long period, thousands or tens of thousands of years, or because of a series of years with little rain that did not let the well refill. 54. This scale is ordinal and relative, not linear and not logarithmic, and goes from 1 to 10, classifying minerals by comparison. It is based on the capacity of a material to make scratches in another. Talc is 1 and diamond is 10.

112

Rome in Egypt’s Eastern Desert

Table 3.3. Hardness (Mohs) of the rocks present in the vicinity of Dios. The hardness-values (which may vary according to the exact composition of the rocks) are not results of experiments on the spot, but in other parts of the world. They are thus indicative only. Formation Code gd mvb

osm

Rock

Hardness (Mohs Scale)

Granodiorite

6 to 7

Chlorite-schist

highly variable

Amphibole-schist

5 to 6

Sericite-schist

highly variable

Serpentine

2.5 to 3.5

Talc

1

Magnesite

around 4

Microgabbro

5 to 6

Basalt

7 to 7.5

rocks found at Dios (Table 3.3), we understand better what the epimeletes was up against: except for talc, magnesite, and serpentine, all the rocks are as hard as or harder than the steel of the points (in bold in Table 3.3).55 Before reaching the bedrock, the well at Dios first goes through recent layers of wadi bottom alluvium and/or wind-blown sand, which are easy to penetrate. Part of the layer of the basement is perhaps particularly easy to drill (talc, magnesite, serpentine), but makes the maintenance of the walls of the well difficult and may need lathing or tubing. The rest of the rocks are much harder, which makes the digging difficult, but assures much better stability of the well-shaft, even without lathing. It seems that the epimeletes of the wells has encountered some of these hard rocks when making his analysis, and was penetrating hard bedrock, which confirms that this was an exploitation of a water-carrying layer in the deep underground.56 At present, certain questions remain unanswered: Was the well dug into formations of serpentine and talc, before it arrived in the harder rock? Did the epimeletes encounter a dike of basalt? Did the wellshaft pass through hard rock only? Did the operation succeed in augmenting the production of the well? Without studies on the spot (geomorphological prospection, excavation and stratigraphy of the well) we can unfortunately not know. Even if some questions remain unanswered, this example still shows the great potential of interdisciplinary research between the known geological data and the information from the ostraca in the study of the exploitation of the Eastern Desert of Egypt. It permits us, even without studies on the spot, to identify clearly the functioning and supply of a well during antiquity, something that would have been impossible while working in only one of these disciplines.

55. See also the paper on stomoma referred to in note 7 above. 56. The only other phenomenon which might explain such a difficulty for the diggers would have been a progressive formation in the bottom of the well of very pure calcium carbonate, but this seems excluded by the local geology.

4 Claudius Lucilianus, prefect of an ala and of Berenike The excavation season of December 1999–January 2000 at Didymoi brought to light several inscriptions, one of which, registered under the inventory number 940, gave us the damaged name of a new prefect of Berenike. Only the cognomen remains. In reading it, one might hesitate between Κα]ι̣κιλιαν[-- and Λο]υ̣κιλιαν[--, but careful examination of the lettering requires the second of these: of the letter preceding kappa, only the bottom of the hasta is visible, and this stroke is too far from the kappa to belong to an iota (iota in this inscription is tightly spaced with neighboring letters); in some lights, one may also pick out a curve toward the left, which is the start of the left branch of the upsilon. A search of the DDbDP turns up just one attestation of the cognomen Lucilianus (which is, moreover, generally uncommon) in the papyri. It appears in P.Bas. 2, a contract dated to 190, in which two cameleers of Soknopaiou Nesos undertake to transport three requisitioned (?) camels: ὁμολο[γοῦμε]ν παρειληφ[έ]ναι παρʼ ὑμῶν καμήλους ἀπο[ ὑπὸ το]ῦ κρατίστου ἐπάρχο[υ] Κλαυδίου Λουκιλιανοῦ κτλ. (ll. 6–8). Claudius Lucilianus for long appeared in lists of the prefects of Egypt, but G.  Bastianini, in the addenda et corrigenda that he published in 1980, placed a question mark after his name, on the basis of reservations expressed per litt. by J. D. Thomas:1 besides the content of the papyrus, the honorific epithet for the prefect of Egypt was, at this period, not κράτιστος, but λαμπρότατος and, moreover, when κράτιστος is applied to the prefect of Egypt, it modifies the title ἡγεμών, not ἔπαρχος. Bastianini concludes, “Forse, Claudius Lucilianus è un praefectus militare. La questione richiede approfondimento ulteriore.” Subsequently, he placed Claudius Lucilianus squarely in the list of “eliminated” prefects, with the notation “forse un prefetto di tipo militare.”2 The inscription of Didymoi has a good likelihood of settling the question.

1. Bastianini 1980: 84 and n. 5. 2. Bastianini 1988: 516.

113

114

Rome in Egypt’s Eastern Desert

Figure 44. I.Did. 9. © A. Bülow-Jacobsen

I.Did. inv. 940 (= I.Did. 9; fig. 44) Sandstone plaque, 27 × 29 × c. 5 cm thick. Average height of letters: 2 cm. US 12011. Found among the rubble in a destruction layer of room 120 (a room in the northwest corner of the praesidium, immediately next to the chapel). Around 190 (on the basis of P.Bas. 2).

4

8

– – – – – – – ][ Λο]υ̣κιλιαν̣[1–2] [ἐπάρχου Ὄρ]ους καὶ εἴλης Ἡ̣[ρα-] [κλιανῆς Ἰ]ούλιος Κλήμης [κουράτωρ] πραισιδίου ἔκτι- vac. [σε σὺν τ]οῖς σὺν αὑτῷ στρατι[ώταις εὐ]χ̣αριστίας χάριν ἐπ[1–2] [ c. 6 Ἀμ]μώνιος Μέλα[] [ c. 8 ]ανός vac. (ἔτους?) (.)[ c. 7 ] [ c. 8 ]ο̣υ Καίσ̣[αρος ]ι vac. [

Claudius Lucilianus

115

2–3.

The name of the prefect was probably in the genitive, introduced by ἐπί (“under…”); so it appears in the Greek inscription on the altar of Yarhibol found at Berenike (SEG XLVIII 1977): ἐπὶ Αἰμιλίου Κέλερος ἐπάρχου Ὄρους Βερενείκης καὶ εἴλης Ἡρακλιανῆς. In my view, actually, the gratitude expressed in line 7 is not addressed to the prefect (in which case the inscription would be the dedication of a statue of him, which is most improbable in the context of our little garrison), but to a divinity: this is, moreover, always the case with the expression εὐχαριστίας χάριν in Egypt.

3.

After εἴλης, there is the bottom of a vertical hasta, which would suit an eta very well. At least three other prefects of Berenike were also prefects of the ala Herculiana, which came from Palmyra to Koptos between 183 and 185: not only Aemilius Celer, but also Vettius Gallianus (O.Xer. inv. 888) and a man whose (Greek) cognomen ended in -νίας and who appears in O.Did. 38 (probably later than 212: the curator who wrote this letter is an Aurelius).

4.

Other soldiers in the Roman army of Egypt were called Iulius Clemens, including some in the Eastern Desert (O.Max. inv. 253, O.Xer. inv. 4); it is a banal combination.

5.

κουράτωρ is more likely than an accusative as the object of ἔκτισε (“built the … of the praesidium”).

7–8.

ἐπ[. One might hesitate between ἐπ’[ἀ|γαθῷ and ἐπ[ο|ίησεν or ἐπ[ο|ίησαν. So far, the sequence εὐχαριστίας χάριν ἐπ’ ἀγαθῷ has appeared in Egypt only at Philae, in the later principate: I.Philae 174 (200); I.Philae 184 (II–III); I.Philae 271; I.ThSy. 328 (Philae, 186: cf. Bingen, CdE 65 [1990] 159); I.ThSy. 330 (Philae, third century); in all of these cases, the gratitude is addressed to a divinity. The phrase now appears in another dedication from Didymoi, which also comes from the chapel, and which has a terminus post quem of the reign of Caracalla (I.Did. 5). Ἐπ[οίησεν/-σαν would have the merit of providing a justification for the appearance of the personal names in the nominative in fine and would mean “has/have made” (the object that the curator had caused to be built?); but it has to be said that this verb is almost always preceded by its subject in Greek inscriptions.

8.

The letter after Μέλα could be ς (cf. the tightly curved sigma of -τίας in the previous line), or rather, in the view of Jean Bingen, who proposed this solution, φ, a phi of which it appears, in some lights, that we can see the top of the hasta extending as far as the first vertical of the pi at the preceding line. Ammonios (son of Melas?) was apparently named along with several companions. These were probably soldiers (at least in part) who assisted the curator.

9.

(ἔτους?). The remains of the symbol L are visible, and perhaps even (in some lightings) a break in the edge of the broken side which would be the top of the kappa that we would expect in the number of the regnal year. Then, a short titulature of Commodus (Αὐρηλίου Κομμόδ]ου Καίσ[αρος τοῦ κυρίου; I cannot say just how the line breaks ran).

10.

]ι. The most natural reading is ]ο̣ι, but theta instead of omicron is possible if required. Perhaps Φαρμου]θ̣ι λ̣ [ ?

116

Rome in Egypt’s Eastern Desert

It is extremely likely that the Λο]υ̣κιλιαν̣[ of our inscription and the Claudius Lucilianus of P.Bas. 2 are one and the same person: as well as the relative rarity of the cognomen Lucilianus there is the presence of the honorific epithet κράτιστος, which, as J. D. Thomas stressed, is not suitable for an ἔπαρχος Αἰγύπτου. Moreover, there are relatively few instances in Egypt in which an ἔπαρχος is given the epithet κράτιστος, and in almost all cases they concern the prefect of Berenike: P.Aberd. 149.3 (II–III); P.Turner 34.2 (216), O.Krok. Ι 47.57; 64.5–6 (109); O.Dios inv. 712 (second century); O.Xer. inv. 251 and 618 (161), cf. Chapter 17; O.Xer. inv. 601 (219); O.Xer. inv. 237 (third century). The only exception is SB VI 9526.48–49 (200), where it is a praefectus castrorum (ὁ κράτιστος ἔπαρχος τῶν στρατοπέδων). For this reason, it is reasonable to think of restoring ἐν τῷ Κοπτεί]τῃ νομῶι in P.Bas. 2.9, lines 8–9 of which were published as follows: ἄρ[ρενας τρεῖς οὓς κ]αὶ ἀποκαταστήσωμεν καὶ παρα|δώσωμ[εν ἐν τῷ ]τῃ νομῷ. The requisitioned camels would be going to work in the desert of Berenike, like their fellow of P.Lond. II 328 (p. 75), who was borrowed from his master in 163 εἰς κυριακὰς χρείας τῶν ἀπὸ Βερνείκης γεινο(μένων) ποριῶν. The owner of this animal is a villager from Soknopaiou Nesos; it is also from this village that the cameleers charged with bringing the three camels of P.Bas. 2 came (the document says nothing about their owners).3 However, the width of the lacuna as estimated by the first editor, Rabel, is too great for the restoration Κοπτεί], as examination of the photo confirms. Moreover, the hypothesis that the cameleers were undertaking to take the camels to the Koptite nome makes the restoration of the participle in the lacuna of line 7 problematic, and it is difficult as well for reasons already set forth by Rabel:4 the meaning appears to demand “requisition” (the solution finally adopted by Rabel), but it is hard to see what verb would suit, as the whole prefix ἀπο- would instead suggest verbs meaning “send back.”5 I will add that the authority that orders the requisitions of camels (as indeed with all military requisitions) is always the prefect of Egypt, which Claudius Lucilianus was not (or is no longer); thus, for example, in the correspondence sent in P.Flor. II 278, the anonymous officer who orders the strategoi to supply him with camels reminds them that these orders were issued by the prefect of Egypt.6 As a result, I think that the three camels of P.Bas. 2 were actually being sent back to the Arsinoite for some reason by the prefect of Berenike Claudius Lucilianus, but that, instead of returning them to their owner,7 they were being sent for another task, to which the declarants were ordered to take them. If the prefect of Egypt ordered the requisition, it was the officers who benefited from this order who inspected the animals brought to them;8 we even have a case of a rejected animal in W.Chr. 245.21–22: τὸν δὲ λοιπὸν κάμηλον εἷς (l. ἕνα) ἀποβληθέντα9 φανέντα ἀνεπιτήδειον, “the remaining camel, sent back because he appeared unsuitable for service…”. 3. Soknopaiou Nesos, a village wedged between Lake Moeris and the desert, had practically no agricultural land, with the result that its inhabitants specialized in other activities than agriculture, particularly raising camels: most of the sales of camels, declarations of camels, and attestations of taxes on camels come from this village (D. Hobson, “Agricultural land and economic life in Soknopaiou Nesos,” BASP 21 [1984] 89–109). 4. Introd., pp. 12–13 and comm. ad 7. 5. Rabel suggests ἀπο[βληθέντας. But ἀπο[λυθέντας (“dismissed, released,” cf. P.Oxy. LVII 3912.16) would be even more appropriate for the available space. 6. κατεπείγοντος τοῦ καιροῦ τῆς πορείας ἧν εὐτυχῶς ἄγειν μέλλω ἐξαυτῆς κατὰ τὴν ἐνκέλευσιν τοῦ λαμπροτάτου ἡμῶν ἡγεμόνος καμήλους οὗς προσέταξεν ἄρρενας καὶ ῥωμαλέους δυναμένους ταῖς πορείαις ὑπηρετεῖν. See also P.Col. X 288 (330), a declaration of camels in which an edict of the prefect of Egypt Flavius Hyginus, ordering the requisition of a fifth of the animals of each owner, is mentioned (I thank R. S. Bagnall for this reference). 7. It is indeed not possible to translate ἀποκαταστήσωμεν καὶ παραδώσωμ[εν] as “we will return them (to their owners),” because it is the expression used for “deliver a supply to the State.” 8. P.Flor. II 278 r°, ii.17; iii.21; iv. 5. 9. This is the reading of Krebs in the ed. princeps (BGU I 266: to be precise, ἀποβληθέντ[α]); Wilcken corrected this to ἀποκληθέντα̣, a correction that he does not claim to be paleographically certain (“scheint mir eher ἀποκληθέντα dazustehen”) and which does not provide a satisfactory meaning; moreover, unlike ἀποβάλλειν, ἀποκαλεῖν is almost entirely lacking in the papyrological corpus. Preisigke already confirmed Krebs’s initial reading (P.Bas. 2.7n.).

5 Vibius Alexandros, prefect and epistrategos of the Heptanomia A new epistrategos of the Heptanomia appeared recently in a Leipzig papyrus published by Ruth Duttenhöfer (P.Lips. II 146). The papyrus is a request addressed to the epistrategos Vibius Alexandros by a citizen of Antinoou polis1 concerning a dispute that he had with the komogrammateus of a village in the Arsinoite nome.2 The request is not dated, but it belongs to a dossier that makes it possible to place it between 18 January and 5 April 189. It happens that the 1992 excavations at Mons Claudianus yielded an ostracon, which is published below, containing a copy of two memoranda dated to 1 March 189 and addressed by the vice-curator of Claudianus to, respectively, Vibius Alexandros and the procurator Tertullus. In this ostracon, Vibius Alexandros is not called epistrategos, but ἔπαρχος, prefect. It is likely, but not certain, that he combined the two offices. The posts of epistrategos and of prefect maintained close relations, which can be explained by the fact that they were hierarchically close in the equestrian career ladder,3 as these Egyptian examples illustrate: –Artorius Priscillus is attested in 109 as prefect of Berenike, according to the ostraca of Krokodilo; later on, he was epistrategos of the Thebaid (the terminus ante quem being 117).4 –P.Brookl. 24, iii.46–47 (213–216):5 an officer whose cognomen Marcus is the only part of his name to survive,6 is prefect of the ala Herculiana and acting epistrategos of the Thebaid: Marco praef(ecto) alae Ḥẹṛ(culianae) ạg̣(ente) p̣arṭes epistrategiạẹ. 1. P.Lips. II 146.1–2: Οὐιβίῳ Ἀλεξά[ν]δρῳ τῷ κρατίστῳ ἐπιστρατήγῳ παρὰ Ἀντ[ωνίου Δ]όμνου κτλ. 2. This shows that the competence of Vibius Alexandros, which is not specified in the praescriptum of the request, is the Heptanomia and the Arsinoite. 3. Thomas 1982: 161. 4. Cuvigny (ed.) 2006: 303. 5. The detailed commentary appears in the first edition of this papyrus: Thomas and Davies 1977, esp. 56 for the reading Herc(ulianae). 6. On the use of Marcus as cognomen in this period, see P.Oxy. LVIII 3920.5–6n.

117

118

Rome in Egypt’s Eastern Desert –P.Turner 34 (25 December 216) probably acquaints us with the successor of this Marcus. The papyrus contains a copy of a request addressed to him by an Alexandrian citizen who had landed property at Diospolis Parva.7 The beginning of the preamble is drawn up as follows: [Οὐετ]τίωι Γαλιανῶι [τ]ῷ κρατίστωι ἐπάρχωι Ὄρους διαδεχο(μένῳ) καὶ τὴν ἐπιστρ(ατηγίαν), “To his Excellency Vettius Gallianus, prefect of the Desert and acting epistrategos.” We should understand this as “prefect of the Desert of Berenike,” a title that can also be abbreviated to “prefect of Berenike,” as the Desert of Berenike was the major part of the Eastern Desert adjacent to the Thebaid. In O.Xer. inv. 888 (no exact date), the same Vettius Gallianus is at the same time prefect of the Desert and prefect of the ala Herculiana, a combination found also in the case of three other individuals (Aemilius Celer between 185 and 212, Claudius Lucilianus in 190,8 and -nias in O.Did. 38).9 Vettius Gallianus is attested as ex-epistrategos in P.Erl. 19.10: Οὐεττίου Γ̣αλλιανοῦ τοῦ ἐπιστρα(τηγήσαντος).10 The papyrus is not dated, but in it there is a reference to year 25 of Caracalla (216/217); David Thomas concludes from this that Gallianus had become epistrategos on a regular basis after 25 December 216.11 – Marcius Hermogenes is an uncertain case: praefectus classis Alexandrinae in 134, he may have become epistrategos of the Heptanomia afterward.12

We may confidently discard the case of Ti. Iulius Alexandros, epistrategos of the Thebaid in 42 who, according to Baillet,13 left a graffito in the Valley of the Kings although he was still only a military prefect: the author of the graffito I.Syringes 1733 calls himself Ἀλέξανδρος ἔπαρχος κάστρων Θηβῶν. But a quick look at the facsimile (pl. LXVII) shows that the writing is a Byzantine cursive, and not of the first century, as Baillet wrote.14 Since the first version of this chapter, which appeared in 2002, the awareness that the prefecture of Berenike was not a military post but a procuratorship (cf. Chapters 3 and 6), allows us to formulate the problem of the combination of offices of Vibius Alexandros in slightly different terms. The examples of prefects of Berenike who were at the same time prefects of the ala stationed at Koptos showed that a Roman equestrian could combine a military position with a procuratorship. In the case of the epistrategia, the only example until now of such a combination was that of Marcus in P.Brookl. 24; and Marcus was only acting epistrategos, while Vibius Alexandros is full epistrategos. But what is the definition of his prefecture? Is it a military command? The ostracon does not say so. As Vibius Alexandros appears there as in charge of the supply of a metallon of the desert, one thinks naturally of the prefecture of Berenike, recalling the fact that the first known prefect of Berenike, P. Iuventius Rufus, a tribune of a legion, was 7. Diospolis Parva (modern Hiw) is located about 50 km downstream from Kaine (Qena). 8. Cuvigny (ed.) 2006: 304. 9. In 2002, in the French version of this chapter, I wrote at this point: “Bien qu’il cite la préfecture en premier, c’est évidemment à l’épistratège que s’adresse le requérant de P.Turner 34, comme le montre plus loin le vocatif ἐπιτρόπων [μέγισ]τε (également employé en P.Lips. 146, 4–5).” But documentation found afterward has allowed it to be shown that the prefecture of Berenike was a procuratorship, just like the epistrategia (see Chapters 3 and 6). 10. The restoration of the aorist participle is based on the mention of the epistrategos in office in the previous line: Σεπτιμίου Πατροφίλου τοῦ κρα(τίστου) ἐπιστρ̣(ατήγου). 11. Thomas 1982: 198. 12. Thomas 1982: 209–10. 13. I.Syringes, pp. 441–43. This identification was not rejected by Thomas 1982: 52. 14. R. S. Bagnall (per litt. May 12, 2020) observes that the loan-word κάστρα is in any case completely foreign to the language of the Egyptian documents during the principate, and that there is no pre-fourth century attestation apart from the very dubious one in I.Kanaïs 59bis (Chapter 30).

Vibius Alexandros

119

at the same time director of “all the mines and quarries of Egypt.”15 But the documentary attestations of prefects of Berenike do not stretch further to the north than Ophiates (the metallon of Wadi Samna), which is located 45 km in a direct line south of the latitude of Mons Claudianus; and, of the some 9000 ostraca found at Claudianus, none mentions a prefect of Berenike or the Desert of Berenike. I would thus rather suppose that Vibius Alexandros was a military prefect and that his mandate included the organization and supervision of work in the metalla of the middle part of the Eastern Desert: Mons Claudianus has also produced two drafts of letters on the progress of work, addressed by the quarrymen to the predecessor of Vibius Alexandros, the prefect Antonius Flavianus, whose title ἔπαρχος is not defined any more precisely.16 I wonder if this region may have been encompassed for some period under the general name of Porphyrites: this hypothesis would take account of the fact that the employees of the familia who worked at Mons Claudianus belonged, at least under Antoninus, to the numerus of Porphyrites, arithmos of Claudianus.17 One could even wonder if there was not a short-lived prefecture of Porphyrites, which could have been a procuratorship on the model of Berenike. If Vibius Alexandros commanded an auxiliary unit, this is likely to have been a cohort rather than an ala; in fact, the post-Trajanic ostraca from Mons Claudianus mention only cohorts: – O.Claud. II 266 (mid II): a letter in which a soldier asks a debtor to pay the amount that he owes him to a miles cohortis (στρατιώτῃ χορταρίῳ). – O.Claud. inv. 6038 (mid II): a somewhat enigmatic and fragmentary dipinto. Line 2 runs as follows: β Θρακον ρ[. A mention of the cohors II Thracum, stationed in 131 at Syene and in 143 at Luxor, is probably to be recognized here. – O.Claud. IV 869 is a letter addressed to a centurion of a cohort. – P.Bagnall 8 (185–187): copy of a letter of the prefect of Egypt Pomponius Faustianus to Probus, procurator (metallorum), concerning two soldiers of the same cohort who had been guilty of defection. – O.Claud. inv. 7235:18 mention of a miles cohortis in a petition in Latin dated to 186. – O.Claud. inv. 7363 (222–235):19 dedication made by the soldiers of the cohors II Ituraeorum under Severus Alexander (the last known garrison station of this cohort is Pselchis, in the Dodekaschoinos, in 138). In any case, we do not know the composition of the garrison of middle Egypt. As epistrategos of the Heptanomia, Vibius Alexandros must have had his residence at Antinoou polis, the only Greek city in his epistrategia, or perhaps its neighbor Hermou polis, which was under Diocletian the residence of the praeses of the Thebaid.20 The links between these two great urban centers of middle Egypt and the quarries of the Eastern Desert may be illustrated by a lintel found in 1985 in the antiquities magazine of Hermou polis, a lintel that belonged, according to the inscription painted on it, to a hostelry of the tabularii of Porphyrites and the other quarries (the stone was not in situ and may have been brought from a 15. I.Pan 51 (AD 11) and I.KoKo. 41 (AD 18). 16. O.Claud. IV 849 and 850. But the question becomes more complicated with Antonius Flavianus: an ostracon from Dios found in 2007 showed that he was probably prefect of Berenike (O.Dios inv. 514, cf. Chapter 15, p. 241). It is impossible to say if this ostracon, a letter of the curator of Dios, is contemporary with the letters addressed to him by the workers of Mons Claudianus or not. 17. See O.Claud. III, pp. 36–38 and 81–82. 18. Edition: Chapter 21. 19. Edition: Chapter 12. 20. Thomas 1982: 58.

120

Rome in Egypt’s Eastern Desert

nearby locality, Antinoou polis or Cusae).21 The dipinto confirms the preeminence of Mons Porphyrites in the region, one more argument for supposing that Mons Claudianus was not part of the competence of the prefect of Berenike. The idea that Mons Claudianus could have depended on an authority based in Hermou polis or Antinoou polis is, however, awkward: Kaine, the hub of the roads serving Claudianus and Porphyrites, is much closer to Koptos, the seat of the prefecture of Berenike, than it is to the two large cities of the Heptanomia:22 this observation might lead us to prefer the hypothesis according to which Vibius Alexandros, after exercising a prefecture in the Thebaid, was promoted after 1 March (the date of the ostracon) to the epistrategia of the Heptanomia. O.Claud. inv. 7295 (= SB XXVIII 16941) F.W.I – room 1 NE (5) 14.I.1992 Fig. 45

16.7 × 19 cm

1 March 189

The ostraca dated to the reign of Commodus are not numerous at Mons Claudianus (Table 5.1). Table 5.1. Ostraca dated to the reign of Commodus found at Mons Claudianus inv. 5882

F.SE – s1 NW (14)

year 21

180/181

bottom of a document written in charcoal, of which only the date survives

inv. 8197

F.SE – r1 NW (24)

year 24 (?)

183/4?

bottom of a letter or account concerning σιδήρια

inv. 7235

F.W.I – room 1 SE (4)

consuls

186

petition in Latin

P.Bagnall 8

F.W.I – room 1 SE (7)

prefecture of Pomponius 185–187 Faustianus

O.Claud. III 629

F.W.I – room 1 SW (8) year 29

188/9

receipt for advance

inv. 7295

F.W.I – room 1 NE (5) year 29

1 March 189

copy of letters of the vicecurator of Claudianus

translation of a letter of the prefect of Egypt to Probus (probably the procurator of quarries)

In this series, inv. 7295 is the only one in which the year number is not followed by an imperial titulature, but the mention of a 29th year excludes any other date than to Commodus, which agrees completely with the stratigraphy. To these texts we may usefully add a group of letters addressed collectively either to a prefect (named Antonius Flavianus), or to Probus, “procurator of our lord Caesar,” by the community of quarrymen and smiths; these do not have any preserved date, but this Probus is dated indirectly to c. 185–187 thanks to P.Bagnall 8. These collective letters demonstrate a certain activity in the metallon, which has so far not been able to be connected to any construction carried out under Commodus.23 These work areas had obviously been closed already when Rufus Aristoteles, deputy cu21. Cockle 1996. Uncertainty on the provenance: p. 24. The text of the dipinto is: hosp(itium) tabular(iorum) Porphyr(itis) et aliorum metallorum. 22. Kaine is about 30 km from Koptos and 320 km from Antinoou polis. 23. Maxfield and Peacock 2001: 433.

Vibius Alexandros

121

rator of Claudianus, wrote inv. 7295, in which he recopied in a skilled hand two reports in the form of a hypomnema that he sent respectively to the prefect and to a procurator: on 1 March 189, the date of these reports, the metallon was hardly in functioning condition and the military force in charge barely had the means to guarantee its own survival. The titles of these three individuals require commentary. Ἀντικουράτωρ is to my knowledge unparalleled, which is not a surprise, since it was an exceptional position, created ad hoc in the absence of the titular curator, who had gone back down to the valley for reasons unknown to us. The prefix ἀντι- is the Greek equivalent of the Latin pro- in forming nouns for functions (cf., for example, ἀντιστράτηγος for propraetor, ἀνθύπατος for proconsul); unlike ἀντιστράτηγος and ἀνθύπατος, however, ἀντικουράτωρ is a hybrid, in which the prefix is a semantic calque and the root is a direct borrowing from Latin: it would not have been felicitous, in fact, to call the replacement for a curator procurator, since this word already existed in Latin with other senses; still, this step had already been taken by the author of the receipt O.Did. 62 (end of the second/start of the third century), a donkey-driver to whom the προκουράτωρ of the praesidium of Didymoi had entrusted a sum of money for him to pay it to the curator, who was at Koptos.24 A Latin-speaker might perhaps have devised the title vice curator (cf. O.Bu Njem 101: a vice principalis, head of the post) or subcurator (cf. ἀντεπίτροπος, which translates subprocurator). The conjunction of prefect and procurator appears in the collective letters of c. 186/7 already mentioned. At the date of the present reports, March 189, the prefect and the procurator had been replaced and were now named Vibius Alexandros and Tertullus. The first, as we have seen, has just appeared in a Leipzig papyrus; as for the other, it is tempting to see him as a procurator metallorum. He might then be the same man as Tertullus lib(ertus) proc(urator) m(armorum) n(umidicorum) who, some twenty years later, put up a dedication to the deified Septimius Severus in the quarries of Chemtou.25 Here again, a connection with the two inscriptions of P. Iuventius Rufus is needed, because they show us in place, next to the prefect of Berenike, an ἐπίτροπος τῶν μετάλλων (who in this case is not an imperial freedman, but a freedman of the prefect himself, P. Iuventius Agathopous). The two reports of Rufus Aristoteles, drawn up the same day to point out several scarcities, differ in their content. The second, addressed to the procurator, is very detailed, but it lacks one of the key pieces of information contained in the first (the delay in the delivery of the rations of the current month). This omission may possibly be explained by the fact that it concerned only military rations, which belonged to the prefect’s competence, not that of the procurator metallorum, who was responsible on his side for equipping the servile personnel. The general character of the report addressed to the prefect may perhaps come from the fact that it is only a reminder about an earlier report already mentioned.26 In the report addressed to Tertullus, Rufus Aristoteles refers, on the subject of the donkeys and the familiares, to an original complement that had been reduced dramatically: out of twelve familiares, only two remained; out of five donkeys, there remained only one. He says nothing of the circumstances that produced this depopulation;27 whatever they were, we are distant from the 400 members of the familia reported under Trajan by O.Claud. inv. 1538 (Chapter 11). 24. Since this instance another occurrence of προκουράτωρ has appeared, in O.Dios inv. 626 (first half of the second century according to the stratigraphy). 25. Thus post 211. Kraus 1993: 60 (= AE 1994, 1881). 26. Cf. comm. ad 3–4. 27. It could be at stake in inv. 7316, a long, fragmentary, and nearly incomprehensible letter in which we are dealing with the eparchos, some barbarians, familiares, wheat, and discontented people (ἐγόνγυσαν). It reads - - ἅπαξ ἄραντες τοὺς ὄνους ἀπῆλ[θον?]. Further along, the genitive absolute παρόντος τοῦ οὐηξίλλου shows that the text was written in the same ambience as inv. 7295.

122

Rome in Egypt’s Eastern Desert

Iulius Silvanus, the curator who had departed, leaving a critical situation to Rufus Aristoteles, is known as the addressee of two letters sent by Aurelius Besarion, the curator of Raïma.28 The same Besarion addresses five further letters to two soldiers, Iustus and Rufus: might the latter be Rufus Aristoteles? In one of these letters, Besarion presses his correspondents to be on their guard against the Barbarians.29 We can see that he had good reason for his concern. A semi-literary, bilinear, and unligatured hand; the absence of deletions suggests that these are copies made for the personal archive of the vice-curator, not drafts. The year sign, found in lines 8 and 25, presents a form for which I do not know any parallel ( ).

Figure 45. SB XXVIII 16941. © A. Bülow-Jacobsen 28. Inv. 7352 and 7358, found in F.W.I – room 1 SE (8). 29. Inv. 7251, found in F.W.I – room 1 SE (6).

Vibius Alexandros Οὐιβίῳ Ἀλεξάνδρῳ ἐπάρχῳ. παρὰ Ῥούφου Ἀριστοτέλους ἀντικουράτορος μετάλλου Κλαυδιανοῦ. φανερόν σοι ποιῶ, κύριε, δεδηλωκός σοι τὸ βήξιλλον περὶ ἐπιχρείας τοῦ μετάλλου καὶ σείτου πραιτερίτου μηνὸς Φαμενω⟨θ⟩ καὶ ὑγρῶν οὐκ ὀλίγων. ἔγραψά σοι, κύριε, ἵνα μηδέν σε λανθάνῃ. (ἔτους) κθ vac. μηνὸς Φαμενω⟨θ⟩ vac. ε̅. ——— Τερτύλλῳ ἐπιτρόπῳ. παρὰ Ῥούφου Ἀριστοτ⟨έ⟩λους ἀντικο⟨υ⟩ράτορος ἀποκατα⟨στα⟩θέντος ὑπὸ Ἰουλίου Σιλβανοῦ κουράτορος. φανερόν σοι [ποι]ῶ, κύριε, καταβαίνων ὁ Σιλβανὸς [3–4 δέ]δωκέ μοι κρειθῆς vacat [ c. 6 ] vac. σείτου ἀρτά(βας) γ̅ μά(τια) ε̅ [ c. 10 ] σκέπαρνον, κάδους δύο, c. 10 [ ]ειονα ἄκμονα, ὄνον ἕνα, κ̣[]παλαιὰ ε̅, ἀσκοὺς παλαιοὺς η̅· ἔτι δέ σοι φανερὸν ποιῶ, ἵνα εἰδῇς καὶ φροντίσῃς, ὅτι οὐκ ἔχομεν φαμιλιαρίους εἰ μὴ δύο ἀπὸ δεκαδύο καὶ οἱ λάκκοι κενοί εἰσιν διὰ τὸ μὴ ἔχειν ὑπηρεσίαν καὶ ἀπὸ ὄνων πέντε ἔτι ἕναν ἔχομεν οὔτε ἀσκοὺς ἔχω εἰ μὴ ὀκτὼ παλαιοὺς οὓς κατέλειψέ μοι. (ἔτους) κθ μηνὸς Φαμενω⟨θ⟩ ε̅.

4

8

12

16

20

24 5 l. σίτου

123

14 l. κριθῆς

15 l. σίτου

22 l. ἕνα

“To Vibius Alexandros, prefect, from Rufus Aristoteles, vice-curator of the quarries of Claudianus. I bring to your attention, sir, that the vexillatio has reported to you concerning the equipment of the site and on the delay in the delivery of the wheat for the month of Phamenoth, as well as in that of the liquid goods in a significant quantity. I write to you, sir, so that nothing may escape your attention. Year 29, Phamenoth 5. “To Tertullus, procurator, from Rufus Aristoteles, vice-curator designated by Iulius Silvanus, curator. I bring to your attention, sir, that Silvanus when he went down turned over to me [quantity] of barley, 3 artabas 5 matia of wheat, […], an ax, two buckets (of saqiya), […], one anvil, one donkey, 5 old […], 8 old goatskins. I inform you further, so that you may be informed and take care for it, that we have only two slaves out of twelve, that the cisterns are empty because of a lack of personnel, that out of five donkeys only one remains, and that I have no goatskins apart from the eight old ones that he left me. “Year 29, Phamenoth 5.” 1.

On his original report, Rufus Aristoteles probably started a new line after ἐπάρχῳ (cf. A. Martin, “Τῷ δεῖνι παρὰ τοῦ δεῖνος. Réflexions à propos d’un type documentaire,” PapCongr. XXIV, 661–75).

124

Rome in Egypt’s Eastern Desert

2.

Ἀριστοτέλης is not a common name in Egypt, and until this text it had been attested only in the Ptolemaic period.

3–4.

φανερὸν ποιεῖν is normally construed with an infinitive clause, interrogative or introduced by ὅτι/ὡς (so in line 20), or simply with a direct object in the infinitive; I have, however, found in prose authors some rare examples of constructions with a participial clause: Dem. Mid. 150; Jos. Ant. Jud. 6.91; 8.170; Plut. Aem. 6.8; Ant. 66.7. They support the solution adopted in the translation, which makes of τὸ βήξιλλον the subject of δεδηλωκός, while περί depends on δεδηλωκός, an attested expression. If this interpretation is right, the vice-curator limits himself here to reminding the prefect of the existence of an earlier report made by “the vexillum”; this agrees with the vague character of περὶ ἐπιχρείας τοῦ μετάλλου. A. Bülow-Jacobsen suggests another interpretation to me: correct to δεδηλωκώς (although this would be the only interchange of ο/ω in the text) and understand: “having set forth for you [in another document] the complement of the detachment, I inform you about …”

4.

βήξιλλον. Vexillum has two meanings: flag or detachment (it is then a synonym for vexillatio). One could be tempted to understand it here as a synonym for vexillarius (as one can say in French “enseign” for “enseign carrier”), but there is no parallel for this. The word recurs in inv. 7316, in an equally unclear context (see footnote 27). The noun ἐπιχρεία is attested only in the papyri. Its oldest occurrences come from Mons Claudianus: here, in O.Claud. II 376 and IV 756, where it clearly has the meaning of “equipment, gear” (ἀπογόμωσις τῆς ἐν τῇ ἁμάξῃ ἐπιχρείας), which it also has in P.Lond. III 948 (236), P.Köln I 52.13, 61 (263), P.Oxy. XXXIV 2718.11 (458), and probably still in P.Lond. IV 1347.3 (710) (contra WB s.v.). In O.Claud. 376, one might be in doubt between this concrete sense and that of “need,” which it has in SB VI 9471.4, and which makes it an equivalent to the simple χρεία (παράσχες … εἰς ἐπιχρ[ε]ίαν ζυγοῦ κάρνου … ξύλον ἰτόεινον). Here, ἐπιχρεία probably refers to supplies, the detail of which is given in the second report.

5.

πραιτέριτος is used here the same way as in Ch.L.A. XVIII 662.5–6 (Chapter 21) and in P.Panop. Beatty 2.199: ὑπὲρ πραιτερίτου δωνατίου (“on account of arrears of the donativum”); more often, this adjective is used in reference to soldiers who suffered a delay in the benefits they were entitled to: P.Panop.Beatty 2.112; W.Chr. 424.16 (στρατιώτας οὐκ ἐφ’ ὀλίγῳ χρό[ν]ῳ ἐπὶ ταῖς χρεωστουμέναις ἀν⟨ν⟩ώναις πρετερίτους); P.Michael. 29.5, 15.

6.

ὑγρῶν, the “liquid provisions,” as opposed to σείτου (cf. P.Oxy. XXVII 2474: ἐκφορίοις ὑγρῶν τε καὶ ξηρῶν, “rents in liquid and dry goods”). Probably this refers to oil and wine.

11.

ἀποκατα⟨στα⟩θέντος is a better correction than ἀποκατα⟨τε⟩θέντος: ἀποκατατίθημι is known only from two occurrences in Apollonios of Rhodes, where it expresses the act of putting down an object that a person was carrying on himself (chest, helmet); ἀποκαθίστημι is in contrast common in the papyri, where it usually means “restore, pay something owed”; among the definitions listed by the WB, it is einen Vormund bestellen  that comes closest to the meaning required here by the context (the example cited in support is M.Chr. 88, ii.18: ἵνα τοῖς παιδίοις δύο ἐπίτροποι ἀποκατασταθῶσι, “so that two tutors may be appointed for the children”). I think that the definition provided by Preisigke is too restrictive and that in the case of

Vibius Alexandros

125

M.Chr. 88, as in that of the ostracon, ἀποκαθίστημι is simply equivalent to καθίστημι (appoint to a function). 13.

We expect φανερόν σοι [ποι]ῶ, κύριε, ⟨ὅτι⟩ καταβαίνων ὁ Σιλβανὸς κτλ.

14.

[3–4 δέ]δωκε. [ἀποδέ]δωκε, [παραδέ]δωκε?

14–15. It is difficult to see why the quantity of barley has not been written right after the word κρειθῆς, but it seems to appear in the next line; in fact, there is no room in the lacuna for ἀχύρου (the third member of a classic triad) and the statement of a quantity. It may be noted that σείτου is aligned under κρειθῆς. 17.

]ειονα ἄκμονα.]ειονα can in principle be only a comparative or a noun. On the basis of study of the original, the traces before ειονα might correspond only to one of the following letters: κ, ρ, σ, φ, χ. If we are dealing with a comparative, βρα]χ̣{ε}ίονα comes to mind, but it can be excluded for semantic reasons (βραχύς is not used in the papyri to refer to the short height of an object) as well as morphological (the comparative used in the papyri is βραχύτερος); π]ρ̣{ε}ίονα, “a saw,” fits better in the context.

19.

Here begins a new series of complaints filling out the preceding list, but reiterating insistently the single donkey and the eight old goatskins.

6 Procurator Montis In 219, a beneficiarius restored a small sacred complex at Koptos, consisting of a chapel and a garden; this pious initiative is commemorated in two parallel inscriptions, I.Portes 86 and 87, the second of which is badly damaged. The first eight lines of I.Portes 86 run as follows: [ἐπ]ὶ̣ τοῖς εὐτυχεστάτοις καιρ̣[οῖς] | τοῦ κυρίου ἡμῶν Αὐτοκράτορο[ς] | Μάρκου Αὐρηλίου Ἀντωνείνου | Εὐτυχοῦς Εὐσεβοῦς Σεβαστοῦ | (ἔτους) β// Μεσορη κ̅, ἐπὶ | Γεμινίῳ Χρήστῳ ἐπάρχῳ Αἰγύπτου | καὶ Οὐαλερίου Ἀπολιναρίου ἐπιτρό|που Ὄρους κτλ. The date is thus expressed not only by the regnal year (Mesore 20 of year 2 of Elagabalus, or 13 August 219), but also by mention of the prefect of Egypt currently in office and another high official named Valerius Apolinaris, whose title has not drawn adequate attention until now: ἐπίτροπος Ὄρους. It occurs only here.1 Although this individual has never been included in the lists drawn up for the prefects of Berenike,2 he has sometimes been regarded as such. Lesquier correctly translated his title procurator Montis Berenicidis,3 but thought that this referred to a procurator metallorum whose competence had been reduced, at this late date, to the mines and quarries of the desert of Berenike.4 On the other hand, although he refers to Lesquier, André Bernand translates the Greek title as “préfet de la montagne,” probably under the influence of the parallel provided by the Tariff of Koptos (I.Portes 67), which he refers to, and where the prefect of Berenike (ἔπαρχος Ὄρους Βερενείκης) appears, like our ἐπίτροπος Ὄρους, as a direct subordinate of the prefect of Egypt. R. S. Bagnall did not question this interpretation of the title;5 but ἐπίτροπος is not the semantic equivalent of praefectus. In 1998, during the first season in Didymoi, we found a copy (or a draft) of a letter of the local curator addressed to the same Valerius Apolinaris. The name gives the appearance of utter banality, but, curiously, this is not in fact the case in Egypt, as Bagnall remarks (1996: 145, n. 6). The prosopography of 1. It is plausibly restored in I.Portes 87, the parallel inscription, which is damaged and in which the name of the dedicant and that of the procurator Montis present a difference in comparison to I.Portes 86: Valerius Apolinaris becomes ] Claudius Apol[, and the dedicant, Marcus Aurelius Apollonis, becomes Marcus Aurelius Se[- - -]nis (cf. Bagnall 1996: 150, n. 23). 2. The last that had appeared when I prepared the first version of this chapter was in Devijver 1974: 464, n. 79. See now the prosopography in appendix to Chapter 3 of this volume. 3. Lesquier 1918: 153, n. 2. 4. As opposed to the ἐπίτροπος πάντων τῶν μετάλλων τῆς Αἰγύπτου of AD 11 (I.Pan 51); Lesquier 1918: 240 and n. 4. 5. Bagnall 1996: 145.

127

128

Rome in Egypt’s Eastern Desert

the Roman army of Egypt compiled by Cavenaile and Criniti records only one homonym of the procurator, a soldier of Upper Egypt, a standard-bearer of the cohors I Lusitanorum, who left a proskynema (undated) at Talmis (SB I 4566).6 Finally, among all of the unpublished ostraca from the Eastern Desert, I note only one Valerius Apolinaris, probably a soldier, in the garrison of Mons Claudianus in the years 140–145.7 In the prescript of the ostracon, Valerius Apolinaris is addressed as ἐπίτροπος Σεβαστοῦ, i.e., procurator Augusti. O.Did. inv. 159 = O.Did. 40 16 × 12 cm Fig. 46

around 219

Figure 46. O.Did. 40. © A. Bülow-Jacobsen

6. Corrected by Speidel 1988: 792. 7. His is the name appearing in the amphora dipinto O.Claud. inv. 6860, and he writes in very bad Greek the letter O.Claud. inv. 6502; the marks of punctuation that he puts between words show that he was a Latin speaker.

Procurator Montis

129

Didymoi is the Didyme of the Antonine Itinerary, the third station of the road from Koptos to Berenike, after Matula and Laqita (Phoinikon); cf. Chapter 1, p. 48. Didymoi is 67 km from Koptos. The ostracon was found in a floor of room 43, one of those that abut the south wall of the fort; this floor, US 14302, which is bounded on the west by a wall of the original barracks, is a layer of sebakh covered by a paving that forms the last floor of the room; it corresponds to a phase of reconstruction and reoccupation of the military buildings, a phase that one would be tempted to date, after a detailed study of the archaeological material, to the third century (information communicated by the excavator, M. Reddé). Οὐαλερίῳ

4

8

12

Ἀπολιναρίῳ ἐπιτρόπῳ Σεβαστοῦ. παρὰ Ἰσιδώρου κουράτορος Διδύμων. δηλῶ σοι, κύριέ μου, κατέβην ἐν Κόπτῳ Μεχειρ ὀγδόῃ καὶ εἰ⟨κά⟩δι· περὶ τῶν τριῶν κάδων ὧν ἔδωκα ἀλλαγῆναι ἀνήνεγκα αὐτὰ ὅσα ἐποίησαι ὁ Πετρώνιος ὁ Βάρβαρο[ς· μ]εμαρτυροποίημαι τὸν π̣[ 2–3 ]πα ἐνώπιον τῶν [ c. 11 ]ι̣ω̣των [ c. 15 ]α – – – – – – – – –

8 l. ἐποίησε “To Valerius Apolinaris, procurator Augusti, from Isidoros, curator of Didymoi. I inform you, sir, that I went down to Koptos on the 28th of Mecheir. Concerning the three pots that I gave to be exchanged, I brought back all of those that Petronius the Barbarian had made. I called as witness the … in front of the soldiers (?)…” 4.

δηλῶ. δ ex μ corr.

6.

On the κάδος, a pot for the water wheel, see Bonneau 1993: 114. These containers, which wore out rapidly, were of ceramic or bronze. Our letter shows that the water-lifting machines in the praesidia were under the responsibility of the curator praesidii. Among all the ostraca of the Eastern Desert that we have found, only a few documents from Mons Claudianus mention κάδοι; they confirm the impression that these objects were subjected to minute administrative control. This is particularly the case in SB XXVIII 16941 (189), a letter of the vice-curator of Claudianus (ἀντικουράτωρ) to the procurator Tertullus, to inform him of the scarcity of food and supplies in which the curator left him when he departed; in the list figure two saqiya pots (Chapter 5); O.Claud. inv. 5395 is a mutilated letter in which it is a matter of business concerning pots (one of which has been sent to “Egypt”), in which the procurator (ὁ ἐπίτροπος, without specification) is involved; O.Claud. ΙΙ 385 is a letter by means of which a curator (perhaps that of the satellite praesidium of Raïma) asks Claudius Alexandros (perhaps a centurion), who is located at Claudianus, to send him some pots, because the two sent earlier had arrived broken.

130

Rome in Egypt’s Eastern Desert

8–9.

ὁ Πετρώνιος ὁ Βάρβαρο[ς. Mayser cites no instance of a personal name with a noun in apposition in which each is preceded by an article, only divine names, e.g., τῆι Ἴσιδι τῆι κυρίαι (Grammatik II.2, 5, ll. 10 ff. and p. 110, ll. 41–45). This emphatic construction may betray the vehemence of the curator, who appears to be seeking to exonerate himself from any accusation. The Roman name of this Barbaros is noteworthy; but this description may be only a nickname without reference to an ethnic origin. Against what I wrote in the first version of this chapter, this ostracon is not contemporary with the period (around 264) when the Barbarians came to the praesidia of the Berenike road to get supplies for themselves.

10.

τὸν π̣[ or τὸν τ̣[. The most suitable restoration in this military context is τὸν π̣[ρίγκ]ι̣πα. On the rare mentions of this rank in the Eastern Desert, see Chapter 28, pp. 445 f.

11.

σ]τ[ρα]τ̣ι̣ω̣τῶν?

Until the time this ostracon was found and inspired me with the first version of this paper, the scraps of official correspondence that we had discovered in the praesidia of the Desert of Berenike showed that information and orders often circulated directly between the prefect of Berenike and the commanders of the forts (curatores praesidiorum). It fitted with the full form of this prefect’s title, “prefect of the military posts and of the desert of Berenike” (praefectus praesidiorum et Montis Beronices), as it runs in ILS 2699. Hence, the title ἐπίτροπος bestowed in I.Portes 86 and in O.Did. 40 on Valerius Apolinaris led me to formulate the following hypothesis: I remarked that the ἐπίτροπος Ὄρους appears at the very moment when the prefect of Berenike, and perhaps also the epistrategos of the Thebaid, disappear from our sources:8 I.Portes 86 dates from 219, while the last attestation of a prefect of Berenike comes on 25 December 216 (P.Turner 34.2),9 and the last dated epistrategos of the Thebaid has a terminus post quem of 216/7 (P.Erl. 19).10 It is thus by no means impossible that the unprecedented (and perhaps short-lived) office of procurator Montis was created between 217 and 219 to replace those of the epistrategos of the Thebaid and the prefect of Berenike, which were so closely connected that they were sometimes held by the same person, whether in succession11 or at the same time: Vettius Gallianus, the praefectus Montis of 216, was in effect at the same time “acting epistrategos” (P.Turner 34), and he wound up becoming epistrategos, to judge from P.Erl. 19, where he is called “ex-epistrategos.” The new title of procurator Montis thus drew its administrative status from the epistrategia of the Thebaid (it was a procuratorship), and its geographic definition from the prefecture: Montis (Vettius Gallianus is called simply ἔπαρχος Ὄρους, but this abridged title is also found since the reign of Trajan in several unpublished ostraca from the desert of Berenike). We may add that if the epistrategos of the Thebaid had still existed at the moment when the beneficiarius 8. Cf. Thomas 1982: 65–66 and 191: although the last known epistrategos of the Thebaid is no later than 216/7, the last epistrategoi of the Heptanomia date to 280–282, and the last reference to the ἐπιστρατηγία of the Heptanomia to 289; Thomas therefore does not exclude the possibility that the office of epistrategos of the Thebaid continued in parallel (ibid., 66); it is only in 298 that it is possible to state with confidence that the position no longer existed: the Thebaid had by then become a wholly separate province, governed by a praeses and divided into Ἄνω and Κάτω Θηβαίς, each of these districts being under the authority of an ἐπίτροπος. 9. [Οὐετ]τίωι Γαλιανῶι [τ]ῷ κρατίστωι ἐπάρχωι Ὄρους διαδεχο(μένῳ) καὶ τὴν ἐπιστρ(ατηγίαν). 10. Thomas 1982: 65. P.Erl. 19.9–10: ] Σεπτιμίου Πατροφίλου τοῦ κρα(τίστου) ἐπιστρ̣(ατήγου) [ | ] Οὐεττίου Γ̣αλλιανοῦ τοῦ ἐπιστρα(τηγήσαντος) [ 11. Artorius Priscillus was prefect of Berenike in 109 (according to several ostraca of Krokodilo) and epistrategos of the Thebaid around 115. Since the Ptolemaic period, the involvement of the epistrategos of the Thebaid in Red Sea commerce is made evident by the combination of his office with a second one called ἐπὶ τῆς Ἐρυθρᾶς καὶ Ἰνδικῆς θαλάσσης.

Procurator Montis

131

dedicated his painted garden, it is the epistrategos who would normally have to have been mentioned after the prefect of Egypt, as we find in some other inscriptions that come from cities in Upper Egypt, including Koptos;12 if the prefect of Berenike rather than the epistrategos of the Thebaid acts in the Tariff of Koptos, it is first because the mentions of the prefect of Egypt and the prefect of the Desert do not, in this document, serve the purpose of dating it, and then because this inscription was set up at a customs barrier, outside of the city and already in the territory that belonged to the competence of the ἔπαρχος Ὄρους. These administrative changes fit into the troubled period beginning with the assassination of Caracalla and continuing until the first years of Elagabalus’s reign. This is what I wrote in 2000. However, further finds of ostraca in Dios and in Xeron Pelagos have shown the weakness of this hypothesis. Two crucial facts emerged. First, it is not necessary that the official in command of the desert should be termed ἐπίτροπος Ὄρους to be addressed as ἐπίτροπος Σεβαστοῦ: the tabulated prosopography of the prefects of Berenike which I annexed to Chapter 3 reveals that, from the reign of Hadrian on, there is a tendency among the authors of reports or petitions to address these prefects as procurator Augusti (ἐπίτροπος Σεβαστοῦ).13 Second, in the report of a nighttime incident sent to him by the curator of Xeron Pelagos, the same Valerius Apolinaris, ἐπίτροπος Ὄρους in I.Portes 86, is called ἔπαρχος Ὄρου⟨ς⟩, just as were his predecessors (O.Xer. inv. 731). The conclusion, of which I already had a glimpse in a 2007 study on Sulpicius Serenus (Chapter 3), is inescapable. The prefectura of the desert, which modern historians, after Lesquier, tend to consider a militia, was, from the point of view of the Roman administration, a procuratorial post, probably at least since the reign of Claudius, which Sherwin-White has shown to be a turning point in the organization of the “procuratorial system.”14 It is interesting to browse the prosopography of the prefects of Berenike (Chapter 3, pp. 105–9) in the light of Sherwin-White’s article. Prefects nos. 1–2 date to the reigns of Augustus and Tiberius, when a territorial praefectura was an “extension of (…) military duties (…) and are not to be reckoned as part of a systematised procuratorial career.”15 Both are tribuni of the legio III Cyrenaica, which was probably based then in Thebaid. For Pinarius Natta, whose cursus is known, this was his first military post. Artorius Priscillus (no. 9, Trajan) is the next prefect of Berenike whose cursus is preserved. By then the procuratorial system is in place: the praefectura of the desert is a procuratorial post which one enters after the three militiae. Artorius Priscillus does not cumulate his praefectura with that of the ala Vocontiorum, which was probably stationed then in Koptos.16 From Claudius on, Sherwin-White remarks, praefectus disappears as the title of governors of small provinces, and is replaced by procurator, which, in the preceding reigns, was restricted to financial posts. Praefectus is kept however for territorial prefects, whose jurisdictions are not provinces, but “special areas within the regular provinces.”17 This is the case of the prefect of Berenike. Under Hadrian, Sulpicius Serenus is the first prefect of Berenike to whom the title ἐπίτροπος is undoubtedly bestowed (in the form of address ἐπίτροπος Σεβαστοῦ). But he is also probably the first prefect of Berenike who cumulates this procuratorship with the command of the ala in Koptos. Sherwin-White does not mention such an evolution, which reminds one of the situation under Augustus and Tiberius (an “extension of (…) military duties 12. At Koptos: I.Portes 70 (103); I.Portes 82 (210): just nine years before the mention of the procurator Montis. 13. There are four occurrences: O.Dios inv. 90; O.Did. 27; O.Dios inv. 1460; O.Did. 40. 14. A. N. Sherwin-White, “Procurator Augusti,” Papers of the British School at Rome 15 (1939) 11-26. Cf. also S. Demougin, “À propos d’un préfet de Commagène,” ZPE 43 (1981) 97-109, at 98, quoting Sherwin-White, Pflaum, and A. H. M. Jones, who have insisted “on the introduction, under Claudius, of the use of the word procurator to refer to an equestrian agent of the provincial administration, whatever his competences were.” 15. Sherwin-White 1939: 13. 16. At least horsemen from this ala are known in the desert under Trajan (Cuvigny [ed.] 2006: 335 f.). 17. Sherwin-White 1939: 20.

132

Rome in Egypt’s Eastern Desert

(…)”). Actually, this combination is an exception among territorial prefects.18 This model, perhaps inaugurated by Sulpicius Serenus, was maintained until the beginning of the third century (Vettius Gallianus, no. 25). We do not know if Valerius Apolinaris, who directly succeeds him, was at the same time prefect of the ala Herculiana. Valerius Apolinaris is the last precisely dated prefect of Berenike, but it is not possible to say if this post was abolished between his time and the abandonment of the praesidia which occurred c. 270, because too few ostraca remain of this last period of occupation. His title ἐπίτροπος Ὄρους in I.Portes 86 remains unique19 and enigmatic. It points to a contamination between the specific name of the office, ἔπαρχος Ὄρους, and the generic title ἐπίτροπος Σεβαστοῦ, which refers to the status of all the equites having a public office which is not a militia (in its restricted meaning of command of a military unit), and which was often used as a form of address in the prescripts instead of ἔπαρχος Ὄρους. On the one hand, as already remarked, the mention of the epistrategos was expected in this inscription from Koptos. Perhaps, because the prefect of the desert Valerius Apolinaris was presumably only actingepistrategos, and the stone-cutter was obviously running out of room on the stone (see I.Portes, pl. 54), it was decided to coin this curious title which had the advantage of being shorter than διαδεχόμενος τὴν ἐπιστρατηγίαν.

18. J. F. Gilliam, “The dux ripae at Dura”, TAPA 72 (1941) 157–75, at 167: “About none of these praefecti do we know much more than that they existed. Most of them seem to have this in common: they were military officers directly charged with the protection of certain areas which presented unusual problems. Their offices were more territorial and more local in character than that of a commander of an individual unit, who was primarily concerned with his unit and only secondarily with the region in which it was stationed.” 19. I leave aside the doubtful case of ἔπαρχος Ὄρους κ̣α̣ὶ̣ ἐπ̣[ίτ]ρ̣ο̣π̣ο̣ς̣ Ὄ̣ρους referring to Artorius Priscillus (O.Krok. 41.47n. and Chapter 3, pp. 101 f.).

7 Greek ostraca from Mons Claudianus, revisited The ostraca published in 1986 in the French version of this chapter with the title “Nouveaux ostraca grecs du Mons Claudianus” were reprinted in the Sammelbuch (SB XVIII 13335–13361). As several of them were of minor importance, I have preferred to exclude them from this chapter, concentrating instead on what the later excavations of Mons Claudianus and Umm Balad made it possible to add to the contribution of those that I have kept. This update has allowed me to make a number of corrections and to make progress on the question of the architect Hieronymos and the chits of zeuge. Prosopographical connections with ostraca found later have shown that the ostraca republished here are likely to have come from the essentially Trajanic dump of the South Sebakh.

I. The architect Hieronymos SB XVIII 13335. Letter. 10 × 10 cm. There is no hope of identifying the two correspondents in this letter with persons mentioned in other ostraca from Mons Claudianus: their names are too common. On the other hand, the three candidates for a bonus have appeared in the ostraca from the South Sebakh. It is curious that the people in charge of the payment of the bonus have native names, while the recipients have Greek, Greco-Egyptian, or even Latin names.

4

8

Πάχνουμις Πετέχωντι τῶι ἀδελφῶι χαίρειν. ἀνάγκην ἔσχον δηλῶσέ σοι ὅπως μὴ δῦς σπονδὴν τοῖς ὑπογεγραμμένοις· οὔτε γὰρ ἔλαβον χαλκὸν περὶ αὐτῶν οὔτε οἶνον. εἰσὶ δὲ{ν} τὰ ὀνόματ(α)· ῾Ιερ{ι}ώνυμ(ος) Ἱερ{ι̣}ωνύμ(ου) ἀρχιτ(έκτονος)

135

136

Rome in Egypt’s Eastern Desert Mᾶρκος Σωκράτους Ἰσίδωρος Ὡ̣ρίωνος.

3 l. δηλῶσαι 4 δῦς for δοῖς, l. δῷς ἀρχιτ(έκτων) ed. pr.

7 l. ἔστι

ονοματ

8 ϊεριωνυμ ιεριωνυμ αρχιτ

“Pachnoumis to his brother Petechon, greetings. I have to let you know not to give a bonus to the persons mentioned below, for I have not received either money or wine for them. Here are their names: Hieronymos son of the architect Hieronymos; Marcus son of Sokrates; Isidoros son of Horion.” 3.

ἀνάγκην ἔχειν, an equivalent of the classical ἀνάγκη εἶναί τινι, is well attested in papyrus letters between the first and seventh centuries; P.Mich. VIII 490.6–7 (second century) offers a parallel to the epistolary formula ἀνάγκην ἔσχον σοι δηλῶσαι.

4.

Surtax or bonus, the σπονδή was traditionally paid in wine or in cash (see S. Eitrem, Symb. Osl. 17 [1937] 27–41): here, both means of payment are intended.

8.

ἀρχιτ(έκτονος). On the resolution in the genitive, see infra.

9.

Mᾶρκος Σωκράτους. This person is attested several times in the South Sebakh. His name appears on several amphoras and on a chit of zeuge of bread (inv. 1980).

Hieronymos Father and Son Jean Bingen devoted a section to “Hierônymos, fils de Hierônymos” in his chapter “Les architektônes du Mons Claudianus” in O.Claud. I. He had an inkling that the question of the architect Hieronymos was tangled (O.Claud. I 26, introduction and comm. ad 1). The ostraca discovered at Umm Balad in 2002–2003 momentarily thickened the mystery. Under Trajan, three architects (the word is to be understood as meaning “engineers”) worked at Mons Claudianus: Hieronymos, Herakleides, and the Alexandrian Apollonios son of Ammonios, whose patronymic is known from I.Pan 38, the dedication of an altar that he erected for Zeus Helios Great Sarapis. Hieronymos is not a very common name. It is therefore likely enough that the architect Hieronymos who lived at Umm Balad1 during the first phase of the occupation of this site was the same man and that after the early abandonment of Umm Balad, he came to Claudianus. The patronymic of the Hieronymos of Umm Balad is never mentioned, but this Hieronymos receives letters from another architect, Apollοnios, who sometimes calls him τῷ φιλτάτῳ, sometimes τῷ υἱῷ (O.KaLa. inv. 650). Conversely, in a letter that he did not send, Hieronymos calls Apollonios τῷ π[ατρί] (O.KaLa. inv. 442). Although these address formulas may refer to a simple difference in age, and not necessarily biological relationship, the letter O.KaLa. inv. 691, addressed by a centurion to Hieronymos, suggests that the latter was in fact the son of Apollonios. This centurion writes: ἔδοξέν σο̣[υ]|[τῷ πα-]

1. On the metallon of Umm Balad, see Chapter 1, pp. 20 f. This site, founded during the reign of Domitian under the prefect Mettius Rufus (88–92) and abandoned under Trajan, was first called either Domitiane or Kaine Latomia, then, after the damnatio memoriae of Domitian, only Kaine Latomia. I realized only recently that Kaine Latomia and Domitiane were not exclusive, but represented two abbreviations of the original name.

Greek ostraca from Mons Claudianus, revisited

137

τ̣ρὶ ⸌Ἀπολλωνίωι⸍ ἀπομερίσαι σ⟦ε⟧⸌οι⸍ δύο ἀνθρ[ώ|πους εἰς τὸ ἔργ?]ον τοῦ λάκκου Πράσου, “your father Apollonios has decided to give you two men for the work on the cistern of Prasou.”2 As for the architect Apollonios, in the ostraca from Umm Balad, he is likely to be the Apollonios son of Ammonios of Mons Claudianus. We do not know for certain from where Apollonios wrote to Hieronymos when the latter was living at Umm Balad in the reign of Domitian or at the start of the reign of Trajan: Was Apollonios already at Claudianus (mentioned in one of his letters, but in a fragmentary context)? A collective letter of the sklerourgoi of Porphyrites, addressed to Maximus, a close associate of Hieronymos, rather gives the impression that Apollonios was working at Porphyrites, where his son Hieronymos was also known. Without naming Apollonios, the sklerourgoi write: οὐκ ἔχωμεν (l. ἔχoμεν) ἀηδίαν πρὸς τὸν πατέρα αὐτοῦ ἢ καὶ πρὸς Ἱερώνυμον, “we have no complaint against Hieronymos or against his father” (O.KaLa. inv. 394.5–7). Apollonios, the senior architect, may have for some time been engaged in organizing, whether directly or through the intermediary of his son Hieronymos based at Umm Balad, the works of extraction and of the construction of infrastructure in the entire zone of Claudianus and Porphyrites. The resolution ἀρχιτ(έκτων) in SB XVIII 13335.8 acted for a long time as an epistemological brake on understanding this scenario. But if we connect the description as architect not to the person named but to his father, the situation becomes much clearer. The architect Hieronymos son of Apollonios had a son with the same name as himself, whom he had brought with him to Mons Claudianus. And in fact we find a Hieronymos son of Hieronymos and a Clemens son of Hieronymos mentioned together in several lists of names.3 In the money account O.Claud. inv. 1533, the two brothers appear in the sublist headed παῖδες, where their names are, moreover, immediately preceded by that of Isidoros son of Horion, who is precisely the third recipient mentioned in SB XVIII 13335. In O.Claud. inv. 1533, the παῖδες seem to receive 2 obols, while those called ἄνδρες get 1 drachma 2 obols. But in O.Claud. I 26, a list of three ἄνδρες, we find the sons of Hieronymos, Clemens and Hieronymos, with their companion Isidoros son of Horion! Is this because in the meantime these paides had reached adulthood? Around 152/3, under Antoninus, a Hieronymos is often mentioned in the dossier of the tabularius Athenodoros, a part of which was published in O.Claud. IV. It is never specified that this Antonine Hieronymos is an architect, but he is clearly a respected individual, who controls the major equipment used for work in the quarries (cf., e.g., O.Claud. IV 873; 882; 891; 894). Even if we have some cases of long service in the workforce at Mons Claudianus, as I show below, it seems to me difficult to suppose that the elder Hieronymos, who around 110 was the father of two sons old enough to be apprentices, was still in post forty years later. The Antonine Hieronymos is perhaps the παῖς Hieronymos, who in his turn became an architect like his father Hieronymos and his grandfather Apollonios.

II. The zeuge of bread These little chits belong to a series that swelled over the course of the excavation seasons and now amounts to a total of 336 items. Fourteen of these documents were also found at the metallon of Umm Balad. The format is uniform: a name (occasionally two) in the nominative or genitive, then a number of ζεύγη (“pairs”): normally 55 in the Trajanic South Sebakh, 50 in the Antonine dumps and at Umm Balad. The names are those of native workers (quarrymen and smiths). The pairs are pairs of loaves of bread, although ἄρτων is rarely made specific. 2. A praesidium on the road of Porphyrites, perhaps Badiya. 3. These sons of Hieronymos do not appear in the ostraca of Umm Balad.

138

Rome in Egypt’s Eastern Desert

Pairs of loaves of bread are well attested in the Greek papyri of Egypt at all periods, particularly in the context of the distribution of salary or of allowances.4 I do not know what physical form a pair of loaves took: were they simply two separate loaves, or two attached loaves, or a single large loaf?5 These two loaves (or this double loaf) probably corresponds to the two meals of the day, the ἄριστον and the δεῖπνον.6 The loaves of the native workforce of Mons Claudianus were made by the women of the worker’s family, who lived at Kaine, using the wheat from the monthly ration of 1 artaba.7 This system, which we understand thanks to the Antonine-period entolai,8 seems apparently to have been in operation already in the reign of Trajan (even though there are no entolai under Trajan). The preparation of bread in house made it possible to avoid carrying vast amounts of fuel to Claudianus. We may also suppose that the workers were more confident in their wives not to abstract secretly a portion of their wheat ration. An Antonine ostracon shows, indeed, that the recipients of bread checked that the number of loaves distributed to them did in fact correspond to the amount of wheat owed them.9 What exactly is the nature of these documents? The first thought that comes to mind is that they are labels allowing identification of the bread that the native workers received each month. But these tiny ostraca offer no means (a notch or hole) for attaching them to a sack or basket. Were they placed inside the container? That is the conclusion I reached earlier, but if that is right, they could not have identified the owner. I think rather that they allowed the worker who received them to verify the contents of a large sack (itself suitably labeled with his name to be easily identified): the number of loaves, but also of other objects, usually baskets (see below SB XVIII 13350, introduction). The idea of a large sack came from one of these chits, which exceptionally accompanied the bread of two soldiers and offered a distinctive format.10 After their names, we read: ἐν ἑνὶ σάκκ(ῳ) ζεύγη ξ ἄρτων, “in 1 sack, 60 pairs of loaves.” This hypothesis has the virtue of taking account of some of these chits that although complete do not give a name, but only a number of zeuge. Other questions remain unsettled. Why 55 or 50 pairs for 30 days, and not 30 as for the soldiers whom I have just mentioned? Why was the artaba yielding 55 pairs under Trajan, but only 50 under Antoninus? Had the volume of the artaba shrunk a bit? Or had the loaves become a little larger?11 It would be useful to discuss these questions in the context of the entirety of these chits and taking into consideration all of the cases, at least in the Eastern Desert, especially those in which a number of pairs or of loaves is mentioned in connection with a volume of grain expressed in artabas or in matia. SB XVIII 13340 and 13341. 5 × 4.5 cm and 6 × 5.5 cm. The texts, their layout, and the hand are identical. Apollonios son of Dionysios has not appeared elsewhere in O.Claud. 4. I know of no example of this way of counting bread outside the Greek documentation from Egypt. 5. See O.Krok. I 16, introduction. In the same spirit, at my baker “une demi-tourte” is the name of a round loaf which weighs half as much as a large loaf called “tourte.” 6. Michel 1997: 109: “On prenait partout deux repas, le matin avant d’aller au travail, plus ou moins consistant selon le labeur envisagé, et le soir le repas principal.” 7. See Chapter 14. 8. See Chapters 9 and 10. 9. O.Claud. inv. 7298. The authors of this letter, sent on mission to the praesidium of Raïma, received a half-artaba of bread that contained only 48 loaves: in other words, 24 zeuge, even though at this period the artaba translated to 50 zeuge. 10. Only two chits for zeuge are in the names of soldiers. The ration was then 30 per person each time, not 55 or 50 zeuge. 11. In WO I, pp. 755–57, Wilcken already discussed the relationship between the artaba and zeuge of bread and observed that this was extremely variable.

Greek ostraca from Mons Claudianus, revisited

139

Ἀπολλων(ίου) Διονυσί(ου) νε SB XVIII 13342. 7 × 4 cm. This Harpochras without a patronymic could not be identified with any of the numerous men with that name known from the South Sebakh. Ἁρποχρᾶτος ζεύγη νε SB XVIII 13343. 6 × 5.5 cm. The fact that a name as commonplace as Ammonios is not followed by a patronymic confirms the hypothesis offered above, according to which the chits were not labels serving to identify a sack, but rather statements of the contents. Ἀμμωνί(ου) ζεύγη ν̅ε̅ SB XVIII 13344. 2.5 × 6.5 cm. The South Sebakh has produced two further chits with the name Aithiopas. In one of them, the name is followed by the patronymic Herakleides (inv. 1974). Αἰθιοπᾶτο(ς) ζεύγη ν̅ε̅ SB XVIII 13346. 7 × 5 cm. Ἀπολλων(ίου) Ὅνου ζεύγη νε 2.

Ὀνοῦ ed. pr. The most likely interpretation is that this is an Egyptian name Ὁνης, which renders Ḥwn (Dem. NB, 778). But the genitive of this name is normally Ὁνέους or Ὁνῆτος. Apollonios son of Hones is not elsewhere attested at Claudianus.

140

Rome in Egypt’s Eastern Desert SB XVIII 13347. 5 × 9 cm. Διοσκ(ορᾶτος) Βοθ̣ύν(ου) ζεύγ(η) νε

1 Διοσκ( ) ed. pr., Διοσκ ostr., Βοθυν 2.

2 ζευγ

This is Dioskoras son of Bothynos, who is attested in the order for payment O.Claud. inv. 646 along with Protarchos, the latter probably being named in SB XVIII 13349 (infra). Inv. 646 begins with δὸς σκλ(ηρουργοῖς), followed by the names of the recipients in the dative, which confirms the occupation of the persons whose names appear on the chits of zeuge. The name Bothynos, which reappears in two other ostraca from the South Sebakh (with the same patronymic in each case?), is a very rare nickname (ὁ βόθυνος, “the hole”). I have found only one other attestation, in an epigraphic will from Kos: [τὸ]ν̣ λεγόμενον Β̣όθυνον τὸν πατρῷον μου δ̣[οῦλον, “my slave, inherited from my father, nicknamed Bothynos” (M. Segre, Iscrizioni di Cos I [Rome 1993] no. ED 200.19 [first century BC]) SB XVIII 13348. 8.5 × 6 cm Ἱερακ( ) Φακᾶς ζεύγ(η) νε

1 ιερακ 2.

2 ]υγ Ἱερακ( ) Φακᾶς. Both names Hierax and Hierakion are well attested in the Trajanic ostraca. Hence Ἱερακ(ίων) Φακᾶς, the latter being a surname, unless Φακᾶς is a mistake for Φακᾶ (genitive), which opens the possibility that it is the patronymic of Hierax or Hierakion. This man has not appeared elsewhere in O.Claud. The nickname Φακᾶς, which refers to marks on the skin, is not common. In all of the ostraca of the praesidia of the Eastern Desert, I know of only one other attestation, an amphora inscription O.Claud. inv. 7914 (Κανω[π--]|Φακᾶ[). In the valley, the only other attestion is P.Thmouis 1, col. 146.13, where Phakas is the name of a slave. SB XVIII 13349. 5 × 4 cm [Πρώ]ταρχος [ζε]ύγ(η) νε

1 ]ταρχος ed. pr. 1.

2 ]υγ

[Πρώ]ταρχος also figures on another chit for 55 zeuge (inv. 4107). Since this name is not common at Mons Claudianus, he is probably to be identified with the sklerourgos who is the beneficiary of the order for payment inv. 646 mentioned earlier. This may be the same man as Protarchos son of Horion, in an amphora inscription from the South Sebakh.

Greek ostraca from Mons Claudianus, revisited

141

SB XVIII 13350. 5 × 4 cm. In this chit, numbers are given for containers, sacks and basket(s), as well as for the zeuge. There are several other examples of this phenomenon, all from the reign of Trajan, in the corpus of chits of zeuge: under Antoninus, the method of provisioning must have changed, and the sakkoi contained only loaves. The containers most frequently mentioned are σφυρίδια (baskets, 39 occurrences); less often, we find κόϊκες (baskets made of doum-palm fiber), μαρσίππια (sacks), χιλώματα (haversacks, cf. Chapter 39). One might think that these containers were used to hold the pairs of loaves, but since on a few occasions these words are preceded by καί, that hypothesis cannot stand. Their content is not stated, except on two chits, where onions and arakos are mentioned. Κάστωρ Πεβτ(ος) ζεύγ(η) νε, σφ̣υρί̣δ̣(ια) γ̣̅, μαρσιπ̣[] 1 l. Πεβῶτ(ος)

2 ζευγ

σφυριδ

“Kastor son of Pebos, 55 pairs, 3 baskets, x sack(s).” SB XVIII 13351. 6.5 × 5 cm. The two men in this chit presumably belonged to the same family. They cannot be identified with any homonyms in the Trajanic ostraca from Claudianus. [Ἑ]ρμογένης καὶ Ἥρων ζεύ(γη) ρι “Hermogenes and Heron, 110 pairs.” SB XVIII 13352. 5 × 4 cm. Μαῶς ζεύ(γη) ν⟦ε⟧δ 1.

The Egyptian name Maos is not found in other texts from Mons Claudianus.

2.

The number of pairs, 55, has been corrected to 54. If for some reason the expected number of loaves had not been put into the sack at Kaine, it was important for the label to specify this so that the worker would not feel cheated.

142

Rome in Egypt’s Eastern Desert

III. The other texts SB XVIII 13353. Inventory. 8.5 × 8 cm. List of foodstuffs and domestic utensils, which make up an assemblage that calls to mind P.Oxy. XXIV 2424 (II–III), an inventory in which the following items (along with others) appear: ἄρτοι,12 σφυρίς, σφυρίδιον, κεράμια, βαυκάλια, κερβικάριον, and ξύστραι. In the ostracon, the last two items are preceded by the remark καὶ παρὰ Διονύσιν. Had these objects been entrusted to Dionysios on deposit (παραθήκη)? The names of the objects are sometimes in the nominative, sometimes in the genitive. The helmet and the shield indicate that the owner was a soldier.

4

8

1 μτα

ζεγυ

ἄρτων μάτ(ια) ϛ ζεύγ(η) γ, ξύστρα, κεράμιν κριθῆς α (ἥμισυ), ἄλλο κεράμιν κλασμάτ(ων), πέτασον, σφυρίδιν, κερβικάριν, κάση παλαιά, θύσκαι δύω, βαυκάλις δύω. καὶ παρὰ Διονύσιν· ἀσπίδα, ἄρτων σιρός. 2 l. κεράμια

3  κλασματ

6–7 l. δύο

l. βαυκάλεις

“Loaves: 6 matia, 3 pairs, a scraper, 1.5 amphora of barley, another amphora with pieces of bread, a petasos, a small basket, a pillow, an old helmet, two incense-burners, two water jugs. “And with Dionysios: a shield, a jar of loaves.” 3.

The first meaning of κλάσμα is “piece, fragment.” In papyrology, it appears only in accounts of the sixth to eighth centuries, where it can refer to the act of giving the details of payments to be made, but can also refer to a tax (see F. Morelli, CPR XXII 26.4n. and 5n.; Delattre and Fournet 2014: 218 ff.). But in the present instance, it probably refers to “pieces of bread”; this is how the word is used in the Gospel of Mark, to refer to the pieces left over from the multiplication of loaves (ἧραν κλασμάτων δώδεκα κοφίνους πλήρεις, 6:43); κλάσμα is regularly used to refer to eucharistic bread (Lampe, s.v.), but an epigram of Lucillius, a contemporary of Nero, uses the word in a pagan context: ἂν δὲ παραρπάζῃς ἄρτους καὶ κλάσματ’ ἀναιδῶς, “but if you cheekily pinch my pieces of bread,” Anth. Pal. 11.153. In present-day Egypt, loaves or pieces of bread are crisped or dried in the oven in order to make sure they remain usable for a long time.

4.

The only other papyrological attestation of πέτασος (broad-brimmed hat) is UPZ 158 A.108, and this ostracon remains the only mention in the ostraca of the Eastern Desert.

12. It is tempting to restore the noun ζεύγη in lines 9 and 10 (cf. P.Lond. 190.31–32, where pairs of loaves and of cakes, λαγάνια, are similarly accounted for in succession).

Greek ostraca from Mons Claudianus, revisited

143

5.

κάση. A borrowing from the Latin cassis, known otherwise in the papyri through the diminutive form κασίδιον. The shift of a Latin noun in –is to the Greek first declension is attested in Egypt only for classis: cf. the examples collected by Daris, Lessico latino, s.v. κλάσση; it is also possible that it might belong to the declensional pattern of θάλαττα, since the nominative κλάσση, deduced by Daris, does not appear in the texts, while the accusative form κλάσσαν exists (SB IV 7354.6).

6.

A censer, θύσκη, appears in military baggage in P.Oxy. XIV 1657.13 (end of the third century).

6–7.

βαυκάλιον and βαυκαλίδιον are very well attested in the papyri, but not βαυκάλις, which according to Athenaeus was the Alexandrian name for a certain type of vase (11.784b). On βαυκάλις, see now Fournet 2009: 25–27.

10.

σιρός. This term survives today in the word “silo.” The first thought that comes to mind is that this σιρός is a movable container that has been put on deposit with Dionysios. One thinks in that case of a large jar, an interpretation that seems to find some support in Hesychius, who among the glosses he gives for σιρός includes πίθος. G. Husson (1983: 252 ff.) has collected the papyrological instances of σιρός, a word for which she sticks to the definition given by Columella, who describes a sirus as a ditch dug into the earth, in which cereals were stored. Husson finds confirmation for this description in SB V 7756. 3 and 16 (AD 359), where the mysterious expression γρι καὶ σιρώματος should be the name of the tax to finance the “digging and upkeep of silos” (a bold interpretation by Rémondon, which Jean Gascou does not venture to support: “Nouveautés documentaires et littéraires sur Clysma,” in Brun et al. (eds.) 2018, n. 37). All the same, it remains the case that the σιροί mentioned in the papyri do not seem to be movable objects, but fixed spaces, even if nothing indicates specifically that they were excavated in the ground. The πίθος of Hesychius could, for that matter, be partly buried. Some other occurrences of σιρός have since appeared in the Eastern Desert, where, whenever the contents of the σιρός are known, it is wheat, σῖτος. These texts also do not give the impression that these siroi were transportables. Thus, several late ostraca from Dios are slips with the names of those who stand night guard εἰς τοὺς σιρούς. In O.Claud. inv. 5745, there is mention of the door, or rather the wicket, of the siros (τὸ θύριον τοῦ σιροῦ). These considerations require us to understand the present passage to mean that an unspecified quantity of bread had been deposited in a silo that was located at Dionysios’s house. SB XVIII 13354. 8 × 5 cm.

List of four men, like many others found at Mons Claudianus. It may concern the four occupants of a kella (Chapter 11, p. 195). Aside from Ammonios son of Soros, these individuals appear in other lists of names from the South Sebakh.

4 1–2 πετερμουθ

Πετερμοῦθ(ις) Νίλου Πετερμοῦθ(ις) Σώρου Ἀμμώνιο(ς) Σώρου Ἆπις Ἁρμαχίων(ος) δ̅ 3 αμμωνιο

4 αρμαχιων

144

Rome in Egypt’s Eastern Desert

“Petermouthis son of Nilos; Petermouthis son of Soros; Ammonios son of Soros; Apis son of Harmachion: 4.” 2–3.

Σῶρος is presented in Trismegistos as a variant of the Semitic name Σοῦρος. In the present case, at least, the analysis of the Dem.NB 685 (ns-ḥr, “er gehört dem Horus”) seems preferable to me. The DDbDP offers only two occurrences of Σῶρος in the valley, at Thebes and Elephantine. In O.Claud., Soros is known only as the patronymic of three native workers, Ammonios, Petermouthis, and Sarapion.

4.

Ἁρμαχίων(ος). The suffix -ων for this Egyptian name, which is better known in the form Ἅρμαχις, is attested only at Mons Claudianus, in several Trajanic ostraca. The same individual is called Ἆπις Ἅρμαχις (read probably Ἁρμάχιος) in O.Claud. inv. 1593. SB XVIII 13355. Order or attestation of payment. 7 × 6 cm Φαόνις Παχνούμι[ος ] Σάνσνος Ἀπολλων̣[ίου ] vacat γί(νονται) (δραχμαὶ) δ (τριώβολον)

1–2 Φαόνης ed. pr.

2 l. Σάνσνως

3 γι  

“Favonius son of Pachnoumis […]; Sansnos son of Apollonios […]. Total: 4 drachmas 3 obols.” 1.

Φαόνις, which was read Φαόνης in 1986 because of a small intrusive stroke next to the iota, was not then identified as a Latin name. Favonius son of Pachnoumis is well attested at Mons Claudianus. His name (unexpected with an Egyptian patronymic) is habitually written Φαουόνις or Φαουώνι(ο)ς. The man appears in four ostraca from the South Sebakh (thus presumably Trajanic), but also in trenches where texts dated to years 21 and 22 of Hadrian and the first years of Antoninus have been found (South West Sebakh, West Sebakh: for the dates of these dumps, see O.Claud. III, pp. 110–15). Among these ostraca are five entolai in his name, two of which are in the name both of him and of his brother Pachnoumis. They had another brother, Mikkalos, who has the same stratigraphic profile as Favonius. At the period of the entolai, it was his sister Tbekis who was making the bread. In O.Claud. IV 833.34 (c. 140–145), he appears in a list of workers receiving hardened points. A systematic study of the prosopography of Claudianus, which remains to be undertaken, will make it possible to see if there are many other cases of a comparable longevity among the native workforce.

2.

Sansnos son of Apollonios is attested in four ostraca from the South Sebakh, two of which are lists of four men in which he appears with Favonius. In inv. 1533, he appears among the παῖδες along with Hieronymos son of Hieronymos.

Greek ostraca from Mons Claudianus, revisited

145

SB XVIII 13356. Two names. 4.5 × 3 cm. These two men, probably brothers, are native workmen (pagani), as several chits of zeuge in their names, as well as an amphora dipinto, inv. 2952, in the name of “Phthaus son of Pebos, ergodotes,”13 show. They are mentioned together in a chit, which assigns to them 1 keramion for the pair. The content is not specified, but it could be water: in the organization chart published in Chapter 11, the individual ration of the native workmen is ½ keramion, and I consider in this chapter (p. 212) the possibility that the amphoras of water were distributed to groups of workers. Among the ostraca of the South Sebakh mentioning Phthaus, one is precisely dated to year 13 of Trajan (inv. 1545). Unlike his brother Habos, who is attested only in the South Sebakh, Phthaus, like Favonius, is present in the later dumps. O.Claud. inv. 8394 (West Sebakh) is an entole in his name, in which he is still described as ergodotes. Φθοῦς Πέβως Ἅβως Πέβως 1 l. Φθαῦς

1–2 l. Πέβωτος

1.

Φθοῦς. The normal spelling of this name, at Claudianus as in the Nile valley, is Φθαῦς (Pȝ-ḏw, “der junge Vogel, das junge Tier,” Dem.NB 347, citing the Greek variants Πταῦς, Φθεύς). The vocalization ου is attested only in the compounds Ψενφθοῦς and Φθουσνεύς, but these have other etymologies.

2.

Ἅβως. The name is epichoric in the Theban region (Ḥbȝ, Dem.NB 779), which is not the case for the patronymic, Πέβως (Pȝ-ȝbȝ, “the panther,” Dem.NB 154). SB XVIII 13357. Two names. 5.5 × 4.5 cm. Μαρεις Τρίουμφος

1.

Μαρεις. An Anatolian name; the ending -εις is specific to Cilicia (Bull. 1961, 79). On Cilician names at Mons Claudianus under Trajan, see O.Claud. I 84, comm. ad 6 and 11.

2.

Kajanto 1965: 278, lists four attestations of the cognomen Triumphus, otherwise unknown in Egypt. The same man is mentioned in O.Claud. IV 752.17 (Trajan), interpreted as a list of building workers, none of whom bears an Egyptian name. Since Triumphus appears in a subgroup of names with no patronymic or alias, he may have belonged to the familia. His name is preceded there by the descriptor λι(θοφόρος).

13. Published in Maxfield and Peacock 2006: 180.

8 An inscription of an ἐργοδότης in a quarry at Mons Claudianus A chance walk on 25 January 1991 led me to put my foot right in the middle of the rock inscription published here, thus giving me the opportunity to express to Father Martin, with whom I visited the site before the start of excavations, how precious to me is the memory of our excursions in the Egyptian deserts.1 If the lovers of texts who visited Mons Claudianus ever since the nineteenth century had missed this inscription until I discovered it, the reason is probably that chance did not lead them to it at a moment when the light was right. It is located on the upper face of a bench of granodiorite, 55 m from the little house which, built around a split basin, overhangs the second tributary wadi of the Umm Hussayn to the east of the fortified village (Quarry Peacock 75). It serves as a sign, giving the name of the work site and that of the person in charge; no other quarry inscription at Claudianus combines these two pieces of information. The inscribed area is cut in a rectangle 0.69 × 0.30 m. The third line is inscribed in a shallow cutting that probably results from the erasure of an earlier text. The letters of the three lines, between 6 and 7 cm in height, are aligned one under the other; as line 1 is one letter longer than the two following ones, the final υ of l. 2 and the final ι of l. 3 are aligned under the space between the ο and ς of l. 1. Under the main text, at 90 degrees, is the beginning of an incomplete inscription. Ε, σ, and ω are lunate; α has a broken bar in ll. 1 and 4; the middle stroke of ε is detached, a trait that appeared toward the end of the Hellenistic period and continues during the imperial period (M. Guarducci, Epigrafia Greca I, p. 380). The paleography does not offer any exact basis for dating, but the name of the quarry is found in two ostraca from the South Sebakh, a dump formed for the most part under Trajan. Σεραπίωνος ἐργοδότου Νικοτύχαι 1. An earlier version of this chapter appeared in a volume in honor of Father Maurice Martin, S.J., in 1992.

147

148

Rome in Egypt’s Eastern Desert

Perpendicularly, 0.24 m below the o of l. 3: Σαρ “Quarry of Serapion, master of works. Nikotychai.” 1.

This could be the ergodotes Serapion, who appears in the amphora inscription O.Claud. inv. 4106, which also comes from the South Sebakh.

2.

The name of the quarry Νικοτύχαι has been recognized recently in two ostraca (Chapter 1, pp. 34–36): not a lot of evidence, but easily explicable by the fact that the exploitation of this rocky bench was clearly suspended shortly after it began. The compound name can be found in the personal Νικότυχος and Νικοτύχη, of which the masculine is much rarer than the feminine (one Νικότυχος to six Νικοτύχη in the first eight volumes of LGPN, one Nicotychus and four Nicotyche in ED Clauss-Slaby). A Νικότυχος, presented variously as a deacon and a scholastikos, appears as the recipient of two letters attributed to Nilos of Ankyra. None of these attestations is earlier than the imperial period. The name is not otherwise known in Egypt.

Should these names be connected with Τύχη or τυγχάνειν? Since the ostraca of Claudianus record a quarry with the compound theophoric name Χρησμοσάραπις, and several others with simple divine names, it is preferable to suppose that the series Νικοτυχ- is derived from Τύχη (which is itself well attested as a woman’s name), Νικότυχος then being a secondary masculine formation from Νικοτύχη (cf. the masculine Ἀγαθήτυχος formed from Ἀγαθητύχη [Masson, OGS I, 229]). The association of victory, nike, and of tyche, notions that are even sometimes equivalent, is a commonplace of ancient art and literature; authors love to connect the two terms in the way that our quarry name combines, or indeed in the reverse order: ἐν ἔργμασι δὲ νικᾷ τύχα, οὐ σθένος, Pindar, Fragm. Hymn. 38; εἰς τὸ νικᾶν τύχης μέτεστι, Plutarch, De Fortuna Romana 318F. Cf. also the acclamation formula equivalent to “long live so-and-so!”: νικᾷ ἡ τύχη τοῦ δεῖνος (Z. Borkowski, Inscriptions des factions à Alexandrie [Warsaw 1981] 76). However, I have not found any literary example of Τύχαι in the plural associated with the idea of victory, except for the following passage on a gold lamella of Orphic inspiration, found at Sybaris in a tomb of the third century BC:2 Ἥλιε Πῦρ διὰ πάντ’ ἄστη νίσεαι, ὅτε Νίκαις ἠδὲ Τύχαις ἐφάνης. On the other hand, the verb which expresses the winning of nike is in the texts routinely τυγχάνειν, e.g., μὴ τυχὼν νίκης (Athenaeus, Deipn. 9.16.29). If this quarry is not that of the victorious (or victory-bringing) Fortunes, what are the Victorious Ones? The columns (αἱ κίονες) that were expected to be extracted in this location? For in fact it happened at Claudianus that monoliths received names (cf. the column Φιλοτραιανός in the quarry of Myrismos, Chapter 1, p. 33). One may wonder if the few letters of line 4 reflect a first try at writing the same inscription. Was the stonecutter insistent on writing Σεραπίωνος rather than Σαραπίωνος? But it is the opposite that we see in O.Claud. IV 660, a draft of a collective letter from the workmen to the procurator metallorum, where the divine name Σέραπις is corrected to Σάραπις, the second form being the older (P. M. Fraser, Ptolemaic Alexandria I, 246), and also the more widespread in Greek in all periods: cf. Clarysse and Paganini 2009, esp. 76 f., where it is shown that the spelling Σεραπ-, which appears at the beginning of the 2. H. Diels, “Ein Orphischer Demeterhymnus,” Festschrift Theodor Gomperz (Vienna 1902) 1–15. Taken up in H. Diels, Die Fragmente der Vorsokratiker I, 6th ed. (Berlin 1961) Orpheus no. 21. On the conditions of its discovery and its date, Comparetti, JHS 3 (1882) 114 and 118.

An inscription of an ἐργοδότης in a quarry at Mons Claudianus

149

imperial period, was always the less common, except—precisely—in the documentation of the Eastern Desert, where the number of attestations of the two vocalizations is roughly equal.

Appendix I. Who had their names inscribed in the quarries of Mons Claudianus? Unlike our Serapion, the several individuals who had their names inscribed in the quarries of Claudianus never specify their function.3 Jean Bingen plausibly supposed that the Hieronymos of quarry no. 83, whose graffito he published in his introduction to O.Claud. I 20–26, was the architect of that name. The Epaphroditos of I.Pan 43 is likely to be the imperial slave who was μισθωτὴς τῶν μετάλλων, known from the dedication of the temple of Sarapis (I.Pan 42). Indeed, it is presumably to this same contractor of the quarries that we should attribute the inscriptions of Wadi al-Hammamat I.KoKo. 54 and 55: the first, inscribed on a block that was being extracted, runs as follows: λ̅ Ἐπαφρ(οδίτου) Καίσαρος. One might consider resolving the initial abbreviated lambda, which perplexed successive editors, as λ(ατομία), although this word is regularly abbreviated λατ( ) or, in Latin, lat( ), in the inscriptions and ostraca from Claudianus. But it is more likely that it is the number of the block or of an extracted object, like the figure γ̅ appearing at the top of the inscription cut on a column base coming from the quarry called “of Myrismos,” which is likely to refer to a column no. 3.4 The λ̅ of I.KoKo. 54 is, moreover, in the same position and thus comparable to στυλ( ), “column,” in the parallel I.KoKo. 55. Another quarry of Claudianus uses the name of the emperor, cf. I.Pan 40, where we read (in the revised text of J. Bingen, O.Claud. I 27–41, introd.) Καίσαρος διὰ Ἡρακλείδου. Here, the genitive indicates not the proprietor of the quarry (after all, the entire metallum was imperial property), but that of the enormous block slated to be extracted from it, on which the inscription is carved, and thus the client who had ordered it. As to Herakleides, he is the architect, well known at Claudianus,5 as another occurrence of this formula on a fragment of a shaft proves, διὰ Ἡρακλείδου ἀρχιτέκτ(ονος) (I.Pan 41). The mention of the ergodotes Serapion in the present inscription probably has the same force as that of the architect Herakleides in I.Pan 40 and 41; both are in charge of the workplace; an inscription from Miletos, to which we will return later, demonstrates the connection of the functions of an ἀρχιτέκτων and an ἐργοδότης: τοῦ μέρους τοῦ θεάτρου, οὗ … ἐργοδοτεῖ ὁ ἀρχιτέκτων Μηνόφιλος, “the part of the theater where the architect Menophilos is directing the works.” At Claudianus, the sphere of responsibility of an architect must have been broader than that of an ergodotes, covering perhaps several work zones and several ergodotai. The texts give no precise indication of such a hierarchical organization, and no ostracon establishes the link between an architect and a particular quarry, but the statements of quarry personnel, to which we shall come later, attest the presence of one or several ergodotai in some quarries. We can exclude the possibility that the architects or ergodotai acquired for their personal accounts the concession of work zones from which they would have derived a private profit: everything that a number of Antonine ostraca from the correspondence exchanged between quarrymen and the procurator metallorum tell us speaks against any notion that at Claudianus there was a collection of independent enterprises, comparable to the coloni who undertook contractually the exploitation of the wells in 3. The quarry inscriptions and quarrymen’s marks of Claudianus were partly published by Meredith 1954. D. P. S. Peacock provides a complete systematic catalogue in Peacock and Maxfield 1997: 216–32 (albeit with numerous typographic errors). 4. Chapter 1, § 74 and Fig. 15. 5. O.Claud. I 27–41.

150

Rome in Egypt’s Eastern Desert

the mines of Aljustrel:6 on the contrary, the community of quarry workers with their foremen seems responsible as a whole for each column quarried, from whatever latomia it may have come. Mons Claudianus, apart from the short-lived episode of the misthotes Epaphroditos,7 must have been exploited under direct rule.

Appendix II. Attestations of ἐργοδότης and ἐργοδοτεῖν A. In documentary texts a. In Egypt (excluding Mons Claudianus and the metallon of Umm Balad8): two proskynemata of the imperial period at Wadi al-Hammamat and four papyri. 1. P.Ashm. I 10 (Hawara, 99/98 BC): a sale contract in demotic, agreed between two embalmers and concerning the rights to provide funerary services in a family tomb; in the damaged Greek subscription, the term ἐργοδότου appears. In the Hellenistic period, ἐργοδότης is attested only with the meaning of the person who ordered work in the context of a contract of hire of a specific task to be completed (locatio conductio operis in Roman law). The role of the word in this contract of sale is not evident. 2. I.KoKo. 112: τὸ προσκοίνημα Ἀπολλωνίου τοῦ υἱοῦ Ἀμμωνᾶ ἐργοδ(ότου). 3. I.KoKo. 48:9 τὸ προσκύνη{ασ}μα Μικκάλου Πετεησίου ἐργοδότου. 4. P.Lond. III 1177 (pp. 180–1) = SB XXVI 16652 (Ptolemais Euergetis, AD 113), where a certain Aphrodisios, ἐργοδότης ἀντλητῶν (“foreman of irrigators”) is mentioned six times; he receives a monthly salary of 40 drachmas for his work in one of the water towers of the city. Depending on the month, he has some thirty to sixty men under his command, each receiving apparently an analogous salary. 5. P.Clermont-Ganneau 2 (SB VI 9230), dated by its editor, André Bataille, to the end of the third century. It is a letter concerning the supply and progress of work in the quarries, addressed by an ergodotes to a colleague who is simultaneously ergodotes and cibariator, i.e., person in charge of foodstuffs, quartermaster: Διόσκορος ἐργοδ(ότης) καὶ οἱ σὺν αὐτῷ σκληρουργοὶ καὶ χαλκεῖς Ἀμμωνᾷ ἐργοδ(ότῃ) καὶ κιβαριάτορει χαίρειν.

6. Domergue 1983: 128–31. The difference between direct and indirect operation of the metalla of the imperial period is clearly set out by the same author in Domergue 1990: 301–5. 7. Epaphroditos Sigerianus is the only contractor of quarries known from the ancient world; on this unique function of his, see the discussions of Klein 1988: 35–37 and Hirt 2010: 318 f. It is perhaps significant that the Serapeum built by Epaphroditos remained unfinished and that the columns which would have supported the architrave, on which the dedication is inscribed, were probably never cut. 8. The ostraca from Umm Balad, excavated in 2002 and 2003 and in the process of publication, provide eight instances of the term but tell us nothing about the precise functions and responsibilities of the ergodotes. At Umm Balad, the ergodotai were named Ammonios, Polydeukes, Sachomneus, and Sokrates. 9. Revised by J. Bingen, CdE 47 (1972) 283 f.

An inscription of an ἐργοδότης in a quarry at Mons Claudianus

151

“Dioskoros, foreman, and the quarry workers and smiths who are with him, to Ammonas, foreman and quartermaster, greetings.” In the address, Ammonas’s double title is repeated in the following form: Ἀμμωνᾷ ἐργοδ(ότῃ) κ(ι-) β(αριάτορι) σκληρουργῶν. 6. P.Lond. V 1648.8 and 15 (Hermopolite, 373): Nomination of liturgists in charge of dike maintenance: ἐκβ]ο̣λέας τε χωμάτων κα[ὶ] χ̣ωμογ̣[ρ]αμμ[ατέα] κ̣αὶ ξυλομέτρη̣ν κ̣α̣[ὶ ἐ]ρ̣γ̣ο̣δ̣ό̣τη[ν] τ̣ῶ̣ν̣δε τῶν χω̣μάτων̣, “(I nominate the persons whose names follow as) assigners of dike work,10 village secretary,11 surveyor,12 and ergodotes of the dikes mentioned.” N. Lewis describes the function of this ergodotes in the following terms: “The ἐργοδότης, the foreman on the spot, directed the actual work on the dikes, assigning specific tasks there to the men sent by the ἐκβολεύς.”13

B. Outside Egypt 7. IK I 117 (Erythrai, ca. 200 BC) Decree of the boule of Antioch on the Maeander honoring three judges “lent” by Erythrai and providing for the erection of two stelai bearing this decree. It was an ergodotes who was to be responsible for the production and inscribing of the stelai. 21

[τὴ]ν δὲ ἔγδοσιν τῆς στήλης καὶ τῆς ἀναγραφῆς [τοῦ ψηφί]σματος ποιήσασθαι Διονύσιον τὸν ἐργοδότην

“The ergodotes Dionysios is to let the contract for the stele and the inscribing of the decree.” Later in the text (ll. 28–32) it is specified that the treasurers will provide the money for the work to the same ergodotes Dionysios and to the person designated as ambassador sent to Erythrai to have the second stele set up there. Here, the ergodotes is the commissioner to whom the city has entrusted the direction of the project. The editors translated the word as “städtische Arbeitsdirektor” and call the ergodotes “ein gewählter städtischer Beamter.” The terms designating these commissioners, who were above all financial controllers, vary by place, period, and circumstances (ἐπιστάτης, ἐργεπιστάτης, ἐπιμελητής, ἐργοδότης, ἐσδοτήρ, ναοποιός, curator operis).14 They thus place the stress on various 10. The function of the ekboleis is described in P.Petaus 49, introduction (they were in charge of designating the people of the village for compulsory labor on the dikes). 11. BL X 106. 12. Ξύλον: a measure of length equivalent to three cubits, which corresponds to one side of a cubic naubion; the ξυλομέτρης checks the volume of earth moved by the laborers (W.Chr. 389, introduction). 13. BASP 8 (1971) 22–23. 14. Hellmann 2002: 37.

152

Rome in Egypt’s Eastern Desert

aspects of the functions of a commissioner. Ἐργοδότης focuses on the fact that part of the functions of these commissioners was to choose contractors (craftsmen, enterprises) and to make legal agreements with them in the name of the city. 8. L. Robert, Le sanctuaire de Sinuri près de Mylasa. I. Les inscriptions (Paris 1945) no. 9 and no. 10 (Hellenistic period): – no. 9: a decree of a tribe (συγγένεια) of Mylasa honoring four ergodotai chosen (ἑλομένων) from among its members to have a portico constructed in a sanctuary. Since these have accepted this appointment, they have earned the gratitude of the community “in proceeding to the letting of contracts for the works in a judicious fashion advantageous to the city, and in acting in this entire matter in accordance with divine and human justice”: τάς τε τῶν ἔργων [ἐ]γδόσεις ἐποιήσαντο καλῶς [κα]ὶ συμφερόντως καὶ λυσιτελῶς τῆι συγγενείαι καὶ ἐμ πᾶσιν ὁσίως καὶ δικαίως ἀνεσ{σ}τράφησαν (27–32). – no. 10: another decree of a tribe to thank one of its members; among the reasons for his fellows to feel grateful to him, we find (ll. 13–16) αἱρεθεὶς δὲ καὶ ἐργοδότης μέρους τινὸς τῆς στοᾶς τῆς κειμένης πρὸς ἡλίου ἀνατολὰς προενόησεν αὐτῆς (broken off), “chosen as ergodotes of a section of the eastern portico, he superintended …” 9. W. Blümel, “Neue Inschriften aus der Region von Mylasa (1988) mit Nachträgen zu IK 34–35,” Epigr. Anat. 13 (1989) 7–8 (no. 895).15 A decree of the city of Olymos, dated to the second half of the second century BC, ordering adornments of a sanctuary of Leto (erection of a statue of the goddess, provision of small cult furniture). The expression “the ergodotai” is mentioned twice in a lacunose context (the left part of the decree is lost). Given the context, provenance, and date, it is most likely that these are commissioners entrusted with letting the contract and directing the execution of the project, as in the three preceding documents. 10. ILS 9219 = S. L. Tuck, Latin Inscriptions in the Kelsey Museum (Ann Arbor 2005) no. 64 (Misenum, 138–192). Funerary inscription of a veteran of the Classis Praetoria Misenensis: P. Aelio Theagene veterano ex Cl(asse) [P]r(aetoria) Misen(atium) militavit ergodota (ll. 3–4). The use of the term in a military context is unique (cf., however, 11), a fact that makes it impossible to be sure of its interpretation. Tuck interprets it as “one who farms out work to contractors.” Theagenes would thus have been a specialist in relations with the businesses to which was entrusted work that the personnel of the fleet did not carry out directly (manufacture of sails, maintenance of the ships, supplies, provisions,16 etc.). His responsibility would thus have consisted of soliciting bids, choosing suppliers, negotiating contracts, following the work, and taking delivery. But, unlike the civic commissioners called ἐργοδόται, this appointment would have been permanent. M.  Reddé prefers the other possible hypothesis: Theagenes would have been “peut-être une sorte de contremaître” (perhaps a sort of foreman).17 In this case, he would have been a director of works (of maintenance of the boats?). Perhaps he would even have had competence in naval engineering: the distinction between the work of an architect and that of an ergodotes is not a clean one, as we shall see. Indeed, we also find the term ἀρχιτέκτων (of which the root

15. Published without translation or commentary. 16. Cf. the conductores fenarii in Rom.Mil.Rec. 80.4 (see P.Bagnall 12, introduction). 17. Reddé 1986: 537

An inscription of an ἐργοδότης in a quarry at Mons Claudianus

153

sense is master carpenter) in naval construction.18 It is curious that the Roman army would have used a Greek word to describe this munus. Was it marine jargon or port jargon? 11. CIL X 2063 (Puteoli, second century). Damaged funerary inscription. The decedent was an ergod[ota]. Another sailor? 12. P.  Herrmann, Inschriften von Milet. Teil 2, Inschriften n. 407–1019. Milet. Ergebnisse der Ausgrabungen und Untersuchungen seit dem Jahre 1899, VI, 2 (Berlin–New York 1998) no. 935. Found in the theater of Miletos (Antonine period), the inscription commemorates a consultation of the oracle of Didyma by a group of masons, apparently upset by the technical difficulties of the task for which they had been engaged. They call into question the ἐργοδοσία of the architect (ll. 1–8):

4

8

οἱ οἰκοδόμοι οἱ περὶ Ε[]νι[] Ἐπίγονον, ἐργολάβοι τοῦ μέ̣ρ̣ους το̣ῦ θεάτρου, oὗ ἐργεπιστατεῖ ὁ προφήτης [θε]ο̣ῦ Οὐλπιανὸς ἥρως, ἐργοδοτεῖ ὁ ἀρχιτέκτων Μηνόφιλος, τὰ̣ εἰλήμ̣α[τα] [κ]αὶ τὰ τετ[ρ]άετα κατὰ τῶν̣ κειόν̣ων̣ περιειλῶσιν καὶ ἐνέγκουσ[ιν ἢ] ἄλλην ἐρ̣γ̣οδοσίαν σκέπτωντα̣ι̣;

“The masons of the team of … Epigonos, contractors for the part of the theater where the prophet of the god, the late19 Ulpianus, directs the project and where Menophilos, the architect, supervises the work zone, should they support (?) the arches (?) and herringbone (?) vaults on the columns and …, or should they instead look to another supervision of the work zone?” This is the only document in which an ergepistates20 and an ergodotes (or more precisely, the corresponding verbs) appear in the same matter. I do not think that the architect whose action is described by the verb ἐργοδοτεῖν replaced the ergepistates because the latter had died.21 The participation of a commissioner and of an architect to support him is the norm in the great architectural programs of the cities. P. Herrmann translates ἐργοδοτεῖ “die Bauausführung vergibt und leitet” (“lets the contract and directs the carrying out of the work”). He thus gives the verb two meanings that it can indeed have, but in different contexts and never at the same time: (a) to give a task to be carried out, to let the contract for a task to an enterprise (19, 20, 25); and (b) to direct the works. Architects are not to be confused with magistrates who let contracts and whose direction of the work is not technical but limited to financial 18. In Latin architectus navalis. Cf. Donderer 1996: 46; Hellmann 2002: 33–34. 19. L. Robert, CRAI 112 (1968) 581, n. 4. The correction of L. Robert (Ἥρως →ἥρως) had escaped me when I was working on the first version of this chapter. 20. For a selective list of epigraphic attestations of ἐπιστάτης τῶν ἔργων and its equivalent ἐργεπιστάτης, see Th. Corsten, Epigr.Anat. 10 (1987) 113–17. 21. Nor do I think that it is necessary to suppose that Ulpianus was a “perpetual” ergepistates who had financed the theater works through a foundation that continued to finance them after his death (as Parke 1985: 77 imagines). L. Robert, who showed that Ulpianus was dead, does not say this. The magistrates who could be αἰώνιοι were the gymnasiarchs, the agonothetai, the stephanephoroi, and the agoranomoi, but not the commissioners of public works. Ulpianus must have died after the conclusion of the contract, which nonetheless remained valid; hence the present tense of ἐργεπιστατεῖ.

154

Rome in Egypt’s Eastern Desert

aspects. On the other hand, several of the testimonia that I have collected show that ἐργοδοτεῖν refers specifically to the work of the architects who were in charge of the workplace (16, 26). This work is called, in lines 7–8 and in the following inscription, ergodosia. 13. I. Nysa 40 (second–third century).22 Dedication of a statue, erected by his sons in honor of M. Aurelius Aphrodisios (lines 2–10):

4

8

Μ(ᾶρκον) Αὐρ(ήλιον) Ἀφροδείσιον Ὀνησίμου Ἀντιοχῆ βουλευτὴν μετὰ πᾶσαν ἀρχὴν ἀρχιτέκτονα ἐν πολλοῖς καὶ φθάνουσιν ἔργοις δοκιμεῖα δόντα ἐπὶ ταῖς ἐργοδοσίαις καὶ νῦν ἐπὶ τοῖς τηλικούτοις ἔργοις εὐδοκιμήσαντα ἐπὶ ταῖς τηλικαύταις ἐργοδοσίαις

“Marcus Aurelius Aphrodisios, son of Onesimos, Antiochian, councilor who has fulfilled all civic offices, architect, who has in the past shown his worth in his supervision of works in numerous building sites, and now, in a time of such great works, distinguished by such important supervisions of works …” In the Miletos inscription as in I.Nysa 40, the role of architect corresponds to that of “project supervisor” in modern law: he plays a role in important projects between the “project owner” (in these cases, the city, represented by a commissioner) and the enterprise. We must keep in mind that the project supervisor, thanks to his technical knowledge, plays the role of designing projects, directing their execution, and helping the project owner in drawing up the schedule of duties (which would be part of the solicitation of bids and the contract), the choices of businesses to carry out the project, the approval of the work, and the settling of accounts.23 In the Milesian case, the architectural terms (the meaning of which has not been clearly established) suggest that there may have been some disagreement of a technical nature between the artisans and the architect;24 the craftsmen ask Apollo if they might not need to change ergodosia. This feminine noun is attested only in these two inscriptions from Asia Minor. Until now, it has not been properly understood. In the original version of this chapter, I followed the universallyaccepted definition of ἐργοδοσία given in LSJ (“letting out work”) and understood the passage to mean that the oikodomoi wanted to be hired by another client, or in other words to break the contract and enter into a new one. But the normal term for such a contract is ἐργολαβία, not ἐργοδοσία. For this reason, I do not think that the oikodomoi intended to break the contract (entered into with the city, represented by Ulpianus).25 In fact, ἐργοδοσία refers to the activity of Menophilos as this is expressed several lines earlier by the verb ἐργοδοτεῖν. The oikodomoi are not in agreement with their project supervisor Menophilos and ask the god if they should not demand a replacement for him (lines 7–8).

22. To appear in IK 36/2. ed. pr. (without translation or commentary): W. M. Ramsay, BCH 7 (1883) 270–72, no. 14. 23. Cf. Brélaz 2003; Saliou 2012. 24. Perhaps there had even been an accident (Donderer 1996: 158). 25. Moreover, would it have been easy for them to be hired for a new project after having broken their contract?

An inscription of an ἐργοδότης in a quarry at Mons Claudianus

155

14. MAMA VIII 577 (Aphrodisias, after 212). Funerary inscription bearing a curse against anyone who should open the tomb in order to take it over for himself: ἐπεὶ [εἴ τις] τούτων τι τολμήσας ἢ ἐπιχειρήσας [ἢ] ἐργοδοτήσας [π]οι[ήσει π]αρὰ τὴν γνώμην καὶ βούλησιν ἐμοῦ ἔσται ἐ[πά]ρατος καὶ ἀσεβὴς καὶ τυμβωρύχος, “for whoever shall have dared, or undertaken, or commissioned any of these actions, against my decision and my wish, he shall be cursed and impious and a tomb-robber…” Ἐργοδοτεῖν here has its original sense, “give a task to be carried out” (20, 25), even in a case where this would be a violation. 15. W. H. Buckler and D. M. Robinson, Sardis VII, Greek and Latin Inscriptions (Leiden 1932) no. 18 (Sardis, 459); reedited by M. di Branco, AnTar 8 (2000) 181–208. A promissory oath solemnly taken by the οἰκοδόμοι of Sardes before the defensor civitatis. They swear that they will no longer leave any works, whether public or private, for which they have been contractually engaged (ἐργολαβεῖν), unfinished for their clients (ἐργοδόται).26 16. IGLS II 349.6 (Chalcis, 550/551), cf. D. Feissel, Travaux et Mémoires 12 (1994) 305–07. The inscription commemorates the building of a section of fortifications (the lacuna at the end of the line is restored thanks to the parallel dedication, IGLS II 348): ἐργοδ̣ο̣τ̣ήσαντος Ἰσιδώρου τοῦ [μεγαλοπρ(επεστάτου) ἰλλουστρίου καὶ μηχανικοῦ]. Μηχανικός appears as a synonym for ἀρχιτέκτων starting in the second century.27 D. Feissel interprets ἐργοδοτεῖν in its legal sense (p. 307: [ἐργοδοτεῖν] “signifie qu’Isidôros, en tant qu’employeur, avait adjugé les travaux à des équipes d’ouvriers”), but in my view this is another example of this verb referring to the function of the architect as project supervisor. 17. A couple of Christian funerary inscriptions from the necropolis of Korykos (Cilicia) give ἐργοδότης, without further specification, as the occupation of the deceased (MAMA III 587 and 686). Despite the ergodotes of the well operators in P.Lond. III 1177 (4), the term normally is to be connected with the domain of building and public works. As a descriptor for the deceased, ἐργοδότης does not indicate a contingent occupation (project head, team head): ergodotes, foreman, must be a profession. Two other inscriptions from the same necropolis describe the decedent as κεραμεὺς ἐργοδότης or ἐργοδότης κεραμεύς (MAMA III 627 and 688). My first reaction was to see in them heads of ceramic workshops, but the ergodotai of the other two inscriptions and the existence of combinations of occupations in the funerary inscriptions of Korykos lead me rather to understand, as did Évelyne Patlagean, the two dead men as having had two concomitant professions, that of ergodotes and that of potter.28 On the other 26. M. di Branco invokes as a parallel SB XX 15090 (Oxyrhynchos?, 543), a complaint addressed by a businessman to the defensor civitatis against an absentee workman (ll.  2–5: οὐκ ἠνέσχετο πρὸς τὰ σύμφωνα τοῦ ἑαυτοῦ μισθοτικοῦ ἐξακολουθῆσαι, ἀλλʼ ἀνεχώρησεν ἐκ τῆς ἐργασίας τοῦ ἐμοῦ ἐργαστηρίου, “he has refused to comply with the conditions of his contract, and he has withdrawn from work in my workshop”). Di Branco indeed calls the complainant an ἐργοδότης, but this word does not appear in the text (a master craftsman is never called ἐργοδότης in the papyri). In any event, the legal relationship of this master craftsman with a decamping workman with whom he had concluded an employment contract is not the same as that of the commissioners of work at Sardes, who had entered into business contracts with unscrupulous masons. 27. Hellmann 2002: 33. 28. Patlagean 1977: 167. Some surprising combinations: gardener-goldsmith (κηπεργοῦ αὐραρίου, MAMA III 348b), butcher-accountant (μακελαρίου λογαρίτου, MAMA III 280), sackmaker-potter (σακκᾶ κ(αὶ) κεραμέος, MAMA III 470). Having two occupations was normal in the working classes. Thus, in the 19th century, in Europe, workers often combined several occupations to even out seasonal unemployment in one or the other. Gérard Noiriel mentions the example of a cobbler from Malakoff who was also a market gardener and another who was also a school-teacher. He observes that it was impossible to say which of the two occupations was the principal one (G. Noiriel, Une histoire populaire de la France. De la guerre de Cent Ans à nos jours [Paris 2018] 307).

156

Rome in Egypt’s Eastern Desert

hand, the translation proposed by Patlagean for ergodotes, “bailleur de travail,” does not mean anything and mixes up the notions of project ownership and carrying the work out.

B. Ἐργοδότης and ἐργοδοτεῖν in literary texts 18. Xenophon, Cyr. 8.2.4.8. Far from being specialists, artisans in small cities had multiple competences and were able to carry out the tasks of a joiner, a carpenter, and a mason (which recalls the multiple occupations in the inscriptions of Korykos). Even so, such an artisan thinks himself lucky if he can find enough clients to earn a living: ἀγαπᾷ ἢν καὶ οὕτως ἱκανοὺς αὐτὸν τρέφειν ἐργοδότας λαμβάνῃ. 19. Apollodorus Gelous 20 (fourth–third century BC): Fragment with the participle ἐργοδοτῶν. 20. Phrynichus, Eclogae 322 (second century AD): ἐργοδότης οὐ κεῖται, τὸ δὲ ἐργοδοτεῖν παρά τινι τῶν νεωτέρων κωμῳδῶν, οἷς καὶ αὐτοῖς οὐ πειστέον, “ergodotes is not allowed (sc. in Attic dialect), but ergodotein is found in one of the authors of New Comedy, who are not themselves trustworthy.” The poet implicated is precisely Apollodorus Gelous. 21. Aretaeus Medicus, SD 1.6.6, ed. Hude p. 42 (second century AD). A case of manic madness in a carpenter: τέκτων ἤδη ἐπὶ οἴκου μὲν σαόφρων ἐργάτης ἦν, μετρῆσαι ξύλον, κόψαι, ξῦσαι, ξυγγομφῶσαι, ἁρμόσαι, ξυντελέσαι δόμον νηφαλέως, τοῖσι ἐργοδοτέῃσι ὁμιλῆσαι, ξυμβῆναι, ἀμεῖψαι τὰ ἔργα μισθοῦ δικαίου, “There was a carpenter who was when at home a mentally sound workman, skilled in measuring, cutting, planing, nailing, and adjusting wood, in building a house properly, in negotiating with his clients, coming to agreement with them, and receiving proper compensation for his work.” I do not share the view expressed in LSJ, s.v., that ἐργοδότης was used incorrectly here with the meaning “workman.”29 The remainder of the list confirms that is a matter of relations with his clients, the project owners. 22. Artemidorus, Oneirocriticon 1.56: ὅμοιος δ’ ἂν εἴη καὶ δεσπότῃ ⟨καὶ⟩ ἐργοδότῃ καὶ φίλῳ τρέφοντι καὶ παντὶ τῷ βαστάζοντι. In a dream, “(the horse that someone is mounting) could represent a master, a client, a friend who feeds you, and any person who is a source of support.” 23. Artemidorus, Oneirocriticon 1.24: ἀνδρὶ δὲ χειροτέχνῃ ἀγαθόν· πολλῶν γὰρ ἀκούσεται ἐργοδοτῶν “(dreaming that one has extra ears) is a good omen for an artisan: he will hear many clients.” 24. Pollux, Onomasticon 7.182 (second century AD): ἐν μέντοι τοῖς ἄλλοις τεχνίταις ὁ Πλάτων ἐν τῷ δευτέρῳ τῆς Πολιτείας καὶ τοὺς ἐργολάβους καταλέγει· “ῥαψῳδοί, χορευταί, ὑποκριταί, ἐργολάβοι.” νῦν μὲν οὖν τοὺς περὶ τὴν σκηνὴν λέγει. Ἐργολάβους δὲ καὶ πάντας τοὺς ἐργολαβοῦντας τι ἔργον ἔστιν εἰπεῖν, ὡς τοὺς ἐναντίους, τοὺς ἐκδιδόντας, ἐργοδότας εἴρηκε Ξενοφῶν, τοὺς δὲ ἐργολάβους καὶ ἐργολήπτας Τηλεκλείδης ὁ κωμικός. Οἱ μέντοι ῥήτορες τὸ ἐργολαβεῖν ἐπὶ τοῦ ἐπηρεάζειν λέγουσιν. Τοὺς δὲ ἐφεστηκότας τῇ τῶν ἔργων ἐπιμελείᾳ οἱ μὲν Ἀττικοὶ ἐπιστάτας ἔργων λέγουσιν, Ἐπίχαρμος δὲ καὶ ἐργεπιστάτας. “Among men who have a skill, Plato, in the second book of the Republic, mentions also the ergolaboi (those who undertake work): ‘rhapsodes, dancers, actors, ergolaboi.’ He is thus speaking of theater people. But ergolaboi can be used of anyone who undertakes work under a business contract, while for the other party, who supplies the work to be carried out, Xenophon uses ergodotai. The ergolaboi are called ergoleptai by Teleclides the comic poet. The orators use ergolabein 29. An anomaly that does, however, occur in a passage of Athanasius (28), probably because of a textual corruption.

An inscription of an ἐργοδότης in a quarry at Mons Claudianus

157

(undertake work) with the meaning of ‘harm’. Those who are put in charge of the organization of works, the Attic writers call them epistatai (directors) of works, and Epicharmus also calls them ergepistatai.” 25. Pollux, Onomasticon 7.200: συνθηματιαίους δὲ στεφάνους εἴρηκεν Ἀριστοφάνης τοὺς ἠργολαβημένους, οὓς οἱ νῦν ἐκδοσίμους λέγουσιν· φατέον δὲ ἐπὶ μὲν τοῦ ἐκδιδόντος ἔργον ὁτιοῦν τὸ ἐργοδοτεῖν, ἐπὶ δὲ τοῦ ἐργαζομένου τὸ ἐργολαβεῖν, “Aristophanes calls synthematiaious (‘agreed on’) crowns that were the object of a business contract, what we now call ekdosimoi (‘things given for execution’). One has to say, in connection with the one who gives any task to be carried out, ergodotein (‘to give to carry out’) and in connection with the one carrying out the work, ergolabein (‘undertake to carry out a task’).” 26. Dionysius of Alexandria, Περὶ Φύσεως (third century AD), ap. Eusebius, Praeparatio Evangelica 14.25.7 (fourth century AD). A refutation of Epicurean philosophy, which conceives of the world as being without creator, but instead composed of atoms: εἰ δὲ ἑκουσίως ἐθελουργῆσαι συγκέκλῃνται, θαυμάσιός τις αὐτῶν ἀρχιτέκτων ἐργοδοτῶν προηγήσατο, καθάπερ εὔτακτος στρατηγὸς οὐ συγκεχυμένην εἴασε τὴν στρατιὰν καὶ πάντας ἀναμίξ, ἀλλ’ ἐν μέρει μὲν τὴν ἵππον, ἰδίᾳ δὲ τοὺς ὁπλίτας τούς τε ἀκοντιστὰς καθ’ ἑαυτοὺς καὶ χωρὶς τοὺς τοξότας καὶ τοὺς σφενδονήτας ἔνθα ἐχρῆν διετάξατο, ἵνα ἀλλήλοις οἱ ὁμόσκευοι συμμαχοῖεν. “If in fact they have been assembled to carry out their work voluntarily, a remarkable architect has distributed them and presides over them, or perhaps, like a knowledgeable general, he has not left his army in disorder, in general confusion: he has arranged them where they need to be, the cavalry on one side, the hoplites and javelineers on another, and the archers and slingers separately, so that like-armed troops may fight together” (adapting the translation of E. des Places, Sources Chrétiennes 338). Ἐργοδοτεῖν, expressing the activity of the architect, refers unambiguously here to the supervision of the workplace, not to contracting out to businesses. 27. Gregory of Nazianzus (fourth century) uses ἐργοδότης in an original manner. In two passages in which he is paraphrasing Exodus, he applies the expression πικρὸς ἐργοδότης to the diabolical figure of Pharaoh (Oratio 16 = Migne, P.G. 35, 940 A) and to his fiends (Oratio 45 = Migne, P.G. 36, 644 A), transposing ἐργοδιώκτης and the classical ἐπιστάτης τῶν ἔργων used in the Septuagint (Exodus 3:7 and 1:11). In Exodus, this last expression must be understood in its literal meaning (director of works), and not in the specialized meaning that it bears in the Bauinschriften. This is the first non-Egyptian attestation of ἐργοδότης with the meaning “foreman.” Gregory may use ἐργοδότης with a different meaning in his Oratio 40 (Migne, P.G. 36, 385  B), an exegesis of the parable of the workers of the eleventh hour (Ev. Matt. 20). He comments as follows on the protests of the workers who, having been the first to work in the vineyard, and thus those who worked the longest, expected to receive a higher wage than the last to be hired: πῶς οὖν αἰτιῶνται τὸν ἐργοδότην, ὡς διὰ τὸ τῆς ἰσότητος ἄνισον, “How can they accuse the ergodotes of having been unfair through fairness?” The Latin translation of the P.G. gives ἐργοδότην the legal meaning of employer (operarum conductorem) in an employment contract, but the ancient sources do not use ἐργοδότης to refer to the employer in this type of contract. I think that in this case ἐργοδότης has its original sense, i.e., the party in a business contract who lets a task to be completed. 28. Athanasius (fourth century) compares the tenacity of gold miners who have just found a vein to that of theologians, no less relentless in their work, who are determined to draw all possible spiritual richness found in the Gospel of Luke: Ὥσπερ οἱ τὴν χρυσῖτιν γῆν μεταλλεύειν λαχόντες ἐργάται… τὸν αὐτὸν

158

Rome in Egypt’s Eastern Desert

τρόπον καὶ ἡμεῖς οἱ τῆς ἐκκλησιαστικῆς ἐργασίας ἐργοδόται… (Migne, P.G. 28, 944 A). Nothing in this comparison, which it would take too much space to cite in extenso, justifies ἐργοδόται, probably just a corruption of ἐργάται by a copyist who found it irreverent to call theologians “laborers.” 29. In numerous lexicographers we find the following gloss: ἐργοδόται· οἱ τὰ ἔργα ἐκδιδόντες τοῖς δημιουργοῖς, “ergodotai: those who give tasks to artisans to carry out.” Thus, those who commissioned work, the clients. 30. The rule of the convent of the Theotokos at Constantinople (drawn up in 1110), describes in Chapter 27 the work of the ἐργοδότριαι, the nuns in charge of works in the workroom of the convent (text and translation in P. Gautier, “Le typikon de la Theotokos Kécharitôménè,” Revue des études byzantines 43 [1985] 19–155). Their name comes from the fact that they hand out the work (in the form of raw material to work) to the sisters: ἐργοδοτεῖν ταύτην (scil. τὴν ὕλην) ταῖς ἐργαζομέναις. They then receive back the finished products: παραλαμβάνειν τὰ ἐργοδοτηθέντα εἰργασμένα, “the finished work which had been given out to be executed.” Ἐργοδότης and ἐργοδοτεῖν thus allow two possible meanings. Originally, these words had a legal character and were used to refer to one of the parties in a contract of hiring of a task to be completed (= a business contract) and to his actions, respectively, denoting a legal relationship like that of the Roman locatio conductio operis. Through this action, a task is given to an artisan, or if it involves major works, to one or several enterprises, to carry out. The ἐργοδότης is the person who commissions a task (or his agent in the case of commissioners of works elected by cities); in modern times, in business law (mostly in construction), the person who orders the work is called “project owner.” The other party to the contract, who is to carry out the work (craftsman, business), is the ἐργολάβος, the contractor. Beginning in the imperial period, ἐργοδότης and ἐργοδοτεῖν acquire a second and more concrete meaning, alongside their original one: to distribute tasks to a group of workers, to organize the work in a workshop or on a building site. The verb is attested with this meaning in two inscriptions from Asia Minor and refers specifically to the work of the architect (12, 16, 26); in this period, ἐργοδότης takes on the meaning of the head of a building site or foreman, but only in Egypt, where it is used rarely, except in the imperial quarries. There is no instance of the verb ἐργοδοτεῖν in the Egyptian documentation. Ἐργοδότης is attested for the first time in Xenophon. It was not an Attic word; Xenophon borrowed it from elsewhere,30 probably as a result of his wanderings in Asia Minor, and he is the only author to have used it before the imperial period (24). Its corresponding verb ἐργοδοτεῖν (“to give a task to be carried out in the context of a business contract”) is found just once in a writer of New Comedy, Apollodorus Gelous (fourth–third century BC; 19, 20). In the ancient texts in which ἐργοδότης and ἐργοδοτεῖν express a contractual relationship, it is always a situation of the hiring for a task (business contract, locatio operis), and not for hiring personal services, as is the case when an employer enters into an agreement with personnel whom he employs (employment contract, locatio operarum). It is only later, in Modern Greek, that ἐργοδότης refers to the employer in an employment contract, whether this employer is a private individual or the head of a business, whether large or small; by contrast, someone who assigns a task to a craftsman to carry out is no longer called an ἐργοδότης as in ancient Greek, but a πελάτης (client).31 The use of ἐργοδότης to refer to a foreman might explain this slide toward the mod30. W. G. Rutherford, The New Phrynichus. Being a Revised Text of the Ecloga of the Grammarian Phrynichus (Hildesheim 1968) 456. 31. I thank Despina Chatzivasiliou for this information.

An inscription of an ἐργοδότης in a quarry at Mons Claudianus

159

ern meaning of head of a business: the foreman, like the head, distributes tasks among the workmen. He is called ἐργοδότης not because he employs them, but because he gives them work to do. In the Hellenistic period, ἐργοδότης appears in a handful of Bauinschriften from Asia Minor in a specialized meaning: that of the commissioner entrusted by a city to organize the carrying out of public works. In all periods, other terms were used more commonly to refer to this characteristic function of civic life, particularly ἐπιστάτης and ἐργεπιστάτης. These commissioners had an intermediate role between the project owner and the foreman. They acted in the name of the true project owner (the city that had appointed them), and entered into the detailed letting of contracts for tasks to businesses; this is probably the reason for the choice of the term ἐργοδότης to refer to them in a small number of Hellenistic inscriptions. But when it came to large-scale public works, these men lacked the technical knowledge to supervise the progress of the work. Their responsibilities thus focused on financial control of the project, with the technical direction and supervision entrusted to an architect, whose work in turn could be referred to by the verb ἐργοδοτεῖν (in one inscription of Asia Minor from the second century AD [12], the meaning of which is clearly confirmed in the third century by a philosophical treatise of Dionysius of Alexandria [26]) and by the abstract noun ἐργοδοσία, attested only in two inscriptions of the imperial period from Asia Minor (12, 13). The ἐργοδοσία of the architect corresponds to project supervision in modern operations. This supervision is today entrusted to an architect, an engineer, a design office, or the business itself, depending on the nature and importance of the project. The project supervisor finds technical solutions for the needs of the project owner, directs the work, follows their progress, and checks them.32 He assists the project owner in the choice of organizations, but he does not himself let the contract and is therefore not part of the contract of hire. In Greek, an ἀρχιτέκτων remains an ἀρχιτέκτων; he is never called ἐργοδότης. In Egypt, under the early empire, ἐργοδότης is used in a particular sense, that of site manager, head of a team. The term corresponds roughly to that of raïs in contemporary Egyptian usage. The context is always that of work that requires numerous workers (the quarries of Mons Claudianus and 4–6). As far as our present documentation shows, this use of the term is specific to Egypt in the imperial period. It may have appeared first in the great imperial metalla of the Eastern Desert, where the legal and geographical conditions demanded a particular organization of the local manpower: it was for the same reason that the technical term κιβαριάτης was devised. Although at a humbler level in terms of scale and technical expertise, the work of the ergodotes-site supervisor corresponds to that expressed, in the case of an architect, by the verb ἐργοδοτεῖν (which is not attested in the Egyptian documentation). One may even wonder if the supervision of the site, ἐργοδοσία, was not carried out by lower-status architects; M.-Chr. Hellmann remarks that in the site accounts, the daily wage of the architects is on the same level as that of the workers whom they supervised, and that often they were not those who planned the programs that they had to carry out in the context of their ergodosia.33 In documentary papyri of the fifth and sixth centuries AD, ἐργοδότης is no longer used; instead it is replaced by ἐργοδιώκτης. This term is known already in the third century BC: the first attestations occur in a papyrus from the Arsinoite nome dated to 256 BC (P.Petrie Kleon 51) and in the Greek translation of Exodus in the Septuagint, which also dates to the reign of Ptolemy II. The papyrus is a letter in which the δεκατάρχοι of the stonecutters (λατόμοι) complain to the ἀρχιτέκτων Kleon about the ἐργοδιώκτης Apollonios, who has assigned them to work on hard stone, while saving the softer stone 32. Thus it is only after approval of the architekton that the epistates issues orders for payment. 33. Hellmann 2002: 50: “en fait, ces architectes dont nous connaissons les salaires étaient de simples employés considérés comme des chefs de chantier.” The same idea is developed by Jacquemin 1990: 84.

160

Rome in Egypt’s Eastern Desert

for his own men.34 In literature, ἐργοδιώκτης is attested only in the Old Testament and in theological commentaries. The return of ἐργοδιώκτης in papyri from the Byzantine period, after a long period in eclipse, is in my view connected to its biblical use. This late documentary usage is not limited to Egypt: thus, one finds a rock inscription in the name of a site boss called ἐργοδιώκτης in a limestone quarry near Bostra; this document dates, at the earliest, to the fifth century.35 Although the papyri of the Hellenistic and Roman periods use ἐργολάβος, ἐργολαβία, and ἐργολαβεῖν in connection with carrying out tasks through a business contract,36 the “Egyptian” meaning of ἐργοδότης lacks the juridical connotation that it had in the authors of those periods: no source gives the impression that ἐργοδότης-site manager owed its use to the fact that he entered into a contract of ἐργολαβία with the workers under his direction. No business director (even if the Egyptian ergodotai were such heads, something for which there is no proof) is ever called ἐργοδότης. Letronne, Fitzler,37 and Bataille38 indeed recognized that the ἐργοδότης of the Egyptian documentation is only a team leader; but Letronne was captive to the legal meaning, which caused him to restore ἐργολ[άβου] instead of ἐργοδ(ότου) mistakenly in one of the inscriptions from Wadi al-Hammamat (3), in such a way as to create a bizarre pairing ergolabos-ergodotes that is impossible in this context.39

Appendix III. The ergodotai of Mons Claudianus A. Prosopography A minimum of fifteen different ergodotai have been identified:40 Ἀπολλώνιος, Ἀπολλῶς (perhaps the same man), Ἁρποχράτης, Ἔννις, Παλαπούηρις, Πατωοῦς, Πελέας, Πετέχνουμις, Πετενεφώτης son of Παλαπούηρις, Πετενεφώτης son of Ζμενάχνουμις, Πίηγχις, Σάνσνως, Σεραπίων, Σωκράτης, Φθαῦς son of Πέβως, Φμοιπελέας, Χρυσέρως: these are the names of the ergodotai of Mons Claudianus, mostly Egyptian or Greco-Egyptian. Among them are epichoric names of the Theban region (Πετενεφώτης) or of Aswan (Πελέας, Φμοιπελέας). One name is Latin: Ἔννις (Ennius),41 which comes from the same archaeological context, in the southwest dump, as several ostraca dated to year 4 of Antoninus; but it is troubling that the only bearer of this nomen, in the entire Claudianus prosopography, is Ἔννις Πίηγχις, holder of 34. The site manager Apollonios was thus also a business director, but this is not the function that made him be referred to as ἐργοδιώκτης. 35. SEG XLV 1988 (Studies in the History and Archaeology of Jordan 5 [1995] 400–1 [P.-L. Gatier]), cf. Bull.Epigr. 1996, 494. 36. And also in the secondary sense, which Pollux attributes to the orators, of harmful business (24). 37. Fitzler 1910: 134. But to refer to the Egyptian ergodotes as a Beamter, as he does, is infelicitous. 38. Aegyptus 31 (1951) 207–8. 39. Letronne, Recueil des inscriptions grecques et latines de l’Égypte II (Paris 1848) 433: “Ainsi l’ἐργοδότης était l’homme de l’ἐργολάβος.” 40. In the absence of patronymics, we cannot be sure that there were not multiple ergodotai with the names Sansnos, Sokrates, or Pienchis. 41. O.Claud. inv. 7182. To follow the reasoning that leads to the identification of the ergodotes Ennis with the ergodotes Pienchis, one must know that the dump south of the camp is considered to have been formed in the main from debris dating to the reign of Trajan; in 1990, we recognized, thanks to prosopographical links, that it had been contaminated, at least in its upper layers in sub-units d6 and d7, with texts from the reign of Antoninus: the same individuals and the same type of texts, specifically some entolai, appear in the southeast corner of the fort (with specific dates ranging from year 5 of Antoninus to year 6 of Septimius Severus); the campaign of 1991 finally showed that the whole site seems to have been invaded by this mainly Antonine documentation.

An inscription of an ἐργοδότης in a quarry at Mons Claudianus

161

a ticket for bread,42 who is likely to be either the same person as Πίηγχις ἐργοδότης, or his son: three lists of names mention Πίηγχις ἐργοδότης and a fourth, which belongs to the same dossier,43 an Ἄνις Πίηγχις, Ἄνις being evidently a faulty writing of Ἔννις.44 Although the individuals named Ἔννις Πίηγχις, Πίηγχις ἐργοδότης, and Ἄνις Πίηγχις appear in the ostraca from the south dump (where the Antonine occupation is visible in a more limited fashion), it is not impossible that these texts are in fact contemporaneous with the amphora inscription in the name of Ἔννις ἐργοδότης, which was found in an Antonine context. In fact, the bread chit of Ennis Pienchis mentions the number of 50 pairs of loaves of bread, like the tickets from the reign of Antoninus, instead of 55, the typical amount specified in the tickets found in the south dump and thought to date to the reign of Trajan. But it must be observed that the individuals who surround the ergodotes Pienchis in the lists of names where he appears do not reappear in the Antonine prosopography of Mons Claudianus,45 where Ennis himself makes only one appearance. The ergodotai Harpochrates, Phmoipeleas, Apollonios, and Apollos, these last two being perhaps a single individual, unquestionably belong to the Antonine phase of the exploitation of the quarry. Phthaus son of Pebos, who issues a receipt dated to year 13 of Trajan (O.Claud. inv. 1545), is the only precisely dated ergodotes. Finally, a remark about Χρυσέρως, a rare name in Egypt,46 which allows us to think that the six attestations of this name in O.Claud., all of which come from the south dump, refer to the same man; in any event, the Chyseros son of Trophimos listed third among the other workmen assigned ἰς τὰς πλάκας is certainly the same as Chryseros ergodotes in another list of names (inv. 3822), for these two lists provide another onomastic connection, in the person of Satornilos son of Gaius.

B. What information do the ostraca give us about the function of ergodotes? The ostraca from Mons Claudianus provide 102 occurrences of the occupation name ergodotes. Among these are 37 tituli on amphoras; 8 lists of names; one label for ζεύγη of bread; 14 rosters of quarry workers; 5 organization charts for the distribution of water; 11 collective letters addressed by the ergodotai along with their fellow workers to the procurator metallorum or to a prefect (military? of Berenike?);47 some ergodotai figure as creditors or debtors to one of their colleagues in 4 entolai, to which one may add an entole issued by an ergodotes. Some other texts that do not belong to groups (receipts, requests, orders for delivery) bear witness to the economic and administrative role of the ergodotes. a. The number of ergodotai The ostracon that gives us the best view of the composition of the microsociety of Claudianus is the great organization chart of the distribution of water, inv. 1538+2921 (Chapter 11), which, being virtually complete, encompasses the entirety of the personnel, from the centurion at the top down to the 212 42. O.Claud. inv. 2947. Several of these tickets are published in Chapter 7. The ergodotes Sansnos also has his name on a ticket for 50 pairs of loaves of bread. 43. O.Claud. inv. 4178+4182. This is the dossier of the “lists of four (or five).” 44. The mistake may be caused by the fact that Ennius does not otherwise occur in Egypt, whereas Annius is well attested in Syene, as R. S. Bagnall remarks. 45. With the exception of Bekis son of Sokrates, a combination of names too commonplace to connect Ennis Pienchis to the worker community of the reign of Antoninus with any certainty. The Antonine prosopography of Claudianus is essentially supplied by the archives of the kibariatai, which seem slightly later than year 4 of Antoninus. 46. There are only four papyrological attestations (cf. CPR VI.2, 80.11–12n.; add P.Stras. VI 592, ii.5). 47. O.Claud. IV 848–863.

162

Rome in Egypt’s Eastern Desert

sklerourgoi (stonecutters) from Alexandria, to whom are added the 130 who have come from Syene, 40 smiths and 25 temperers whose origin is not specified, the other large group being that of the familia, which amounts to no fewer than 400 individuals. All in all, 924 individuals are accounted for by the water service on a given day at Mons Claudianus. Of these 924 men, 7 are ergodotai. Along with the two highest-ranking persons present on the site, the centurion and the decurion, a certain Herakl( ) appears, who is likely to be the architect Herakleides, particularly because in another organization chart he appears alongside one Apoll( ), who is probably his contemporary the architect Apollonios. I shall not go into the difficult question of the familia here. It is only necessary to note that in the quarry Myrismou, where a team of paganoi and a team belonging to the category of familia work side by side, the ergodotes is included among the paganoi. If we consider only the 421 individuals of this last group, we find that on average there was one ergodotes for 60 men. From the same text, it emerges that nine quarries, of variable importance, were in use at that moment. There was not, therefore a system of assigning one ergodotes per quarry. There was a hierarchy among the ergodotai: Sansnos is the only one among them who is mentioned by name and who has an entitlement to one keramion of water, while his colleagues are simply called “the six others” and only receive a half keramion each. The amounts distributed to the centurion, the decurion, and the architect have been left blank. Most of the other employees, whether military or civilian (e.g., veterinarian, tesserarius, tirones, all of the paganoi workers), receive a half-keramion, which was thus the base ration. The members of the familia were entitled to only a third of a keramion; the stratiotai and the cavalrymen get somewhat more favorable treatment: they receive 1/2 + 1/3 (thus 5/6) of a keramion. Sansnos thus has a double ration, while his colleagues, the other ergodotai, are treated on the same footing as the workmen whom they supervised. The fact that there was not an ergodotes assigned to each quarry also emerges, it seems, from another series of texts, the rosters, which enumerate the specialists for a particular quarry:48 of the 14 complete examples, only 6 mention the presence of one or two ergodotai. It is nonetheless not impossible that in some cases where there is no explicit mention of an ergodotes he is implicitly counted among the sklerourgoi. To judge from these texts, an ergodotes might have under his orders 59, 56, 49, 42, or 26 workers. O.Claud. IV 647 and 648 are two rosters drawn up at least three months apart in the quarry “Myrismou”; in them, we find two ergodotai for a total workforce of respectively 45 and 89 men. These numbers may, however, be deceptive. The quarries that appear in the organization chart inv. 1538+2921 are all worked by both paganoi and members of the familia; but a third roster for Myrismou, which is also complete, makes a distinction, among the specialists, between the paganoi and the familia (O.Claud. IV 649); the ergodotes is in that case counted among the 42 paganoi. Did his authority also extend to the familia? Do the other rosters, less carefully, mix the two juridical categories together? We have no means to answer these questions. b. Primus inter pares The hypothesis that the ergodotes might have been present in a quarry without being mentioned separately in the roster rests on the fact that his status does not seem to be much higher than that of the men he directed. This appears also in the strongly indigenous onomastic repertory of the ergodotai. We have also seen how the ergodotes Chryseros was registered in a list of workmen without any identifying mark 48. O.Claud. IV 632–690.

An inscription of an ἐργοδότης in a quarry at Mons Claudianus

163

(he is not the first name on the list49 and his title is omitted); in the rosters, the position of ergodotes/ergodotai occupies the first place, alternating with the position of sklerourgoi. The ergodotes is listed first, right before the sklerourgoi, in four of these texts (three of which concern Myrismou); in three or four cases, it is listed second, preceded by the sklerourgoi. In four cases, it falls as low as fourth, fifth, or sixth position, with the sklerourgoi always listed first. The same fluctuation can be seen in the prescripts of the collective letters addressed by the workers to the procurator metallorum or the eparchos: in four cases, the senders describe themselves as “the ergodotai and the sklerourgoi who are working in the quarries of Claudianus (or of Porphyrites),” but in one case they are instead “the sklerourgoi, the ergodotai, and the smiths.” These prescripts bear witness to the same spirit as the letter P.Clermont-Ganneau 2 (“Dioskoros, ergodotes, and the sklerourgoi and smiths who are with him”), and one could not put it better than Bataille did: “On sent dans sa lettre qu’il se considère comme étroitement solidaire de ses hommes et il menait probablement la même vie” (One feels in his letter that he sees himself as united with his men and that he probably leads the same life as they).50 The letter inv. 6252, of which the recipient, the ergodotes Apollos, is respectfully called κύριε by his correspondent, is an exception.51 All the same, the ergodotai occupied a distinct position, just like the more prominent members of religious associations, and needed to entertain their comrades on feast days: an account of πόσις lists their contribution of four kolophonia of wine; two other accounts from the same dossier mention an imperial freedman and a centurion among the contributors.52 c. The economic functions of the ergodotes Of the ergodotes at work, the body of texts from Mons Claudianus gives us even a limited view only of his administrative and economic roles. O.Claud. IV 819 is a note written by the ergodotes Sansnos to request a skin (of water?) for the quarry of Apollo; O.Claud. IV 818 is another request, for an unknown object, coming from Phthaus. The ergodotes needed to be literate. In inv. 5811, the ergodotes Phmoipeleas is asked by Peleas for four pairs of loaves of bread. Although we have no explicit example from Mons Claudianus of the combination of the functions of kibariator and ergodotes, the latter must sometimes also have assumed the function of manager of provisions, like Ammonas in P.Clermont-Ganneau 2.53 The kibariatai54 were chosen in rotation and usually for a short period among the workers (probably the most influential and best educated), to make the trip to the valley for supplies, collect the monthly wages, and manage the current bills of their comrades. This original institution fitted readily into the picture that we have sketched of this milieu of Egyptian quarry workers, a community that seems to have had a sense of shared responsibility and functioned in an egalitarian manner.

49. The same observation is true for the dekania-lists, where the head of a team, the dekanos, is not the first named. 50. Aegyptus 31 (1951) 208. 51. An exception all the more remarkable in that the term κύριος is not frequently used in the letters found at Mons Claudianus: it appears mainly when civilians address soldiers or the procurator metallorum, or when soldiers address someone of a higher rank. 52. Respectively O.Claud. inv. 5370, 5019, and 5912. 53. The Latin origin of the word κιβαριάτωρ should not mislead us: although the interference of the Roman administration in the work of the quarries latinized their vocabulary, we should not follow Bataille in concluding that the kibariator was a soldier (Aegyptus 31 [1951] 209–10). 54. The bureaucracy of Mons Claudianus privileged the form κιβαριάτης instead of κιβαριάτωρ, which is used in P.Clermont-Ganneau 2 and the ostraca from Pselchis (O.Wilck. 1128–46).

9 The amount of the wages paid to the quarry-workers at Mons Claudianus O.Claud. inv. 47511 FSE – r1 S (14) Fig. 47

136–146 silt clay

ἐντολ(ὴ) Πάχουμις μηνὸς Θωθ. ὀψωνίου (δραχμαὶ) μ̅ζ̅. τούτων προχρίας (δραχμαὶ) κ, ἀποχὴν λαμβ(άνεις) ἐλαίου κοτύλας γ̅ φακοῦ μάτιν α κρομμύων μάτ(ιν) α οἴνου κεράμιν α συνβολῆς (δραχμαὶ) γ τὸν σῖτον ἰς ὄρος δαπάνη (τετρώβολον) τὰς λυπ(ὰς) ἰς ὄρος.

4

8

1 εντολ 10 

14.5 × 10.5

l. Παχούμιος 11 l. λοιπ(άς)

2 3 l. προχρείας λαμβ l. εἰς

6 ματ

8 l. συμβ-

9 ϊς, l. εἰς

“Instructions from Pachoumis for the month of Thoth. Wages: 47 drachmas. Deductions: advance 20 drachmas (you get a receipt); 3 kotylai of oil; 1 mation of lentils; 1 mation of onions; 1 amphora of wine; symbole: 3 drachmas. My wheat to the desert. Dapane: 4 obols. The rest to the desert.”

1. SB XXIV 16173.

165

166

Rome in Egypt’s Eastern Desert

Figure 47. SB XXIV 16173. © A. Bülow-Jacobsen

Pachoumis worked in the imperial quarries at Mons Claudianus. He was either a quarryman, a stonemason, or a smith. He belonged to the group of workers called pagani. The workforce at Mons Claudianus was divided into two categories, the familia and the pagani. The pagani were the native, free, skilled workers who came mostly from Syene, from Alexandria, and probably especially from the Theban region, if their names are anything to go by. The two categories of workers were mercenarii: both pagani and familia received a money-payment which is always called opsonion.2 In addition to this sum they received a ration of victuals which was not the same for the two groups: the pagani were entitled to 1 artaba of wheat plus a wine ration, the quantity and distribution of which is, however, uncertain. The familia received 1 artaba of wheat, lentils, and oil and, once a year, a set of clothes.3 The pagani had themselves to procure oil, lentils, and other goods from the Nile Valley; the price was deducted from their wages. The text published above gives some examples of other deductions that could be made from the wages, such as reimbursement of advances on pay and contributions to associations (dapane, symbole). The paying of the pagani and the familia respectively created two different kinds of documentation issuing from two different administrations, i.e., the entolai for the pagani and the receipts for advances on pay for the familia.4 The formulae of the receipts for advances on pay did not contain the amount of the opsonion, which is therefore unknown. On the other hand, in the entolai which each paganus-worker wrote every month to the quartermaster (kibariates), the 2. In the everyday language of the Hellenistic period, ὀψώνιον replaces the classical μισθός, which continues in literary usage (Launey 1950: 726). On the other hand, scribes setting up work-contracts nearly always use μισθός instead of ὀψώνιον, in my opinion because it was felt to be more solemn (Hengstl 1972: 44 and n. 54 for locatio operarum; p. 60 for locatio operis). 3. The only attestation is O.Claud. III 432. 4. The receipts for advances are now published in O.Claud. III.

The amount of the wages paid to the quarry-workers

167

amount of the monthly wages is hardly ever omitted. The entolai are usually not dated, but the receipts for advances, with which they are normally found, often are, and they can thus be dated to the decade 136–146.5 As a documentary type the entolai are extremely well represented among the ostraca from Mons Claudianus: out of a total of some 9,100 registered ostraca, about 1,240 are entolai.6 It is a new type of text in papyrology, a product of the special conditions that decided the organization of the work in the imperial quarries. These entolai will be published in the series Ostraca Claudiana, but their quantity is such that this publication may not appear for several years. For this reason, I have thought it useful to publish one here, in order to submit in advance to historians the information they contain concerning the wages in money of the free workers in the imperial quarries. Hitherto, from the whole Roman Empire, only the pay of two Dacian miners was known, and we had almost no information on the pay of quarrymen.7

I. The pay scale (1) Pachoumis earned 47 drachmas per month, which is by far the commonest amount attested. In their present, nearly definitive, state the texts mention 106 workers with wages of this size, but their specific occupation is not mentioned once. (2) 37 drachmas 4 obols. This rate of pay is attested for nine individuals, two of whom are smiths, one being a χαλκεύς, the other a στομωτής, a steel-temperer. The other seven do not state their craft. I cannot exclude the possibility that the lower pay-rate is a question of age, since in three cases the less well paid are sons, included by their father on his entole. An abnormal amount of 37 drachmas is found in three entolai, of which at least two were submitted by Sarapion the Younger – in the third case the name of the worker is missing. Since the same Sarapion, in other entolai including duplicates of those of 37 drachmas, states his pay as 37 drachmas 4 obols, we must assume that the 37 drachmas are an error, the reason for which escapes us. (3) 28 drachmas. This is the lowest rate, known for five workers only. In four cases it is the pay of a son and once of a brother included on the entole of a worker who earns the normal 47 drachmas. The fifth case is that of Isidoros, who was in the habit, instead of giving his patronymic (which was also Isidoros, as we know from other documents), of putting παῖς after his name. I shall leave aside the well-known ambiguity of the word παῖς and its derivatives (“slave” or “young boy”?); suffice it here that Isidoros, with his patronymic, is not a slave, but probably, like others who earn 28 drachmas, an apprentice. Incidentally, the low number of apprentices is surprising. 5. There are no receipts for advances earlier than 136, by which year we could, therefore, believe that the two management systems were established. However, Wilfried van Rengen told me that an entole dated to Hadrian’s 13th year has been found in Porphyrites (inv. 553). The latest receipt dates from 197, but the nature of the archaeological contexts makes it impossible to date with confidence entolai other than those found together with receipts from 136–146. 6. The uncertainty about the exact number of entolai results from the existence of related texts (accounts, receipts) which are more or less close to the entolai in form or function. 7. There is perhaps an exception: PSI VIII 962, B.25–39 (131/2) is a work-contract through which an inhabitant of a village in the Herakleopolite nome binds himself to assist his employer in μεταλλευτικὴ ἐργασία for one year. For this he receives 180 drachmas, of which 160 are paid in advance. This gives a monthly total of 15 drachmas, which is not much, even if one should add various payments in kind specified in a lacuna. But the text presents many uncertainties, and we do not know whether the work was full-time, nor whether mining or quarrying was concerned.

168

Rome in Egypt’s Eastern Desert

The wages of 47 drachmas, 37 drachmas 4 obols, and 28 drachmas belong to the same pay scale. The proof of this is that fathers who are paid 47 drachmas often include sons or younger brothers on their entole who earn 37 drachmas 4 obols. There are rare traces of another pay scale, in which a maximum of 48 drachmas (received, for example, by a blacksmith8) corresponds to lower grades of 38 drachmas, 38 drachmas 2 obols, and 38 drachmas 4 obols (the last two are found within the same month). It would take us too far from our subject here to enter into the details, but the facts at our disposal suggest that the rise in wages was introduced around 151, perhaps after an interruption of the quarrying. The uncertainty concerning the amount of the lower wages must derive from the disorder that ensued when operations restarted with a different pay scale. We may thus conclude that wages of the native, skilled workers employed at Mons Claudianus depended only on the age of the worker, not on his craft, since quarrymen and smiths received the same pay.9 The pay scale was very even, and one quickly reached the top level of 47 drachmas. The workers at Mons Claudianus were better paid than those of the Nile Valley. On the basis of the table of monthly wages compiled by Drexhage,10 one can calculate that the average monthly pay for a civilian in Egypt in the second century was a little more than 25 drachmas. The maximum during the same period was 40 drachmas, but this was uncommon (only two cases out of twenty-two). Not until around 250 do we find wages of above 40 drachmas. It should be noted that these monthly payments are often lower than the results obtained by multiplying by thirty certain daily payments, which may run as high as 4 drachmas per day. The impression is that payment per day was often more advantageous, either because the man did not work regularly or because a temporary employee weighs less heavily on the budget of the employer, who could thus afford to appear more generous. These 47 drachmas represent about half of the pay of a legionary infantryman, who earned 300 denarii a year, equal to 1,200 drachmas, or, on a monthly basis, 100 drachmas per month, but who had to pay for his own wheat. Traditionally, the monthly ration per man in Egypt is one artaba, or 38.78 liters, of wheat. The fixed, official price for an artaba of wheat was at this time eight drachmas. The market price was a little higher, at an average of 9.2 drachmas in the first half of the second century.11 These prices mean that the 47 drachmas were more or less equivalent to the price of five artabas of wheat. So, since one artaba is enough to nourish one active, grown man, the better paid among our workers received enough to feed five people.12 There is never any question of deductions from the wages because of religious feasts, holidays, or illness. One cannot, however, completely exclude that such deductions were made in disguised forms, e.g., as payments to an official.

8. This shows that the “high” wages of 47 and 48 drachmas were not reserved for quarrymen. 9. Many different crafts are mentioned in the O.Claud., but they are hardly ever used to distinguish individuals. For this reason, we do not know the occupations of our pagani, except for a very few cases where a sklerourgos (quarryman), a stomotes (steel-temperer), or a mechanikos (machine operator) are mentioned. The pay of the mechanikos is not known. 10. Drexhage 1991: 425–29. I have selected for my calculations only the 22 cases which seemed certain. 11. Based on Nos. 21–24 of the table of market prices of wheat in Lower and middle Egypt by Duncan-Jones 1990: 151. The published texts do not give a trustworthy impression of prices in Upper Egypt (Duncan-Jones 1990: 152), but perhaps it was not much different, in spite of the impression given by certain ostraca concerned with taxation, where it would seem to be much lower. In a unique ostracon from Mons Claudianus (O.Claud. inv. 1077, a letter) there is discussion of a sale of wheat during the reign of Trajan or Hadrian. Two possible prices of two or three staters respectively are mentioned in a somewhat obscure context, i.e., 8 or 12 drachmas. 12. The calculation is easily made on the basis of the data collected by Foxhall and Forbes 1982.

The amount of the wages paid to the quarry-workers

169

II. Comparison with the wages of the Dacian miners The table of Drexhage mentioned above gives instant confirmation that nearly all wages in drachmas in Roman Egypt are sums divisible by four, or, in other words, they can be reduced to a round sum in staters or tetradrachms. This preference for sums divisible by four is seen also in the entolai, where advances on pay are often 20 or 40 drachmas and transfers between individuals are very often divisible by four.13 Now, the only amount from the period 136–146 which is divisible by four is that of 28 drachmas, which is also the least attested, and one has to ask what calculations were behind the peculiar amounts of 47 drachmas and 37 drachmas 4 obols. They are apparently not the results of a multiplication of a daily pay14 since, at Mons Claudianus, the monthly wages are not divisible by thirty, the number of days in every Egyptian month, nor, for that matter, by any lower number of days, if feast-days were after all deducted.15 At almost the same time, but far from Egypt, in Dacia, three illiterate workers were hired to work in the mines of Alburnus Maior. Copies of the contracts with which they bound themselves to their employer were written on tabulae ceratae and show, among other things, the length of the employment and the total wages which would be paid to them in several unspecified instalments. The last column in the table below shows that in the two cases where there are enough data to make the calculations, the total is remarkably close to the payment in money received by the workers at Mons Claudianus. The first edition of the “Transylvanian triptychs” is found in CIL III.2. The three contracts of employment that are of interest to us are TC IX, X, XI, pp. 948 ff. I use the latest, and best, text, established after consultation of the originals by Röhle.16 The texts are reproduced by Noeske, who also gives a detailed commentary (Noeske 1977: 396–404). Table 9.1. Dacian mining contracts that mention wages CIL No.

Date Hired

Expires

Length

Total Pay

Pay for 30 Days (in drachmas) (1)

TC IX

23 Oct. 163

13 Nov. 164

388 days (2)

90 denarii (3)

27.835

TC X

19 May 164

13 Nov. 164 (4)

179 days

70 denarii children (5): 10 denarii

46.927 6.703

TC XI (6)

in lacuna

13 Nov.

half a year?

105 denarii

70?

Remarks (1) It is known that the denarius was statutorily equivalent to the tetradrachm or stater, so that the drachma was statutorily equivalent to the sestertius.17 It is thus possible to convert the Dacian wages into Egyptian monthly wages and to see that the result is very close to 28 and 47 drachmas. The conversion is, of course, theoretical and serves for verification only, since nothing points to the Dacian wages being paid in monthly instalments.

13. It should be remembered in this connection that taxes were payable in tetradrachms (Gara 1976: 77 and 144). 14. This, as we shall see below, has been believed by several scholars in treatments of the wages of the Dacian miners. 15. There were, of course, holidays, but probably not the same number every month. 16. Noeske 1977: 398, n. 681. 17. Gara 1976: 14.

170

Rome in Egypt’s Eastern Desert

(2) Ciulei is mistaken when he sets the duration of this contract at one year and thirteen days (Ciulei 1983: 49). I do not understand how Noeske obtains a sum of 386 days. If we include the day when the contract was written (ex hac die) and the day on which it expires, as he does for TC X, I get 388 days in the following way: October 163 November, April, June, September December, January, March, May, July, August, October February (NB: 164 is a leap-year) November 164 Total

30 × 4 31 × 7

9 120 217 29 13 388

(3) The original supplement [sept]aginta was made in analogy with TC X, but was not cogent, because the duration of the contract is not the same in the two tablets, and the amount would be abnormally low. Later the supplement [non]aginta has been preferred (Mrozek 1975: 72; Noeske 1977: 397). (4) The date 19 May, instead of 20 May, is based on a new reading by Röhle (Noeske 1977: 400, n. 703)18 which extends the duration of the contract to 179 days instead of 178. Ciulei wrongly arrives at five months and thirteen days (Ciulei 1983: 49). (5) Röhle returns to the original reading liberisque instead of cibarisque as proposed by Carcopino.19 Carcopino’s correction was widely accepted and is still upheld by some (Röhle 1968: 188–91). Noeske accepts liberisque (1977: 402), and this reading does, indeed, seem necessary from the facsimile in CIL III.2, p. 948: i and e are certain.20 The pay of these children is much below that of Mons Claudianus and suggests that the children were younger than our pais Isidoros and the sons of workers. Unlike mining, where certain operations require more agility than physical strength, quarrying is unsuitable for child-labor. Children were also employed in Egyptian gold-mines, as we hear from Agatharchides in the second century BC (De Mari Erythraeo 26). (6) The date of contract TC XI is lost, so Noeske does not take much interest in this text, where the daily pay cannot be calculated. Mrozek, on the other hand (1977: 104), tries to show that the contract was concerned with a period of six months. He argues that the employment contracts traditionally expired on the Ides of November, since TC IX and TC X suggest that they were set up at the end of May for an employment of about six months, or before the Ides of November for an employment of about one year. I find this argumentation interesting, on condition that we accept that TC IX and TC X are representative of the time of the year when contracts were normally set up. Since the payment stipulated for this period 18. The facsimile in CIL, which supports the reading liberisque (cf. following remark [5]) does, admittedly, not confirm the reading XIIII as made by Röhle (in Acta Musei Napocensis 6 [1969] 515 f.). 19. J. Carcopino, “Note sur la tablette de Cluj CIL III nr X p. 948,” Revue de Philologie 63 (1937) 103. 20. The problem is whether the facsimile is accurate or if liberis is a wishful reading. A good photograph would dispel the remaining doubts, but those I have seen are, unfortunately, not of an adequate quality (Inscriptiones Daciae Romanae I, p. 234; Apulum 7 [1968] 324). Röhle claims to have seen “klar und deutlich die Buchstaben liberisque” on a photograph which has been sent to him (Röhle 1968: 189). Besides Noeske, M. Kaser, Das romische Privatrecht I (2nd ed.) (Munich 1971) 570, n. 72, also accepts the reading liberisque, which, on the other hand, is rejected by H. Kloft, “Arbeit und Arbeitsverträge in der griechisch-römischen Welt,” Saeculum 35 (1984) 215, n. 75, although not on paleographical grounds. I. I. Russu, Inscriptiones Daciae Romanae I 41, does not choose between the two.

The amount of the wages paid to the quarry-workers

171

is 105 denarii, Mrozek deduces that the annual pay was 210 denarii, which would correspond, on Mons Claudianus, to a monthly pay of exactly 70 drachmas. We do have a single example of this amount, in an account which is related to the entolai. This text, O.Claud. inv. 8429, shows deductions from the wages specified for two consecutive months. The first two lines run: Φαο̣φι ὀψ(ωνίου) (δραχμαὶ) ο, Ἁ̣{τ̣}θυ⟨ρ⟩ ὀψ(ωνίου) (δραχμαὶ) ο, “Wages for Phaophi, 70 drachmas, wages for Hathyr 70 drachmas.” This ostracon would confirm Mrozek’s hypothesis, but the conclusion is fragile: TC IX and TC X do not concern a period of exactly a year or six months, so why should this be the case with TC XI?

III. Is monthly pay the result of a multiplication of daily pay or a division of an annual salary? The basis for the calculation of the miners’ pay has been much discussed. In my opinion both Mrozek and Noeske are following a false trail when they try hard to show that the wages were the result of a multiplication of the daily pay. Thus, Noeske starts with the postulate that the wages mentioned in the contracts result from a multiplication of a certain number of working days by a daily pay and that this daily pay must be payable in the coinage of the day. As 90 denarii divided by the number of working days does not result in a round figure, he concludes that the daily pay was 4 asses and 2 quadrantes and introduces the idea that the period of work included non-paid days of rest (Noeske 1977: 397–8). Several arguments can be put forward against this: – It is doubtful that the round figures of 70, 90, or 105 denarii result from such a multiplication. – It was unimportant to be able to pay the workers by the day: since these workers were employed for a longer period, they were not day-laborers. – The Egyptian papyri suggest that the monthly wages in money were not on the same level as the day-by-day payment. The two types of payment are used in different situations, where both the nature of the work and its duration varied. We have already seen that the daily wages were sometimes proportionally higher than the monthly ones and that, on the other hand, the monthly wages were generally divisible by four, i.e., could be paid in staters. Consequently, the monthly wages are not the result of a multiplication of the daily pay, but are based on a different calculation, although we should not forget that the wages at Mons Claudianus differ from the norm in Egypt, inasmuch as two out of the three pay-rates cannot be paid in staters. The fact that the amounts are the same every month on Mons Claudianus suggests that the number of working days was not taken into account. In spite of all the subtleties they use to calculate the daily pay of the workers, Mrozek and Noeske speak in terms of wages for a year or half a year.21 As a matter of fact, the concept of annual pay is attractive and would explain the atypical monthly wages paid at Mons Claudianus. Let us begin with the fact that 70 denarii are the pay for half a year minus seven days, in TC X. This would mean that yearly pay was 140 denarii.22 This results in a striking coincidence. 21. Mrozek 1968: 318: on the basis of TC XI, “We can calculate that the annual pay of a miner was 210 denarii or 2.3 sesterces or more than 9 asses a day.” Mrozek 1977: 104: “der jährliche Verdienst des Restitutus agnomine Senioris betrug also 210 Denare, was pro Tag ungefähr 2.3 Sesterzen ausmacht”; Noeske 1977: 402: “wenn Memmius Asclepi für ein halbes Jahr Arbeit im Bergwerk 70 Denare verdient...” 22. Thus, put by Domergue 1990: 345.

172

Rome in Egypt’s Eastern Desert

140 denarii, or 560 sesterces, are not divisible by twelve, but 141, or 564 sesterces or drachmas, are and the result is 47, the monthly wages at Mons Claudianus. 141 is of course an unexpected amount for a year,23 but perhaps we can be allowed to believe that the sum was adjusted in order to make it payable in monthly instalments.24 If we try the same model on the payments of 28 drachmas (= 7 staters or denarii) a month, it would be derived from an annual sum of 84 denarii, which is the pay of the worker in TC IX; the six supplementary denarii correspond to the 23 days by which the employment exceeds a year. It should be noted that the payment for these last six days, which by strict reckoning should be 5.293 denarii, is rounded up to the next whole denarius in the worker’s favor. It seems probable that the gap between a strict pay scale and the variable periods of employment left room for bargaining, but no trace of this has been found at Mons Claudianus. The demonstration is less convincing with the amounts of 37 drachmas 4 obols,25 both because the Transylvanian tablets have not preserved documentation of a similar rate, and because the annual rate of 113 denarii, on which it should be based if we apply the same principle, is less suitable for payment in monthly instalments than 114 would have been:26 113 denarii divided by 12 gives 37 drachmas and a remainder of 8 drachmas (= 56 obols), which cannot be divided by 12, unless we use the traditional drachma of 6 obols (instead of 7, as was normal at this time). The result of this would be 8 × 48 ÷ 12 = 4 obols.27 The entolai of Mons Claudianus attest wages which are, in two cases, abnormal. Since the earnings in money of these quarrymen are exactly the same as those of the two miners in Alburnus Maior whose earnings can be known with certainty, I propose that the peculiar amounts on Mons Claudianus are the result of a division into monthly instalments of an annual pay which was fixed for all free workers in the metalla (quarries and mines) of the Empire, where the tradition may have been to employ workers on a yearly or half-yearly basis, as suggested by the Dacian contracts. Such a general regulation of wages, made to work across differences in monetary systems, calendar,28 and methods of exploitation,29 could 23. In the Lex Ursonensis (Spain, first century AD) the annual remuneration (annua) of the municipal employees is a multiple of the aureus (= 25 denarii), see CIL II Suppl. 5439, ch. 62 and the comment by Mrozek 1975: 75 f. 24. This is not normally the case for the yearly wages in Egypt as listed by Drexhage 1991: 430. Only in one case out of twelve is the amount divisible by twelve, but on the other hand it is nearly always divisible by four. 25. I have wondered whether the wages of 37 drachmas 4 obols could be calculated on the same basis as the daily payment of 5 asses, attested in CIL IV Suppl. 4000 (Pompeii, ante AD 79). The inscription is quoted, e.g., by W. Krenkel, “Währungen, Preise und Lohne in Rom,” Das Altertum 7 (1961) 167–78, at 175 (non vidi, reference taken from Szilagyi 1963, at 347). At this rate the payment for thirty days would indeed be 37 sestertii and 2 asses, or at the statutory exchange-rate, 37 drachmas and 3.5 obols. But this is surely a coincidence, since we have seen that the daily rates were used for irregular and temporary employment and were not used as a basis for the wages of permanent personnel. 26. An annual rate of 114 denarii would mean a monthly payment of 38 drachmas, which sum is sporadically attested on Mons Claudianus (see above). 27. This solution supposes that drachmas of 6 and 7 obols, respectively, are used in the same operation. This could be awkward, but is not without parallel: in P.Cair.Mich. 359, Part ii, p. 14, n. 10, Shelton quotes P.Lond. I 131 (corr. in BL I 230), where 85 drachmas 2 obols of bronze are converted into 71 drachmas 1 obol of silver, at a rate of 29 obols for 4 drachmas. As Shelton notes, although the 85 drachmas have been converted at the announced rate into 17 tetradrachms, the last 19 obols were converted into 3 drachmas 1 obol calculating 6 obols to the drachma. Since the 6-obol drachma was better suited to the Egyptian monetary system than that of 7 obols, accountants must have been tempted to use it (Gara 1976: 71, n. 47). 28. The miners were not employed for a round number of months, and it is not known at what intervals they were paid. The contracts use the vague per tempora, for which see Noeske 1977: 399 and 401. At Mons Claudianus, workers were apparently paid at the end of each month, as was usual for everyone who was paid by the month. 29. Mons Claudianus was under direct administration: the workers let their service to the emperor, represented by his procurator. The mines at Alburnus Maior were under indirect administration: the imperial procurator lets the mine shafts to private entrepreneurs who hire and pay the workers (Domergue 1990: 301–5).

The amount of the wages paid to the quarry-workers

173

be used as an argument in favor of the hypothesis of Dušanić, who suggests that there existed a central office in Rome, a hypothesis which he bases on other indications of a centralization of the administration of the metalla throughout the Empire.30 If my hypothesis, by which I have tried to explain the similarity of pay scales in Dacia and at Mons Claudianus, is correct, it is an unexpected and striking example of the economic integration of the Empire such as Augustus wanted it, when he imposed the denarius, or its statutory equivalents, as the universal currency-unit. If it is not correct, it remains interesting that these free miners and quarrymen, working in two provinces so far apart, received the same payment in money, a payment which was above the earnings of other members of their class, like manual workers or farm-laborers.31 One cannot, however, maintain that the earnings of the miners in Alburnus Maior were exactly the same as those of the quarrymen at Mons Claudianus, since we do not know if the miners also received rations in kind. Such rations are not mentioned in any contract, now that cibarisque has been rejected, even though some commentators believe that a grain-ration was a matter of course, so that there was no need to mention it.32

30. Perhaps a tabularium dependant on the procurator a rationibus, see Dušanić 1989: 154 ff. 31. Noeske 1977: 402; Mrozek 1989: 166. 32. Corbier 1980:  81. The locus classicus in this connection is Dig. 38.1.50.1 (thus Mrozek 1977:  105): Non solum autem libertum, sed et quemlibet alium operas edentem alendum: aut satis temporis ad quaestum alimentorum relinquendum: et in omnibus tempora ad curam corporis necessaria relinquenda (“But not only the freedman, but anyone else engaged in performing services, is to be provided with food, or to be allowed sufficient time to earn the price of his food, and all are to be left time to take necessary care of themselves,” trans. S. Jameson in A. Watson, ed., The Digest of Justinian II [Philadelphia 1985]). The argument is open to discussion, firstly because the passage does not concern paid work, but the rather special case of free services; secondly, rations may be of different sizes, as we have seen for the pagani and the familia at Mons Claudianus, and, furthermore, in the Egyptian work-contracts the description of payment in kind is never less precise than that of the payment in money (cf. the numerous examples collected and analyzed by Hengstl 1972, passim).

10 Two ostraca from Mons Claudianus: O.Bahria 20 and 21 The native workmen (pagani) of the quarries of Mons Claudianus, that is, those who did not belong to the imperial familia, received each month a salary in cash (in almost all cases, 47 drachmas per month), a wheat ration, and during some as yet unexplained circumstances, a wine ration.1 They entrusted to one or several of their number to go to the Nile valley to collect what was due to them, and to carry out all needed transactions (buying supplemental foodstuffs, borrowing money, repaying debts, making various other payments). Each worker wrote, or had written for him, his instructions on a potsherd, often in the form of a memorandum addressed to his agent, the kibariates. This memorandum was called an entole (“instructions”). The distinctive conditions of work in the quarries of Mons Claudianus, the remoteness, and the communal organizational structures made necessary in an environment managed by the army and the imperial administration, thus produced (during the years AD 136–146) an original documentary type, with its own diplomatics and jargon, which was hitherto unknown: the entole. At any rate, it was not previously recognized. But in fact, an ostracon supposed to have come from the western desert, O.Bahria 20, turns out to be an entole from Mons Claudianus.2 O.Bahria 20 belongs to a group of 22 ostraca, 17 of which are in Greek and were published by G. Wagner (1987: 86–95). Because I had an opportunity to glance through the “Temporary Register”3 in which this group is inventoried, I observed that in the case of the last three ostraca of the series (nos. 20, 21, and 22, published as O.Bahria 20, 21, and 22), the indicated provenance is not the oasis of Bahria,

1. See Chapter 9. 2. Wagner 1987: 94 and pl. XXXIX. 3. This is separate from the “Journal d’entrée” of which Wagner speaks (1987: 86) – and also from the “Special Register”: the numbers assigned to objects in the “Temporary Register” are numbers created from four numbers, shown in the form of a cross: the first three express a date (day, month, year: date of the inventory), and the fourth is a sequence number within a group. In the case of the ostraca published by G. Wagner under the abbreviation O.Bahria, the first three numbers are 25-748 (= 25 July 1948), while the sequence numbers run from 1 to 22. I thank Adam Bülow-Jacobsen for having explained these niceties to me.

175

176

Rome in Egypt’s Eastern Desert

but the following: “found in Mus Monkey house 1948.”4 It was apparently because they were inventoried at the same time as those found by Ahmed Fakhry in Bahria that the three ostraca of the Monkey House were included in the same lot and placed in the same cardboard box.5 Thanks to the courtesy of the authorities of the Egyptian Museum, I was able to examine O.Bahria 20 to 22.6 There is no doubt that O.Bahria 20 and 21 come from Mons Claudianus, but the same is not true for no. 22, which I shall leave aside.7 In order to support my corrections to O.Bahria 20, I have added the text of an entole from Mons Claudianus issued by the same worker, Artemas son of Artemidoros, from whom we now have a total of six entolai, all in the same hand (which does not need to be that of Artemas itself, because it also wrote the entolai of several other workmen).

O.Bahria 20 (Re-ed. SB XXIV 16060). The opening of the entole of the Cairo Museum presents two particularities that are signs of its early date: first, it is an epistolary formula with the greeting χαίρειν, and without the word ἐντολή mentioned at the start, as is usually the case, for example in the other entole of Artemas published below. Second, the instructions are addressed to two kibariatai rather than to a single one. The amount of the monthly salary is not indicated, but we know from other entolai of Artemas that it was 47 drachmas.

4

8

Ἀρτεμᾶς Ἀρτε[μι]δ(ώρου) Βησαρίω̣ν̣ι {ιω(νι)} Β̣ησα[1–3] κα̣ὶ Ἀπολιναρίῳ κιβ(αριάταις) μηνὸ(ς) Ἐπειφ χαίρ̣ε̣[ιν.] Ἁρβεσχείνι Παχνούμ(εως) (δραχμαὶ) μβ (τριώβολον) Σάνσν̣ω̣τι π 4–6 τι (δραχμαὶ) η πωλήσατε τὸν σεῖτον καὶ πληρώσατε π̣ᾶ̣ν̣ ἰς ἀναπλήρωσιν τοῦ ὀψωνίου μ⟨ο⟩υ· οἷς ἔδωκα ἐντολὰς vac. τὴν ἐπίθεσιν κιβαvac. ριάταις. Χ

1 Ἀρτεμ[ᾶτος] Β̣ησαρίωνι τῷ Βησα[ρίωνος] ed. pr. 2 απολιναριω κιβ: Ἀπολλιναρίῳ Ὥρου ed. μ pr. 3 l. Ἁρβεσχίνει παχνου : Παχῆος ed. pr. 3–4   4 Παταγῆτι ed. pr., πγ̣ητι Cuvigny 1997 6 π̣ᾶ̣ν̣: νῦν ed. pr., 2–3 Cuvigny 1997 7 μου ed. pr. 8–9 ἐπίθεσιν καὶ κερισταῖς ed. pr.

4. Mus = Mus(eum). The late May Trad, Conservator in the Egyptian Museum, kindly informed me about this Monkey House: “… le ‘Monkey House’, une chambre de dépôt située presque en face de celle du directeur, qui après avoir été vidée depuis une quinzaine d’années, a été transformée en bureau pour les ingénieurs” (letter of 26 February 1997). 5. The specific location, which was kindly communicated to me by Sayed Hassan, Conservator, is P. 24-198. As far as nos. 20 and 21 are concerned, there is a slight disagreement between the date assigned to them in the Temporary Register (25-7-48, like the rest of the group, including no. 22) and that written in ink on the sherd (28-7-48 in the case of no. 20, 21-7-48 for no. 21). 6. I thank Dr. Mohammed Saleh, Director of the Egyptian Museum, who gave me permission to see these texts, as well as Mr. Mohammed Gommaa, Deputy-Director, Mr. Mahmoud el-Halwagyet and Mrs. Somaya Abd el-Sammeia, Conservators, who assisted me in consulting them. 7. For this ostracon (which was not photographed), which is broken at top and left, my readings differ a bit from those of G. Wagner but do not advance understanding of them very much. In line 1, I read: ] (δραχμαὶ) β (διώβολον). Line 2: ] κ̣α̣ὶ̣ διὰ χιρὸς β. Line 3: ] vac. Κορνη( ) κ̣α̣ὶ  / η̣ (the reading ὑ(πὲρ) ι (ἔτους) seems doubtful to me).

Two ostraca from Mons Claudianus: O.Bahria 20 and 21

177

“Artemas son of Artemidoros to Besarion son of Besa(  ) and to Apolinarios, quartermasters for the month of Epeiph, greetings. To Harbeschinis son of Pachnoumis, 42 dr. 3 ob.; to Sansnos … 8 dr. Sell my wheat and pay the full amount(?) up to the full amount of my salary. My ration of wine(?) to the quartermasters to whom I have given my instructions.” 1.

Βησαρίω̣ν̣ι {ιω(νι)} Β̣ησα[1–3]. The ductus of -ωνι is the same as in ὀψωνίου in line 7 (otherwise, what I read as νι could be read as a cursive καί similar to those in lines 2 and 5). Besarion son of Besarion (var. Besas) is attested at Mons Claudianus. His appearances come mainly from the West Sebakh (W.S.), excavated by Jean Bingen in January 1993, which produced the earliest entolai (years 21–22 of Hadrian). Besarion is not known as a kibariates except perhaps in inv. 8403, a mutilated entole dated to 135–137, which comes from the W.S. and is addressed to several kibariatai (which is an indication of an early date within the dossier of the entolai), one of whom has the name or patronymic Besarion. Like O.Bahria 20, inv. 8403 offers a number of anomalies in its formulation compared to later documents. These are probably to be attributed to the fact that the system was still in the process of being put in place. What causes a problem here is the two characters between the name and patronymic: these seem to be an iota topped by a half-circle identical to that used to represent the raised omega of Ἀπολιναρίῳ in line 2. This anomaly might be explained by dittography: the scribe began again above the final ι as if it were the ρ in the middle. Wagner’s reading τῷ shows that he interpreted the traces as τω, which is paleographically impossible: there is no sign of the horizontal stroke of tau. Moreover, the insertion of the definite article before a patronymic is not a practice found in the ostraca from Mons Claudianus.

2.

Ἀπολιναρίῳ. The raised position of the ω is superfluous, since the word is not abbreviated. The same phenomenon appears in inv. 8464, another entole of Artemas son of Artemidoros, written by the same hand, in which the name of one of the two kibariatai to whom it is addressed is written απολλωνιω. The names of the kibariatai are mentioned or preserved in the five entolai of Artemas. In each case, there are two kibariatai mentioned, which is, as I noted earlier, a sign that we are dealing with the first wave of entolai: under Antoninus, the entolai are addressed to a single kibariates. But the pair mentioned here, Besarion and Apolinaris, is not attested elsewhere.

3.

Ἁρβεσχείνι Παχνούμ(εως). Harbeschinis son of Pachnoumis is well represented at Mons Claudianus, both in 135–137 (ostraca from the W.S.) and in the following decade. He both issues entolai and appears in those of his colleagues, always as creditor: they have sums running 4 to 8 and to 24 drachmas paid to him. Here, he receives an even larger sum, because it amounts to 42 drachmas, 3 obols. In another entole, Artemas has precisely the same amount paid to another recipient (inv. 8520).

4.

I have not been able to reach a certain reading for the word that follows Σάνσν̣ω̣τι. It is clear that it is in the dative, but I hesitate to see in this a mistake for genitive, as this writer does not generally confuse cases. Taking account only of the paleography, πατα-, πελα-, παμ-, πεμ- are possible. G. Wagner read Παταγῆτι, which is paleographically possible, but Παταγης, which

178

Rome in Egypt’s Eastern Desert is supposedly attested only in P.Lips. Ι 97, ii.8 passim (fourth century) is a misreading for Παταπης (BL Ι 213). Perhaps instead we should read Πετάλης, a name attested in 91 at the metallon of Wadi al-Hammamat in a collective proskynema containing names characteristic of the Syenite workers of Mons Claudianus (I.KoKo. 53). In place of -ητι, -αιτι or -ειτι would be possible, and perhaps even better. Παμσάιτι? But this name is not attested in the O.Claud.

5.

“Sell my wheat.” The workers sometimes ask the kibariates to sell their allotment of wheat, either to have more cash or because the wheat distributed by the emperor was often spoiled. In the present case, the amount of the sale will allow Artemas to pay off his creditors in full.

5–7.

The end is difficult to understand, and the other entolai offer no parallels for πληροῦν and ἀναπλήρωσις that could enlighten us, apart from the use of the adverb πλήρης in two entolai that were issued precisely by Artemas and of which the best preserved is published below: they end, after the list of expenses, with (γίνονται) (δραχμαὶ) μζ πλήρης, a clause meaning that there is a zero balance and that the requested expenditures equal exactly the credit balance in money. We could translate it “that is, the full amount of my salary.” It means that Artemas is not expecting any money, in contrast with another of his entolai that ends with λοιπὰς δραχμὰς γ (τετρώβολον) εἰς ὄρος. Otherwise, I have not been able to find a secure reading for the word of two or three letters after πληρώσατε. πληρώσατέ μ̣ε̣ is paleographically acceptable; the meaning would be “and pay me up to the full amount of my salary,” but this makes no sense, because his debt is greater than his salary. It is better to read πᾶν. From this point of view, ἰς ἀναπλήρωσιν means that Artemas asks to have the full amount of his salary used to pay off his debt of 50 dr. 3 ob.

7–9.

οἷς ἔδωκα ἐντολάς. This relative clause might be taken to depend on πληρώσατε or from an understood δότε (as commonly in the entolai): in the first case, one would translate “pay in full (those) to whom I have given instructions” (i.e., Harbeschinis and Sansnos); in the second case, the antecedent of the relative pronoun is κιβαριάταις (and therefore Besarion and Apolinarios themselves: it can happen that the recipients of an entole are mentioned in the third person rather than the second in the body of the text). I have adopted this second hypothesis in my translation, despite the vacat after ἐντολάς, which suggests the beginning of another clause.

8–9.

I still do not know why the wine ration, which the workers could dispose of just like their wheat ration in numerous entolai (transfer or sell it), is called ἐπίθεσις.

10.

This cross, which is not taken into account in the ed. pr., often appears at the bottom of the entolai. I think that it is a check-mark, which perhaps signifies that all of the instructions have been entered into the registers in which the kibariatai itemized the individual workers’ orders. In O.Claud. inv. 6687, published below, it is clearly made by a pen other than that of the body of the text.

Two ostraca from Mons Claudianus: O.Bahria 20 and 21

179

Figure 48. SB XXIV 16061. © A. Bülow-Jacobsen

Another entole of Artemas, found at Mons Claudianus O.Claud. inv. 6687 (SB XXIV 16061; Fig. 48) F.N.I – room 1 SE (6)

10.8 × 9.8

140–145 silt clay

The date is deduced from the archaeological context, which has produced ostraca dated between 140 and 145. This entole is later than the one discussed above: the formulas appropriate for this type of text have now been fixed, and the creditors mentioned belong to the “Antonine generation” except for Gaion, who is none other than the caesarianus known from the ostracon published by Cl. Préaux in CdE 26 (1951) 354–63 (SB VI 9457, now improved as O.Claud. III 417), where he also figures as the beneficiary of the repayment of an advance that he made, but in that case to an employee of the familia. I am saving the lexicographical and prosopographical commentary for the publication of the full body of the entolai. But we should note that Hierakion is known as a kibariates. Ηis name is one of the rare examples of hypocoristics in -ιον, which are normally feminine, borne by a man. The amounts due to Gaion and to Peteesis are not absolutely sure, but these are the only paleographically acceptable readings that allow us to reach the total of 47 drachmas mentioned in 1. 8. The adverb πλήρης (1. 8) stresses that Artemas is not expecting to have any credit balance. At the end of the month of Mesore, he will receive nothing of what he is owed, either in money or in kind. How will he survive the next month? The same way as always, on advances.

180

Rome in Egypt’s Eastern Desert ἐντολὴ Ἀρτεμᾶτο(ς) Ἀρτεμιδώρο(υ) μηνὸ(ς) Μεσορη. Γαίωνι (δραχμαὶ) λ̣α̣, Πελαιᾶτι Μαξίμου (δραχμαὶ) ϛ, Πετεήσι Μάρκου (δραχμαὶ) ϛ̣, θεῶι δραχμαὶ δ, τὸν σεῖτόν μου Ἱερακίωι καὶ τὴν ἐπίθεσιν τῶν γ̅ κεραμίω(ν), (γίνονται) (δραχμαὶ) μζ πλήρης. Χ

4

8 1 αρτεματο

2 αρτεμιδωρο μηνο

3–5 

7 κεραμιω

8 γ̅ ̅ ̅ ̅̅

“Instructions of Artemas son of Artemidoros for the month of Mesore. To Gaion 31(?) dr.; to Peleas son of Maximos 6 dr.; to Peteesis son of Marcus 6(?) dr.; to the god 4 dr.; my wheat to Hierakion along with the supplementary ration of 3 amphoras. Total : 47 drachmas exactly.” 3.

λ̣α̣. The alternative would be λ̣ε̣ or μ̣.

5.

ϛ̣. Or γ̣.

O.Bahria 21 The specific formulation of O.Bahria 20 shows that it belonged to a series of which we have found no trace in our excavations. It is likely that it was picked up in the west dump (W.S.), from which come three other entolai of Artemas and most of the entolai addressed to more than one kibariates: actually, we explored this dump only to a very limited extent. From the W.S. comes also the dossier to which O.Bahria 21 belongs (Wagner 1987: 95 and pl. XL): these are lists of names, enumerating the workmen of the familia, a high proportion of whom have Semitic names, like some of the men mentioned in O.Bahria 21. The handwritings are in the same style (a small, fine, slanted cursive), but differ in detail (clearly in the form of υ); the same care is taken in layout to preserve wide margins. The onomastic connections are not many: O.Bahria 21 includes the name Ζμυρναῖ⟨ο⟩ς and another that G. Wagner proposed to read Σλ̣ασινουρος, which I would instead read Σα̣μ̣σ̣ινουρος.8 This name appears also in O.Claud. inv. 8379, where it is written Σημισμουρος, along with a Σμυρναῖ⟨ο⟩ς. There are no other connections between O.Bahria 21 and this series of Semitic names, but the Οὔλπις Γίγας of O.Bahria 21 may well be the same person as the Ulpius Giganus ex familia mentioned in a Latin letter that also comes from the W.S. (O.Claud. inv. 8433). Like the entole O.Bahria 20, the list of names O.Bahria 21 may be the sole witness to an older period of a dossier that the excavation of the W.S. brought to light.

8. Similarly, read in l. 11 Ἀγαθοκλῆς (η:

) instead of Ἀγαθοκλῖς.

11 The organization chart of the personnel of an imperial quarry according to an ostracon of Mons Claudianus The international mission which excavated at Mons Claudianus from 1987 to 1993, under the direction of Jean Bingen and the auspices of the Institut français d’archéologie orientale du Caire, discovered some 9000 ostraca, largely Greek but some also in Latin. These were found mainly in the site’s numerous ancient dumps. O.Claud. inv. 1538 is the crown jewel of this corpus.1 Almost complete, it supplies a schedule of recipients of water in the metallon, classified by rank, status, and occupation, for one particular day.2 This ostracon is the best preserved of a small archive of fourteen or fifteen similar schedules, which are for the most part very fragmentary and will be published later; I shall, however, need to refer to them from time to time to support difficult readings or the resolution of abbreviations. This entire series comes from the dump which stood on the outside of the south wall of the larger of the two fortified villages built at Mons Claudianus, that of Wadi Umm Husayn;3 this village is called praesidium in the ostraca.

1. I must express my gratitude to my papyrological friends of Mons Claudianus, above all to Jean Bingen, who generously entrusted the study of this document to me. I thank also Rudolf Haensch and Christof Schuler, the directors of the Kommission für Alte Geschichte und Epigraphik, for having invited me to present it in Munich. 2. The ostraca use the term μέταλλον to refer to a zone of operation (for example Mons Claudianus or Porphyrites) and the term λατομία to designate each work area of extraction. 3. This dump is referred to as “South Sebakh” (= S.S) in the publications of ostraca from Mons Claudianus. The most complete description can be found in Maxfield and Peacock 2001a: 109–25.

181

182

Rome in Egypt’s Eastern Desert

A. The text Inv. 1538 + 2921 S.S – e7 N (5) & d7 NE (5) Fig. 49

c. 110 26 × 19.5 cm

The text stretches in its length across a fragment from the neck of an amphora,4 with the writing perpendicular to the wheel marks on the pot. The variable darkness of the ink, sometimes very pronounced, shows that the schedule had been prepared in advance: the scribe left blanks in which to write the exact number of recipients and the amount distributed; certain blanks were, however, never filled in, whether because the recipients were not present on that day, or because the scribe did not get any information about them. The ink of the first round is generally paler than that of the additions, which I indicate by bold type in the transcription, but sometimes the difference is not visible. I have therefore used bold type only when I was certain that there was an addition. The procedure that I have described was not applied systematically. In the third part, from the barber (l. 9) until l. 12, the text seems to have been drawn up all in one pass: it must have concerned fixed, recurring information; on the other hand, the entire end of the document, starting with Φαννίῳ στρατ( ) (l. 14), was added at the moment when the blanks were filled: it in fact seems to bring together the unforeseeable cases, which changed from day to day. For reasons of space, the lines had to be broken in the presentation of the Greek text; the repetition of lines 3 to 6 results from the fact that lines 3 and 4 on one hand, and 5 and 6 on the other, belong together because they consist of groups of superimposed words (cf. p. 191).

Figure 49. O.Claud. inv. 1538+2921. © A. Bülow-Jacobsen 4. This is an AE 3 (= amphore égyptienne 3). On this amphora, see, e.g., Empereur and Picon 1998: 77, or Tomber 1993: 140–41.

(ἑκατοντάρχῃ) vac. δεκ(ουρίωνι) vac. Ἡρακ(λείδῃ) κδ ϛ´· Φαννίῳ̣ [σ]τ̣ρ̣α̣[τ( )] κε̣(ράμιον) α· ἱππεῦσι ϛ̅ κε(ράμια) ε

τείρω(σι) κ̅β̅ κε(ράμια) ια· ἵππῳ δεκο(υρίωνος) κε(ράμια) δ· τεσσαρ(αρίῳ) ´· ὑποτεσσαρ(αρίῳ)´·

ἰατρο(ῖς) β̅ κε(ράμιον) α· σκυτεῦσι β̅ κε(ράμιον) α· κουρ(εῖ) ´· ἐπιμελ(ητῇ) Πανείο(υ) ´· (ἑκατοντάρχου) ´· θησαυροφύλ(ακι) κριθ(ῆς) ´· τηρητ(ῇ) κελλ(ῶν) φαμ(ιλίας) ´·

πλατιαρ(ίῳ) ´· μαγείρ(οις) β̅ κε(ράμιον) α· τηρη(τῇ) σιδ(ηρίων) Καινο(ῦ) Ὑδρ(εύματος) vac. τηρητ(ῇ) ἀνθρ(άκων) ´· τηρητ(ῇ) στάβλω(ν) ´· κτηνοιατρ(ῷ) ´· κιβαρ(ιάτῃ) φαμιλ(ίας) ´·

τηρητ(ῇ) ἄρτ(ων) φαμιλ(ίας) ´· τηρητ(αῖς) κελλῶ(ν) παγαν(ῶν) ἀνδρ(άσι) ζ κε(ράμια) γ´· Οὐμβιδίῳ ἐκ φαμ(ιλίας) ´· ὀνηλάτ(ῃ) (ἑκατοντάρχου) ´· ἁμαξεῖ παγ(ανῷ) ´· Β̣αβ̣ουλλ(ίῳ) vac.

δεκανῷ καμήλ(ων) ´· δρομαδαρ(ίῳ) ´· ὄνοις ε ὑ̣π̣ὸ (ἑκατοντάρχην) κε(ράμια) ι⟦α⟧· ὄνῳ α̅ ὑπ(ὸ) δεκο(υρίωνα) κε(ράμια) β· καμηλ(ίταις) κυρ̣ιακ(οῖς) ἐλαύνο(υσι) ὄνο(υς) vac.

7

8

9

10

11

12

c. 10

η [ σκοπελ( ) κε(ράμια) β β) . [

5 6

]

β ] ξ̣υλοπ( ) β)/ καὶ ἐν κερ(αμίοις)· στρατιώ(ταις) κ̅θ̅ κε(ράμια)

στομω(τηρίῳ) β

θυρο(υρῷ) πραισιδ(ίου) ´· θυρο(υρῷ) ξεν(ιῶν)

ἵππο(ις) ϛ̅ ἱππέω(ν) ἰς ποτισμὸ(ν) κε(ράμια) κδ· ἀναμετρητ(ῇ) μαρμάρ(ων) ´γ´·

vac.

παγ(ανοὶ) κγ φαμ(ιλία) γ Κρηπ(ῖδι) Ἥρας β κε(ράμια) β´

vac.

παγ(ανοὶ) ιγ φαμ(ιλία) γ παγ(ανοὶ) λγ φαμ(ιλία) ϛ παγ(ανοὶ) ιϛ φαμ(ιλία) β Ἄπιδι α κε(ράμια) β Ἥρᾳ γ κε(ράμια) γ Διοσκορί̣ο̣ις α κε(ράμια) γ β)

5 6

παγ(ανοὶ) κ φαμ(ιλία) ε Χρ]ησμοσαρ(άπιδι) ἀσκ(οὶ) β κε(ράμια) α β)

παγ(ανοὶ) λβ φαμ(ιλία) [--] Κρηπ(ῖδι) Με[γάλ- ἀσκ(οὶ) --

παγ(ανοὶ) λϛ φαμ(ιλία) η Μυρισμῷ ἀσκ(οὶ) δ

3 4

[ c. 5 ]

παγ(ανοὶ) ρλζ φαμ(ιλία) [ -παγ(ανοὶ) -- ] φ̣[αμ(ιλία) --] [κα]ὶ̣ ἰ̣ς π(οτισμὸν) ἐν λατομί(αις)· Τραιανῇ ἀσ[κ(οὶ) ]. Μ̣έ̣ϲ̣ῃ̣ ἀσκ(οὶ) vac.

] στομωτ(ηρίῳ) ἀσκ(οὶ) γ· (γίνονται) ἀσκ(οὶ) κα κε(ράμια) β·

3 4

c. 15

[Ἄπ]ιδι χαλ(κεῖ) α̅ κε(ράμια) γ· Ἥρᾳ χαλ(κ-) [ κε(ράμιον) ´ καὶ φυση(ταῖς) καὶ φαμ(ιλίᾳ) ε β̣γ´·

[ date ] [ἰς τὰς] χ̣ρείας χαλ(κέων)· Τρα[ιανῇ χαλ(κεῦσι) --· Μυρισμῷ χαλ(κεῦσι) · Μέσ]ῃ̣ χαλ(κεῦσι) ι̅α̅ ε κε(ράμια) γ· Κρηπῖδι Μεγ̣άλ(ῃ) χαλ(κεῦσι) β̅ []· Χρησμοσαρ(άπιδι) χαλ(κεῦσι) β̣̅ α·

2

1

The organization chart of the personnel of an imperial quarry 183

[κ]ε(ράμια) ξε· Ἀρσινοείτ(αις) ϛ̅ κε(ράμια) γ· Μεμφείτ(αις) γ κε(ράμια) α´· χαλκεῦ(σι) μ̅ κε(ράμια) κ· φυσηταῖς κ̅ε̅ κε(ράμια) η γ´· Φαννίῳ στρατ( ) ἰς πλύσιμ( ) διὰ

Σπινθῆ(ρος) κε(ράμια) β· Ἑρεννίῳ ὑπάγοντ(ι) ἐπὶ κυνηγω̣( ) ζήτη(σιν) ´· Ἰσχυρᾶτ(ι) ἐλθόντ(ι) ἀπὸ Ῥαιμ(α) μετ(ὰ) ἐπιστολ( ) γ´· ἀνδ(ράσιν) ἐκ φαμ(ιλίας)

ὑπάγο(υσι) μετὰ ἁμαξῆ(ς) ι̅β̅ κε(ράμια) δ· ἀνδ(ράσιν) ἐκ φαμ(ιλίας) λοιπ(οῖς) ἀπὸ καταλ(όγου) τπη κε(ράμια) ρκθ γ´.

14

15

16

passim χαλ κε ασκ φαμ πγα 1 μεγαλ χρημοσαρ̅ 2 στομωτ φυσ 4 ]ι̣ ι̣ς) λατομι— κρη) χρημοσαρ̅ 6 κρη) στομ⏑ σκοπελ ο κ — ο ο ο τ ⏑ ⏑ ⏑ ξυλο) 7  δεκ ηρα κερ στρατι 8 τειρ δεκ ιππ ιππε ποτισμ αναμετρη μαρμαρ̅ τεσσαρ̅ υποτεσσαρ̅ 9 ιατρο — λ ο ο δ ο — κουρ επιμε πανει θυρ πραισι θυρ ξεν  θησαυροφυλ κριθ τηρητ κελλ 10 πλατιαρ̅ μαγειρο τηρ σιδ καινο υδρ— τηρητ — — λ τ τ λ τ ⏑ ⏑ ανθρ σταβλ κτηνοιατρ κιβαρ̅ φαμι 11 τηρη αρ φαμι τηρη κελλ παγαν— ανδ ονηλατ ρχ παγ— βαβουλλ 12 δεκαν— καμηλ ο λ κ ο ο δρομαδαρ̅ υ) δεκ καμη κυρια ελαυν ν 13 κατα λ λατομ σκλ εργοδοτ σανσνωτ ομοι⏑ αλλο σοηνιτ σκλ 14 αρσινοειτ τ — τ τ ⏑ τ + μ τ λ δ ο μεμφει χαλκευ στρα πλυσιμ 15 σπινθ υπαγον κυνηγ ζητ ισχυρα ελθον ραι με επιστο αν 16 υπαγ αμαξ ανδρ— λοι) καταλ

καὶ ἀπὸ καταλ(όγου) λατομ(ιῶν)· Ἀλεξαν[δ(ρεῦσι)] σκλη(ρουργοῖς) σι κ[ε](ράμια) ρε· ἐργοδότ(αις)· Σάνσνωτ(ι) κε(ράμιον) α· ὁμοίω(ς) ἄλλο(ις) ϛ̅ κε(ράμια) γ· Σοηνίτ(αις) σκλη(ρουργοῖς) ρ̅λ̅

13

184 Rome in Egypt’s Eastern Desert

The organization chart of the personnel of an imperial quarry

185

Translation5 For the needs of the blacksmiths: (Area)

(Number of Blacksmiths)

for the Traiane [for Myrismos] [for Mes]e for the Large Loading Ramp for Chresmoserapis for Apis for Hera [- -] for the steelworks

[- -] [- -] 11 2 2 1 [- -] [- -] Ø

(Quantity of Water) [- -] [- -] 5 (goatskins), 3 amphoras [- -] 1 (goatskin) 3 amphoras [- -] [- -] 3 goatskins

total

21 goatskins 2 amphoras

[- -]

1/2 amphora

for the bellows men1 and for the familia: 5 persons (or:) for the bellows men and for the familia

2 1/3 5 (goatskins) 2 1/3 (amphoras)

1. i.e., those who operate the bellows. For drinking water in the quarries: (Area)

Pagani

Familia

for the Traiane for Mese for Myrismos for the Large Loading Ramp for Chresmoserapis for Apis for Hera for the Dioskoureia for the loading ramp of Hera for the steelworks

137 [- -] 36 32 20 13 33 16 23 Ø

[- -] [- -] 8 [- -] 5 3 6 2 3 Ø

for the lookout tower(s) […] for the wood sawyers (?)

8 […] 2

(Quantity of Water) [- -] vacat 4 goatskins [- -] 2 goatskins 1 2/3 amph. 1 (goatskin) 2 amph. 3 (goatskins) 3 amph. 1 (goatskin) 3 2/3 amph. 2 (goatskins) 2.5 amph. 2 (goatskins) 2 2/3 amph. […] 2/3 amph.

(Expected Total)1

4 goatskins 2/3 amph. 2 goatskins 1 2/3 amph. 1 goatskin 2.5 amph.2 3 goatskins 3.5 amph. 1 goatskin 3 2/3 amph. 2 goatskins 2.5 amph. 2 2/3 amph.3 2/3 amph.4

1. On the basis of a ration of 1/2 keramion for the pagani and of 1/3 for the familiares, and allowing that the rate of conversion is 5 keramia = 1 goatskin (cf. the metrological notes infra, pp. 208–10). 2. Here we can show that the scribe made a mistake: if we compare the figures of Apis with those of the loading ramp of Hera, where the number of familiares is the same, we can reckon that 10 natives would receive 1 goatskin and 1/2 amphora, which is not plausible: it is clear that the scribe has forgotten to note 1/2 amphora under the heading Apis and that 10 natives receive 1 goatskin. 3. Allowing that these scopelarii receive 1/3 keramion (which suggests that they are not soldiers, but belong to the familia). 4. Supposing a ration of 1/3 keramion. 5. It seems to me clearer to present the information contained in the document in tabular form rather than giving a lineby-line translation.

186

Rome in Egypt’s Eastern Desert (Identity and Number of Recipients)

(Quantity of Water)

for the centurion for the decurion for Herakleides in amphoras for the soldiers, 29 for Fannius, soldier (or strator) for the cavalrymen, 6 for the recruits, 22 for the decurion’s horse for the horses of the cavalrymen, 6, for watering them for the checker of the marbles for the tesserarius for the subtesserarius for the doctors, 2 for the cobblers, 2 for the barber for the sacristan of the sanctuary of Pan  for the doorkeeper of the fort for the doorkeeper of the centurion’s lodging for the guard of the barley granary for the guard of the barracks of the familia for the platearius for the butchers, 2  for the guard of the tools at the New Well for the guard of the (storeroom for?) charcoal for the guard of the stables for the veterinarian for the manager of the supplies of the familia for the guard of the bread of the familia for the guards of the barracks of the natives, 7 for Ummidius, belonging to the familia for the donkey-driver of the centurion for the native carter for Babullius for the dekanos of the camels for the dromadarius for 5 donkeys in the centurion’s transport group for 1 donkey in the decurion’s transport group for the imperial cameleers leading donkeys

vacat vacat vacat 24 1/6 amph. 1 amph. 5 amph. 11 amph. 4 amph. 24 amph. 5/6 amph. 1/2 amph. 1/2 amph. 1 amph. 1 amph. 1/2 amph. 1/2 amph. 1/2 amph. 1/2 amph. 1/2 amph. 1/2 amph. 1/2 amph. 1 amph. vacat 1/2 amph. 1/2 amph. 1/2 amph. 1/2 amph. 1/2 amph. 3 1/2 amph. 1/2 amph. 1/2 amph. 1/2 amph. vacat 1/2 amph. 1/2 amph. 10 amph. 2 amph. vacat

According to the register of quarries: (occupation, geographical origin) for the Alexandrian quarrymen for the foremen: for Sansnos for the other foremen for the quarrymen from Syene for the Arsinoites for the Memphites for the blacksmiths for the bellows men

(number of persons)

(quantity of water)

210

105 amph.

6 130 6 3 40 25

1 amph. 3 amph. 65 amph. 3 amph. 1.5 amph. 20 amph. 8 1/3 amph.

The organization chart of the personnel of an imperial quarry (identity)

(number)

for Fannius, soldier (or strator), for washing, through Spinther for Herennius, leaving for hunting (or looking for hunters) for Ischyras, arrived from Raïma with mail for men from the familia leaving with a wagon for the rest of the men of the familia, according to the register

1 1 12 388

187

(quantities of water) 2 amph. 1/2 amph. 1/3 amph. 4 amph. 129 1/3 amph. 

Notes 1.

[ἰς τὰς] χ̣ρείας χαλ(κέων). I estimate the width of the lacuna at 4 or 5 letters, whence the choice to restore the article. One could also think of resolving χαλ(κείων) (from χαλκεῖον, “forge”), for this part does not take into account the drinking water of the blacksmiths (see the demonstration infra p. 191); it concerns only what was used for the work of the smithy. But that would lead us to resolve the abbreviation in two different ways in the two lines. Τρα[ιανῇ χαλ(κεῦσι) -- · Μυρισμῷ χαλ(κεῦσι) -- · Μέσ]ῃ̣. In the next section, Mese precedes Myrismos; the sequence Traiane-Myrismos-Mese that I restore here appears in the chart inv. 2853, which belongs to the same dossier.

2.

In the lacuna between Hera and the stomoterion, we expect a mention of the Dioskoureia and the loading ramp of Hera (cf. l. 6). However, it does not appear to me that there is enough room for both: did the scribe omit one of them? (γίνονται) ἀσκ(οὶ) κα κε(ράμια) β. In principle, this should be the total amount of water distributed to the chalkeis who, at the smithy, each received 3 keramia of water; and we know that there were 40 of them (line 14). The calculation does not come out quite right no matter what the capacity of the goatskin was (on the ratio goatskin/keramion, see pp. 208–10), but it is possible that of these 40 smiths, some were assigned to the steelworks or were simply ill and remained in the camp. καὶ φυση(ταῖς) καὶ φαμ(ιλίᾳ) ε β̣γ. This information was added as an afterthought, in darker ink. To the difficulty of interpreting these figures is added the ambiguity of the character between them, which has the form of kappa without excluding beta, the hypothesis I have opted for, because a kappa here could only be an abridged form of κεράμιον, but it has no raised epsilon. The fraction seems to indicate that in this case, unlike the situation with the blacksmiths, we are dealing with individual rations of 1/3 which are actually those of the familia and the physetai (although they were pagani) in the fourth section. If ε represents men, we would expect that these five persons received 1 2/3 keramion. In this case, there would be two anomalies: the scribe would have made a mistake in his calculation and forgotten to specify that it was a matter of keramia rather than goatskins, while in this first section of the schedule, the word ἀσκός is sometimes omitted before the amount of water, but this is never the case with its subdivision keramion (the omission could, however, be explained by lack of space). If, on the other hand, ε is the number of goatskins, that would mean that this amount of water (5 goatskins 2 1/3 keramia) is intended for 82 persons (admitting the equation 1 askos = 5 keramia): as well as the 25 physetai listed in the fourth section, 57 familiares would thus work in the smithies. But then why would they not have been taken into account in the second section, which concerns the drinking water for the entire workforce, native or imperial, which was laboring in the quarries and the smithies there? Did the scribe find it easier for his calculations to count separately the drinking water of the two categories of smith workers who were entitled only to 1/3 keramion?

188

Rome in Egypt’s Eastern Desert

4.

[κα]ὶ̣ ἰ̣ς π(οτισμὸν) ἐν λατομί(αις). This resolution fits the remaining traces perfectly and is nicely contrasted with the “needs of the blacksmiths,” or perhaps “of the smithies,” which consisted of water to fill the tempering basins. The restoration of [κα]ί̣ is supported by the entries καὶ ἐν κερ(αμίοις) (l. 7) and καὶ ἀπὸ καταλ(όγου) λατομ(ιῶν) (l. 13). We find ἰς ποτισμόν again in l. 8, with reference to the horses of the cavalrymen.

6.

στομω(τηρίῳ). Unlike the scribe’s usual custom, the suspended omega is reduced to a nearly flat stroke; the same phenomenon occurs with δεκανῷ in l. 12. On the steelworks, see infra p. 193. σκοπελ( ). The skopeloi of Mons Claudianus belonged, according to D. Peacock, to a telephonic or teleoptic network (Peacock and Maxfield 1997: 254 f.). On the papyrological attestations of skopeloi and scopelarii, see Cuvigny (ed.) 2006: 216–28. ξυλοπ( ). Only the top of the ξ is visible, but it is paleographically satisfactory. Probably the same word appears in the schedule inv. 2853.5, where I read ξ̣υ̣λ̣ο̣[πρίσ]τ̣α̣ις (add. lex.).

7.

ἐν κερ(αμίοις). This resolution is confirmed by the schedule inv. 2847. [σ]τ̣ρ̣α̣[τ]( ). The word figures, similarly abridged, in line 14. The most prudent resolution is [σ]τ̣ρ̣α̣[τ](ιώτῃ), but [σ]τ̣ρ̣ά̣[τ](ορι) cannot be excluded (see infra, p. 202).

9.

θησαυροφύλαξ κριθῆς. The thesaurophylax was not, unlike the τηρηταί, a simple employee of the security force, but an administrator who kept the accounts of the deposits and withdrawals in the magazine that he was in charge of (O.Claud. inv. 555 and 7896 attest him).

10.

τηρητ(ῇ) ἀνθρ(άκων). One could just as well resolve ἀνθρ(ακῶνος): “the guard of the storehouse of charcoal.” An ἀνθρακών is attested in an ostracon from the south dump (O.Claud. IV 742). στάβλω(ν). The reference is not to the enclosure for transport animals visible today to the west of the fortified village, as this dates at earliest to the reign of Antoninus.

11.

Οὐμβιδίῳ ἐκ φαμ(ιλίας). On this name, cf. Syme 1968. Does the name of this man go back, even if indirectly, to Ummidius Quadratus, who was governor of Syria from 50 to 60? Β̣αβ̣ουλλ(ίῳ). One might also be tempted to read κ̣αὶ̣ Ἰ̣ουλλ( ), but Β̣αβ̣ουλλ( ) is preferable, to judge by the only surviving parallel passage in another schedule fragment (inv. 3229), where I read ] φαμιλ( ) Β̣αβ̣ουλλι( ) ε[. This sequence suggests that Babullius probably is not the name of the dekanos of the camels; rather, in my view, it is an isolated personal name, next to which no water ration has been entered. The gentilicium Babullius is not otherwise attested in Egypt, but an inscription of Paestum (ΑΕ 1975, 251) records a P. Babullius Sallu[--] who was military tribune of the legio XXII Deiotariana (stationed in Egypt at least until 121) and pro[curator Aug(usti) / missus ad agro]s dividendos veteranis qui su[nt deducti / ---]a eius in colonia Flavia Paesti (these responsibilities at Paestum must date to 71).

12.

δεκανῷ καμήλ(ων). I have rejected a possible resolution δεκανῷ καμηλ(ιτῶν) on the strength of an unpublished ostracon from Umm Balad, a letter written to a centurion by a δεκανὸς καμήλω(ν). At Mons Claudianus, the dekanoi are mainly involved in the transport of water.

The organization chart of the personnel of an imperial quarry

189

ὄνοις ε ὑ̣π̣ὸ (ἑκατοντάρχην) κε(ράμια) ι⟦α⟧. The number of donkeys (ε) was written in place of another number that had been erased; similarly, the number of keramia (ι) has been added over a number that ended in α. καμηλ(ίταις) κυρ̣ιακ(οῖς) ἐλαύνο(υσι) ὄνο(υς). The resolution rests on the schedules inv. 2853.12 (κα[μ]η̣λ̣ί̣τ̣α̣ις κυρ̣ιακ(οῖς) vac. ) and inv. 2918.11 (καμηλείταις κ̣[υριακοῖς?). In O.Claud., the adjective κυριακός is normally applied to animals or to objects (goatskins, water), not to human beings: we would expect that these cameleers would be described as ἐκ φαμιλίας. 13.

The ergodotes Sansnos (see above, pp. 160–63) is well attested by several tituli picti bearing his name, a chit accompanying a basket of bread sent to him, and a letter in which he asks to have a goatskin (of water?) sent to the quarry of Apollo (O.Claud. IV 819).

14.

χαλκευ ̅. Here, the supralinear stroke serves both to mark an abbreviation for χαλκεῦ(σι) and as a number stroke over μ. φυσηταῖς. There is a superfluous horizontal stroke above αι.

14–15. ἰς πλύσιμ( ) διὰ Σπινθῆ(ρος). Are we dealing with πλύσιμ(α) (laundry to be washed: cf. P.Oxy. L 3599.7n.) or a πλύσιμ(ον), “washery”? For a third possible solution, see p. 202. Fannius must be responsible for the operation, thus the recipient of the water in the nominative, but the work is carried out by Spinther (a slave?). 15.

Σπινθῆ(ρος). On the personal name Σπινθήρ and its accentuation, cf. O. Masson, ZPE 91 (1992) 107–11. κυνηγω̣( ) ζήτη(σιν). I initially thought of resolving κυνηγ(ετικὴν) ζήτη(σιν) (on hunting in the ostraca from Mons Claudianus, see O.Claud. II 357 introduction). Probably the hunters’ activity could be described as a ζήτησις (e.g., Pollux, On. 5.9.3: ἐρεῖς δ’ ἐπὶ τοῦ κυνηγέτου ζητητὴς θηρίων); but why in an administrative text would one have used this periphrasis instead of “the hunt,” when it is not elsewhere attested in Greek? Moreover, the scribe systematically uses the raised sign ⏑ to represent omega: the beginning of the word thus should be κυνηγω( ). Would Herennius thus have set out not “on a hunting mission” but in search of hunters who had not returned, ἐπὶ κυνηγῶ(ν) ζήτη(σιν)? In an unpublished letter found at Umm Balad (O.KaLa. inv. 819), the author (a curator praesidii?) informs his correspondent that he has sent two soldiers to look for a man who has disappeared: ἔπεμσα δὲ καὶ δύο στρατιώτας ἐπὶ ζήτησιν αὐτοῦ. ἐλθόντ(ι) ἀπὸ Ῥαϊμ(α) μετ(ὰ) ἐπιστολ( ). μετ(ά) because the tau is raised; ἐπιστολ(ῆς) or ἐπιστολ(ῶν). The style is that of military reports in Latin: cf., e.g., Rom.Mil.Rec. 47, ii.7, reversi … cum epistulis.

B. General commentary I. Date The name of the quarry Traiane provides a terminus post quem, and several converging indications allow us to place the little archive of water distribution schedules to which O.Claud. inv. 1538 belongs in the reign of Trajan:

190 (1) (2)

(3)

(4)

Rome in Egypt’s Eastern Desert One of the schedules was found in a layer that has delivered precisely dated texts, to wit, two orders for delivery dated to the 13th year of Trajan (110); The first three words of line 7 mention, to all appearances, the three most important residents of the metallon: these are, in order, the centurion, the decurion, and Herakleides. Who is Herakleides? Not a soldier, for in this place we would have to be dealing with an officer, and his rank would have been mentioned as it is for the others. It is very tempting to recognize in this person the architect Herakleides, who signed a block of stone from Claudianus found at Rome and datable, by the mention of the prefect Rutilius Lupus, to the years 113–117; 6 we might even narrow this chronological fork to June 113 to July 115, the latter date marking the start of the great Jewish revolt, which completely paralyzed the country. In another schedule of this series (inv. 3069), the mentions of the centurion, of the decurion, and of Herakleides are followed by that of a certain Ἀπολ( ), and it is tempting to resolve the abbreviation as Ἀπολ(λώνιος) and identify the man with another architect known at Mons Claudianus, the Alexandrian Apollonios son of Ammonios, the dedicant of the altar I.Pan 38 under the reign of Trajan (c. 110–114: cf. O.Claud. I 15–18). On the day when the present schedule was written, Apollonios must have been away. Among the individuals employed on specialized tasks and whose juridical status is unknown to us there is an epimeletes of the Paneion. This sanctuary of Pan, of which no trace was found on the ground, is the only sanctuary mentioned in the archive of the water schedules: there is every reason to think that the small temple of Zeus Helios Great Sarapis which still dominates the settlement in Wadi Umm Husayn did not yet exist, or at least as such; but we know from the inscription on its lintel that it was dedicated in 118 by the imperial slave Epaphroditos, who held the concession for the mines (I.Pan 42), and who says that he had it constructed (κατεσκεύασεν). It is probable that this dedication is the epigraphic memorial of ceremonies held to greet the renewal of the working of the imperial quarries under the new emperor, in the aftermath of the victory over the Jews. It is tempting to think that the former Paneion disappeared at the time of the reorganization of the religious complex. Pan of the desert had been the god of the explorers and pioneers: different times, different gods.7

O.Claud. inv. 1538 corresponds to the period of building the Forum of Trajan and the adjacent basilica Ulpia, which necessitated a large number of granodiorite columns from Claudianus: begun in 107, these monuments were inaugurated in 112.

II. Structure of the document The schedule must have begun with a date to a day, but only one text of the group has preserved this;8 in this case it was Hathyr 20, i.e., 16 or 17 November, thus after the season of high heat; the modest level of the water rations in our document suggests that it might come from the same period. Then, our ostracon consists of five sections: 1. Lines 1–2. Water for the “needs of the blacksmiths” (or “of the smithies”), if my resolution is right. The amount of water obviously depends on the number of blacksmiths (χαλκεῖς), which is indicated for each quarry. The amount assigned to their assistants (bellows operators and familiares) seems to be 6. IG XIV 2421 (= IGRR I 550): ἐπὶ Λούπωι ἐπάρχωι Αἰγύπτου διὰ Ἡρακλείδου ἀρχιτέκτ(ονος). 7. On the Romans’ loss of interest with respect to Pan-Min after the first century, see Chapter 29. 8. O.Claud. inv. 3666.

The organization chart of the personnel of an imperial quarry

191

listed at once for all of the quarries at the end of the section and appears, according to the color of the ink, to have been added at the second stage (cf. comm. ad 2). 2. Lines 3–6. Distribution of drinking water, similarly presented area by area; within these workmen, the scribe distinguishes the pagani (“natives”) and the familia (“imperial”). The scribe resorts here to a particular style of presentation: the name of each quarry is followed by the amount of water allocated to it, and above is placed the number of “native” and “imperial” workers who work there. Initially I began my analysis with the preconceived notion that the water in the first section represented not only the filling of the tempering basins for the blacksmiths, but also their drinking water. As a result, I thought that the pagani, in the second section, represented only the sklerourgoi (otherwise, the blacksmiths would have received water three times a day!): since we know from the third section that the “register of the quarries” counted 349 sklerourgoi, I concluded that the Mese quarry, the only area where the number of sklerourgoi was lost in a lacuna, must have employed 39 of them on that day (349 – 310 = 39). Then I noticed that the number of blacksmiths in one quarry, the name of which had been lost, was very high: eleven blacksmiths, a very high number in comparison with two at the Great Loading Ramp, two of Chresmoserapis, and only one at Apis. Then I understood that the names of Mese and Myrismos (which are found in this order in line 4) had been reversed in line 1 and, in fact, the final eta of Μέσ]ῃ was actually present before χαλ(κεῦσι) ι̅α̅. As a result, Mese was a very large area, like the Traiane quarry, and it occupied many more than 39 sklerourgoi: pagani thus referred to the native workmen in the second section, taking all specialties together; this is, after all, the usual sense of paganus at Mons Claudianus. It is then possible to calculate that Mese occupied 421 – 310 = 111 pagani. All the same, this calculation may not be absolutely precise: in the two sections, the scribe does not indicate how many men worked in the steelworks, the stomoterion;9 the form indeed did not provide for recording this number, since στομωτ(ηρίῳ) in line 2 is followed directly by the number of goatskins. This omission may perhaps be explained by the fact that unlike the other sites, the stomoterion, where the first quality steel was manufactured for the entire metallon, functioned with an unchanging complement. It is possible that, of the 421 pagani, some blacksmiths were assigned there (unless a superspecialized staff, foreign to the local workforce, was employed there). If the quantity of water is calculated in the same way for the stomoterion as for the smithies of the quarries, the number of chalkeis will have been 6. In that case, the following table would present the overall situation: Table 11.1. Number of blacksmiths and quarrymen in each work area (Area)

Pagani

Blacksmiths (chalkeis)

Traiane Mese Myrismos Large Loading Ramp Chresmoserapis Apis Hera Dioskoureia Loading ramp of Hera steelworks

137 [105 or 111]1 36 32 20 13 33 16 23 [6?]

[c. 13] 11 [2 or 3] 2 2 1 [2 or 3] ? ? [6?]

Number Deduced for sklerourgoi 94 or 100 33 or 34 30 18 12 31 or 30

Ratio chalkeis/ sklerourgoi if 13 chalkeis, 1:9.5 1:8.5 or 1:9 1:16.5 or 1:11.3 1:15 1:9 1:12 1:15.5 or 1:10

1. According to whether we do or do not suppose the presence of 6 pagani in the steelworks. 9. It is equally probable that, in the differences between the numbers of the “register of quarries” and those of the second section that represent the pagani really present in the area, we must reckon with some who were sick.

192

Rome in Egypt’s Eastern Desert

There seems to have been an effort to avoid letting the number of sklerourgoi exceed 15 for each blacksmith. 3. Lines 7–12. Without a heading. Distribution of water to specific individuals, both humans and animals. It is in this section that officers and other soldiers are mentioned, along with some technicians, artisans, and guards for various places. 4. Lines 13–14, introduced by the formula καὶ ἀπὸ καταλ(όγου) λατομ(ιῶν). Distribution of water to pagani workers; quarrymen and blacksmiths continue to be distinguished, but the place of work is no longer specified; on the other hand, the quarrymen (σκληρουργοί), but not the blacksmiths, are classified by geographic origin. This section stops after the mention of the bellows men (φυσηταί). 5. Lines 14–16. Catch-all section, without a heading, in which the scribe introduces, in darker ink, unforeseen information that reached him at the last moment: Fannius, already named in the third section, needed additional water, a messenger had arrived from a satellite post, a soldier on mission in the desert took a small supply with him, twelve employees of the familia left with a wagon, which allows the scribe to establish the precise number of employees of the familia remaining on site in the metallon, which ends the schedule.

III. Topography O.Claud. inv. 1538 mentions a certain number of places, buildings, areas, and placenames located inside the praesidium, in its immediate environs, as well as in the neighboring mountains and valleys. Only a few of these have been identified on the ground. Some of them need comment. Raïma The most distant entity is Raïma. It is the praesidium with which exchanges of letters were the most frequent; it should therefore be close to Mons Claudianus and have represented probably the first stage reached after leaving the metallon in the direction of the Nile valley. There was a well there, and the site specialized in growing vegetables and herbs. It is probably to be identified with Abu Zawal; on its name see Chapter 1, pp. 54 f. Krepis Most of the work areas mentioned in inv. 1538  are quarries (λατομίαι), with the exception of Dioskoureia, the name of a well,10 and probably also two sites called krepides (Κρηπὶς μεγάλη and Κρηπὶς Ἥρας). How are we to understand krepis (which I have rendered here as “loading ramp”)? The word can refer to “tout ‘socle’, quel qu’il soit.”11 From this one is led to think of loading platforms, a type of rectangular outcrops from which the blocks brought down from the mountain could be pushed onto wagons; but the term could also refer to the supporting walls for some haulage tracks12 (which would have been 10. O.Claud. inv. 490. Perhaps it was under construction at that time. 11. On the variety of architectural meanings of this term, see Hellmann 1992: 242–3. On its meaning at Mons Claudianus, see also Chapter 1, p. 42. 12. For the meaning of supporting wall, see Vallois 1944: 217–18 and n.1.

The organization chart of the personnel of an imperial quarry

193

under construction) or even work zones in which the blocks extracted from the quarry were shaped:13 thus the krepis of Hera would be the zone in which the blocks extracted from the Hera quarry were roughly shaped, while perhaps the “large krepis” would have served several quarries. Identified quarries Just one of the quarries mentioned in the ostracon could be identified with a place, thanks to an inscription found on the spot: this is that of Myrismos (its name is sometimes given as a fixed genitive Μυρισμοῦ, sometimes declined); it corresponds to quarry 22 of D. Peacock.14 One of the two quarries nearest to Myrismou is no. 7, which is very small and was, according to a graffito, called “Apollo Epikomos”;15 it is probably the quarry of Apollo for which the ergodotes Sansnos requests a goatskin of water (cf. comm. ad 13). It is to be noted that in the schedule inv. 2853, the sequence of quarries is Τραιανῇ vac. Μυρισμοῦ vac. Μέσῃ [--- Ἀ]πόλλωνος vac. The other quarry relatively near Myrismou, no. 129, is also very small; like that of Apollo, it dates to the same period of activity as our ostracon, because the epigraphic signature of the architect Herakleides has been found there.16 This document is far from mentioning all the quarries cited in the ostraca from Mons Claudianus with a supposedly Trajanic date. The steelworks Στομωτήριον is an addendum lexicis, but its meaning can easily be deduced from the verb στομόω, from which it derives (“temper iron in a way to turn it into steel,” cf. O.Claud. IV, pp. 257–9). If there was a smithy in each quarry to constantly reforge the tools of the sklerourgoi, the stomoterion was a workshop which made a special steel, called στόμωμα in the ostraca, for the entire metallon. It is interesting to observe that the metallurgists of the Renaissance did not readily find a tempering process that allowed them to obtain a steel sufficiently hard to use in sculpting porphyry with a chisel,17 although the Romans do not seem to have found the hardness of this material a problem. The stomoma is likely to have been this steel especially suited for working with porphyry, which was at the same period being extracted from Porphyrites, not far from Mons Claudianus. The hostelry We do not know if the ξεν(ίαι) of the centurion were located inside the praesidium or in the builtup zone located to its northeast, what the British archaeologists called the “residential and bathhouse complex.”18 The word is resolved in the plural, on the basis of two ostraca from the same dossier of water schedules, in which it is not abbreviated.19 P.Dura 107 (Rom.Mil.Rec. 15), the service roster of an auxiliary cohort, gives the Latin counterpart of this expression: several hospitia are mentioned in this text, notably that of a centurion praepositus and that of a decurion (lines 14 and 22);20 the editors 13. On these zones, see Koželj 1987: 22. 14. See the map of quarries in Peacock and Maxfield 1997: 14. 15. Peacock and Maxfield 1997: 178. 16. Peacock and Maxfield 1997: 189 and 221; O.Claud. I, p. 48. 17. Before this discovery, the stonecutters had to use abrasive methods: cf. Cuvigny and Vagenheim 2005: 115 f. 18. Peacock and Maxfield 1997: 118–34. 19. In inv. 2853.5: ξενίαις (ἑκατοντάρχου); in inv. 2918.7: θυρουρῷ ξενιῶν (ἑκατοντάρχου). 20. Briefly commented on by Davies 1989: 54.

194

Rome in Egypt’s Eastern Desert

do not offer any clear view on the nature of these hospitia: they limit themselves to remarking that in a military context hospitium refers to the quartering of a soldier on the resident, and they translate “at the quarters (billet) of the centurio.” At Mons Claudianus, the decurion does not seem to have disposed of a hospitium, or else he was not important enough to enjoy the services of a doorkeeper. The semantic field of ξενία overlaps with that of hospitium: alongside the first meaning, “hospitality,” a secondary and more concrete meaning of “lodging for guests” developed, something first recognized by Wilcken (P.Brem. 15.4n.); even so, he did not realize that from “lodging for guests” ξενία came also to refer to the quartering of a soldier: he did indeed see that the ξενία to which the decurion invites the recipient in his invitation card P.Oxy. IV 74721 was not a festival, as the first editors had taken it (“to his party”), but he went astray in thinking that it was a reception room (Festraum);22 the decurion will receive his guests in the lodging assigned to him by the army. This is also how we must understand ξενία in P.Mich. VIII 473.12–13, the meaning of which did not appear clearly to G. Husson: in this letter, addressed to the veteran Claudius Tiberianus, a woman close to him (his sister?) recounts her visit to her son Saturnilus: εἰσελθόντες ε̣[ἰ]ς̣ τὴν ξενίαν Σατορνείλου; this Saturnilus, a consumer of ἐπιμήνια, is in all likelihood a soldier and, as such, was lodged in a ξενία where his family could come visit him. The word appears elsewhere in O.Claud., always in the singular, whether it indicates the destination of a goatskin in a dossier of orders to deliver water, constitutes an entry in the service rosters,23 or refers to a particular lodging (inv. 7316: ἰς τὴν ξενίαν α⟨ὐ⟩τοῦ; inv. 7424: καὶ περὶ τῆς ξενείας μου). Ordinary people were housed in groups in the kellai (see infra), while some more privileged people lived in a ξενία.24 In this case we are not dealing with quartering on a resident; from this we may deduce that even in a military context ξενία (and hospitium?) do not necessarily imply that housing had been requisitioned from a private citizen. Could the officers’ apartments in a camp (apart, to be sure, from the praetorium) be called ξενία/hospitium? I cannot say. If this is not the case, we must remember that the praesidium of Mons Claudianus was not properly speaking a military camp, and the centurion did not necessarily live inside its walls; we can imagine that the ξενία in this case refers to the hostel available for visiting guests, on the model of the village ξενία, intended for traveling officials, in P.Oxy. XVI 1853. This meaning of hostel for officials is recognized in the TLL, s.v. hospitium, II, b, β (de hospitio publico magistratuum i.q. praetorium). Does the plural ξενίαι specific to the dossier of water schedules mean only that the centurion disposed of a suite, or that this official apartment allowed him to host important guests, such as the decurion? Although Valerie Maxfield prefers to think that the centurion was housed inside the praesidium, ξενία in the plural appears to me to be an adequate term to refer to the external complex, which she describes as follows: “An alternative interpretation of the complex might be to see it as a mansio, a guesthouse where official visitors could be housed. The series of small, roughly equal-sized rooms each with its own independent entrance, would be highly appropriate for use as bedrooms/guest rooms, and the integration of this accommodation with a small bath suite bears the comparison with so-called mansiones elsewhere in the Roman Empire.”25

21. Καλεῖ σε ὁ (δεκάδαρχος) εἰς τὴν ξενίαν ἑαυτοῦ τῇ ϛ Καλάνδαις ἀπὸ ὥρ(ας) η. 22. He is followed in this by G. Husson, who provides a detailed account of the attestations of the concrete meaning of ξενία (Husson 1983: 178–80). 23. One or two men assigned to the hostel, ξενίᾳ: inv. 1121, 1132. 24. But in the Lausiac History of Palladius, ξενία/ξενίδιον and κέλλα are synonymous, both referring to the cell of an anchorite (Husson 1983: 178). 25. Peacock and Maxfield 1997: 134.

The organization chart of the personnel of an imperial quarry

195

The cells of ordinary people As to the workmen, whether pagani or familiares, they lived in the kellai (κέλλαι παγανῶν, κέλλαι φαμιλίας). On κέλλα (from the Latin cella), see Husson 1983: 136–42. When Geneviève Husson was writing her book, only the kellai of the Nile valley countryside were known, but their definition is perfectly suitable for the workers’ accommodations in the desert: a kella is either a modest independent construction consisting normally of a single room, or a room inside a house; it may serve as a storeroom or rudimentary lodging for humble folks (for example, agricultural workers on an estate: cf. the string of kellai enumerated by number in P.Mich. XI 620;26 cf. also, in the Roman world, the perfect parallel found in Cato, R.R. 14: cellas familiae). This description fits exactly the rows of small independent rooms located in the two fortified villages of the metallon. At Mons Claudianus, the kellai were not numbered but named after one of their occupants: a small number of lists of names headed “kella of soand-so” give an idea of the number of occupants. For the pagani, two kellai housed four men, and one six; two kellai of the familia (identified by the names) were each occupied by five men.

IV. Number, status, and position of the recipients 1. The pagani This term was borrowed by the administration of the quarries in the Eastern Desert from military jargon.27 I have long translated it “civilians” (including in the French version of this chapter), but the translation “native” seems to me now more adequate. Paganus means not only “civilian,” in contrast to “soldier,” but also refers to notions of “free” (as opposed to slave), of “ordinary people” (in contrast to notables), and of “native, indigenous” (as opposed to, at Mons Claudianus, the soldiers and the familia, who often came from lands outside Egypt). From our schedules we learn that the pagani were listed in a register called variously “register of the quarries,” κατάλογος λατομ(ιῶν) (line 13), or “register of natives,” κατάλογος παγανῶν.28 I have excluded here the resolution κατάλογος λατόμ(ων), which one might also propose, because the occupational term λατόμος is totally foreign to the lexicon of the ostraca produced at the quarries of the Eastern Desert. Pagani worked only in the quarries, where the great majority of them carried on the trade of σκληρουργοί, while a minority worked as χαλκεῖς and φυσηταί in the quarry smithies. The occupational term σκληρουργός, literally “worker on hard,” is clearly contrasted in our schedules to these other two trades of the smithy and probably refers to the techniques of extraction and to those of shaping the blocks;29 the sklerourgoi are always by far the most numerous among the team at work in a quarry. They form several groups of unequal size, defined by their geographical origin: 210 Alexandrians, 130 Syenites, 6 Arsinoites, and 3 Memphites, or a total of 349 quarrymen. Leaving aside the remarkable disparity between the first two geographical groups and the second two, this division raises several questions. 26. These may be compared to the numbered rooms in the legionary camp of Vindonissa (M. A. Speidel, Die römischen Schreibtafeln von Vindonissa [Baden-Dättwil 1996] 38). 27. See my remarks in O.Claud. III, pp. 11 f. 28. Thus in the schedule O.Claud. inv. 2853.13. 29. Cf. O.Claud. II 287–288: a request is made for a sklerourgos to come and regroove a millstone for grain that had become smooth, and thus useless. Jean-Michel Carrié has gathered the following testimony from an old worker in the granite quarries of Elba: he had worked there since childhood and had been employed in both extraction and in cutting.

196

Rome in Egypt’s Eastern Desert

If the presence of quarrymen from Syene is predictable, given their (relative) geographical proximity and above all their expertise in dealing with hard materials, that of Alexandrians is less obvious, since the stone quarried in the region of Alexandria was a soft limestone.30 And even if being accustomed to a particular quality of material was not an insurmountable obstacle,31 why so many quarrymen from Alexandria when it would have been easy to recruit from nearer, in the numerous quarries of sandstone and limestone in Middle and Upper Egypt? Was there at Alexandria a work force accustomed if not to the extraction, at least to the shaping of hard materials? While recognizing that the existence of sculptors’ workshops dealing with the stone from Wadi al-Hammamat is not certainly attested at Alexandria, R. Belli Pasqua considers that Alexandria is the most probable location for the finishing of sculptures in classical style executed in an Egyptian stone: “A favore di Alessandria è da considerare il fatto che, trattandosi di un materiale egiziano, sarebbe logico cercane le officine in un luogo in cui esistessero artigiani capaci di lavorarlo. È superfluo sottolineare ancora una volta che la lavorazione della grovacca32 ha una lunga tradizione nel mondo egizio, tuttavia la produzione egizia si discosta, quanto a modi formali ed a concezione stilistica, dall’arte greca. Ciò porta alla necessità di attribuire la produzione dei manufatti di età romana ad artigiani che possiedono la capacità tecnica egiziana, ma che si sono formati alla tradizione formale e stilistica greca. E tanto piú questo era possibile solo ad Alessandria.”33 The fact that the architektones of Mons Claudianus, that is, the chief engineers, were perhaps Alexandrians34 might also explain the presence of workers from the same origin. Moreover, the onomastics of the Syenites, known from other ostraca, notably those of the Antonine period, when the Alexandrian/Syenite distinction was maintained, offers a certain number of names characteristic of Edfu and of Thebes but, curiously, no distinctively Syenite name;35 that observation leads to the question whether the category “Syenite” was not a geographical fiction, encompassing in reality quarrymen who came from all of Upper Egypt, and if the same might not be true of the Alexandrians. But the mention of three Memphites and six Arsinoites, which points on the contrary to care taken in being precise about the geographical origin of the quarrymen, is not easily reconciled with this hypothesis. Apparently, it was the profession rather than geographical origin that held together the much smaller group of the blacksmiths. This difference is partly confirmed by the contemporaneous tituli picti that indicate to which association (πλῆθος) the inscribed amphora belonged; the associations mentioned in this series are respectively those of the sklerourgoi, the Syenites, the Syenite sklerourgoi, and the blacksmiths; but there is never any question of a plethos of the Alexandrians. The O.Claud. show that the χαλκεῖς were regularly assisted by a temperer (φαρμαξάριος) and by a bellows operator (φυσητής), and sometimes also by a hammerer (σφυροκόπος); this structure, verified in several cases of quarries where we know the number of workers, implies that the number of chalkeis and of physetai should be the same, which is far from being the case: for 40 chalkeis, there are only 25 physetai (line 14). Like the chalkeis, the physetai are pagani, but their lower water ration shows that they 30. On the quarries of Alexandria, see Röder 1967. 31. In fact, the latomoi at stake in the letter SB XVIII 13881 = P.Petrie Kleon 51 (Fayyum, 256 BC) are able to work with both types: the dekatarchoi of one of the two teams protest only against favoritism on the part of the ergodioktes, who makes them do all of the work needed on hard stone, while keeping the softer material for his own men. 32. Modern people often but mistakenly refer to the schisty sandstone of Wadi al-Hammamat as “Grauwacke” or “graywacke” or even “basalt.” 33. Belli Pasqua 1995: 42–43. Ibid., 44: it is equally possible that the Egyptian materials were worked in Roman workshops, which would perhaps have imported Egyptian workers; these two hypotheses are not mutually exclusive. 34. This is certainly known for Apollonios (I.Pan 38). 35. These are the conclusions reached by W. Clarysse, who was kind enough to examine a list of Egyptian names used in the ostraca of Mons Claudianus.

The organization chart of the personnel of an imperial quarry

197

formed a lower stratum within the native community. The schedules do not mention the pharmaxarioi: other texts suggest that this profession was a specialty of the familia: the pharmaxarioi, like the sphyrokopoi and the aciscularii, which are also absent from the schedules,36 are probably accounted for as familia. Among the pagani, finally, are mentioned also the seven foremen, the ergodotai;37 one final addition is the native wagoneer mentioned in the third section (line 11), who brings the total of pagani to 422. The south dump also produced lists of manpower in the individual quarries. I give below, for the sake of comparison, the numbers concerning the quarries mentioned in O.Claud. inv. 1538. It is to be observed that these lists and the schedules do not use quite the same categories: the distinction of status pagani/familia is made only once in the lists retained here for our purpose; on the other hand, the lists are more precise in specifying the occupations; they introduce as well a numerous group, that of the ἐργάται (“workers”), which seems to represent the non-specialist workmen of the familia who are at times present in large numbers in the quarries (when there was a large block to be shifted?). By comparing the lists and the schedules, several principles or tendencies nonetheless emerge: (1) The sklerourgoi are always pagani, the group of which they constitute the largest contingent; also among the pagani are the chalkeis, the physetai and, certainly, the ergodotai; (2) the guards (teretai) belong instead to the familia;38 (3) although the familia certainly takes care of the unskilled tasks, as it seems, some specialized techniques are also practiced by the employees of the familia: there are native and imperial sphyrokopoi (one of each status in O.Claud. IV 649); the pharmaxarioi seem instead to belong to the familia (thus all of those employed at Myrismou in O.Claud. IV 649). Table 11.2. Staffing of four quarries on different days O.Claud. Inv. 1538 Traiane Mese

Staffing of Individual Quarries

137 pagani, of whom c. 124 are sklerourgoi

O.Claud. IV 653: 65 sklerourgoi,1 the remainder in a lacuna

c. 105 or c. 111 pagani, of whom 11 are chalkeis

O.Claud. IV 644: 49 men total

36 pagani 8 familiares 2 or 3 chalkeis

O.Claud. IV 647: 45 men total (of whom 30 sklerourgoi, 3 chalkeis, and 2 ergodotai) O.Claud. IV 648: 89 men total

Myrismou

Chresmoserapis

O.Claud. IV 649: at least 109 men total, of whom 42 are pagani (35 sklerourgoi, 3 chalkeis, 2 physetai, 1 sphyrokopos, 1 ergodotes); the main body of the familia consists of 59 ergatai, to whom are added some teretai, 3 pharmaxarioi, 1 sphyrokopos, etc. 20 pagani, of whom 2 are chalkeis; 8 familiares

O.Claud. IV 657: 31 men total, of whom 25 are sklerourgoi, 2 chalkeis, 2 physetai, 2 pharmaxarioi

1. And not 68 as in the edition.

36. Except for the mention of aciscularii in O.Claud. inv. 3069.7. 37. See Chapter 8. 38. This is the case in O.Claud. IV 649.

198

Rome in Egypt’s Eastern Desert

The very marked variations of the numbers are explained by the nature of the work in the quarries, which demanded at each stage a different number of experts in each task or specialty. It was probably the task of the architects to coordinate skillfully, in cooperation with the ergodotai, the deployment of a fixed quantity of manpower in the different areas, which might from one day to another have very different needs. 2. The familia This was, along with the pagani, the other large community at Mons Claudianus.39 Its composition, the legal status of its members, and the means of its recruitment remain a riddle. I have shown elsewhere that it would be too simple to suppose that this familia Caesaris was composed of slaves belonging to the emperor; this is why I prefer to speak of imperial employees. The link with the emperor is corroborated by the use in the ostraca from Mons Claudianus of the adjective κυριακός to describe objects and animals relating to the familia (but perhaps also to the soldiers). If we add up the 388 familiares inscribed on the register (line 16), the twelve who leave the metallon with a wagon, an individual named Ummidius, whose membership in the familia is explicit, and the man in charge of the provisions of the familia, who probably belongs to the same body,40 we reach a total number of 402 imperial employees who receive water distributions in O.Claud. inv. 1538. It is possible that Ummidius and the kibariates should be reckoned separately and that the amount of unnamed workers assigned to Mons Claudianus was 400 men. It is striking to observe that if the pagani (422) and the familia (402) are almost equal in number, seven guards were needed to watch the lodgings of the pagani, while just one suffices for those of the familia: does that mean that the former were given spacious lodgings while the latter were piled up in their barracks? Probably not: the rates of occupation of their respective kellai were the same (cf. p. 195). But a difference in the organization of the storage of provisions explains the difference in the number of guards of the kellai: our schedules mention a guard for the bread of the familia, who has no counterpart for the pagani. Now we know that the latter received their bread individually each month, certainly under Antoninus, but probably already under Trajan. It had been made affectionately for each of them by members of their family, who remained in the Nile valley, as we know from numerous chits with their names on them, from both periods, allowing them to check the number of loaves received.41 It is likely that their provisions, rather than being stored centrally in a single place, were kept in their living spaces, which therefore required a more attentive surveillance. O.Claud. inv. 1538 is the only one of the schedules to specify the number of imperial employees assigned to specific work areas (quarries and other sites). According to the numbers preserved in it, the number of imperial employees in each area was about a quarter of the number of pagani; about a hundred familiares must therefore have been working in these areas. We have already seen what their roles there might be: temperers or hammerers assisting the blacksmiths, guards, aciscularii, or simply laborers. It is not easy to imagine what the other 300 were doing. The schedules themselves do not give any precise indication of the specific occupations of the familiares, with the noteworthy exception of those twelve men who left the metallon with a wagon (line 16): it is evident that this involved the transportation of a block, as we can see from a parallel situation evoked in a letter announcing the arrival of a gigantic cart with twelve wheels and nine servants, all members of the familia (O.Claud. IV 871). Otherwise, the O.Claud. are rather sparing in information on this question; at least they give the impression 39. On the familia at Mons Claudianus, see O.Claud. III, chap. II. 40. This was at least the situation under Antoninus. 41. Cf. Chapters 9 and 14.

The organization chart of the personnel of an imperial quarry

199

that the occupation that required the largest number of familiares was the transportation of goatskins full of water (ἀσκοφορία). The askophoria apparently consisted of loading and unloading the donkeys and camels who provided a shuttle between the wells and the cisterns. But no ostracon mentions more than about forty porters of goatskins (ἀσκοφόροι). A conjecture of D. Peacock may allow us to grasp the solution: it is not impossible that many of the familiares were employed in dragging blocks that were too heavy to be loaded onto wagons. Calculating the number of animals necessary to move such blocks, Peacock thinks that, in some cases, they would have been too numerous to work together, and that as a result it would have been necessary to use human traction.42 But apart from the two mentions of carts already cited and two papyri, one of which speaks of a camel requisitioned in a Fayyum village for “the transportation of the porphyry column,”43 the other of “very numerous beasts of burden” (κτήνη) gathered at Kaine (modern Qena) for “the transport of the fifty-foot column,”44 the papyrological sources are silent on the movement of the monoliths across the desert. Even the manner in which these last two texts are composed does not allow us to know if these columns were actually pulled by the animals: Peacock indeed thinks that the fifty-foot column has to have been dragged by men, and that the animals were there only to supply them. If the number of 400 familiares is in fact to be explained in this way, then we have to suppose that a large proportion of these men were only waiting for one or several monoliths to be ready for transportation, and thus that they were not actually part of the permanent personnel of the metallon. The coexistence of these two unequally treated worker communities and the division of work between them has some resemblances to the structure of the workforce engaged in building the royal tombs on the Theban left bank in the thirteenth and twelfth centuries BC. The role of the pagani is in this context occupied by the “Team,” made up of specialized artisans (quarriers, stonecutters, sculptors, draftsmen, and painters). The Team was assisted by auxiliaries (smdt), who were in charge of supplies (as carriers of water and wood, launderers, gardeners, doorkeepers, and fishermen), but who were also present in the work areas: “De façon générale, les smdt sont définis comme ‘ceux qui portent’ pour l’Équipe. Dans le courant du règne de Ramsès III, ils cessent d’être uniquement préposés aux approvisionnements domestiques pour aider à diverses tâches secondaires du chantier auparavant effectuées par des hommes de l’Équipe ou d’autres hommes ‘de l’extérieur’[…] même leur spécialisation paraît variable.”45 One may readily recognize the profile of the familia. The smdt belong to the category “outside teams” (i.e., who do not live in the village, which was reserved to the Team) to which also belong the cobblers, blacksmiths, doorkeepers, and policemen,46 all professions or offices which have a specific status at Claudianus. Like the pagani of Claudianus, the building workers of Dayr al-Madina worked under the supervision of the royal administration for a ruler preoccupied with eternity; but the connection with him was in their case closer and more personal, if only because of proximity with the capital, Thebes, located just across the river from them. If the village is of similar character to the two fortified villages47 of Mons Claudianus (a settlement in the middle of the desert, enclosed by a circuit wall, with adjoining living units), the workmen of the Team were more pampered by the king: they lived with their families in fairly good-sized

42. Peacock and Maxfield 1997: 264. 43. BGU III 762.16–18 (163): ἐπέμφθη εἰς χρείας τοῦ καθελκομένου κίονος πορφυρειτικοῦ. 44. P.Giss. 69 (118/9), cf. the article of Th. Peña on this text, JRA 2 (1989) 126–32. 45. Valbelle 1985: 131–32. 46. Valbelle 1985: 88. 47. I.e., the large praesidium of Wadi Umm Husayn and the village wrongly called “Hydreuma” by the archaeologists (hydreuma meaning “well” in the Eastern Desert, while this village had only a cistern).

200

Rome in Egypt’s Eastern Desert

houses and, if at times their pay was forgotten, they occasionally received presents, sweets, and little treats.48 And water was not doled out stingily to them:49 the Nile was at hand. 3. The soldiers O.Claud. inv. 1538 is the first ancient text that provides us with precise, complete, and indisputable numeric data on the staffing of a metallon under the principate and which reveals the proportion of soldiers in this population. It is very modest: Total staffing of the metallon50  Military  officers  infantry soldiers  Fannius cavalrymen  recruits Non-military architect pagani of whom – foremen  – quarrymen  – smithy  familia artisans, guards

917 60 2 29 1 6 22 857 1 421 7 349 65 400 35

% 6.5 %

93.4 % 45.9 %

43.6 %

Mons Claudianus was in this period under the command of a centurion (called praepositus: I.Pan 39), assisted by a decurion. The water schedules do not mention a curator praesidii Claudiani, a military position that we find at Claudianus in the ostraca from the Antonine dumps. Just one ostracon from the south dump, which moreover came from a shallow layer and is thus perhaps relatively recent, mentions a κουράτωρ Κλαυδιανοῦ (O.Claud. II 387). It seems therefore that during the period of large-scale quarrying connected with the great imperial commissions, a centurion was assigned to the site and that when activity slowed down, it was thought sufficient to have a curator, as there had been since Flavian times in the praesidia of lesser importance.51 The centurion and the decurion are the only military ranks mentioned. I think in fact that the tesserarius and the subtesserarius (line 8) were not soldiers: it seems that if they were, they would have been listed with the officers, or at least with the soldiers. We will see later that the dromadarius was also probably not a soldier. 48. Valbelle 1985: 152–55. 49. As at Mons Claudianus or even in the quarries of Wadi al-Hammamat in the pharaonic period: A. Gasse, “L’approvisionnement en eau dans les mines et carrières (aspects techniques et institutionnels),” in Menu (ed.) 1994: 171, n. 9. On the supply of water in the village of Dayr al-Madina, see in the same volume the contribution of S. Allam, “À propos de l’approvisionnement en eau de la colonie ouvrière de Deir el-Médîneh,” pp. 1–14. 50. Around ten: I have not counted the imperial cameleers, whose number is left blank, nor the hunter Herennius, who is perhaps already counted among the soldiers, nor Ischyras, newly arrived from Raïma and of unknown status. 51. In 189, there was no more than a vice-curator at Claudianus, who complains to the procurator metallorum that he has only two familiares at his disposition out of a theoretical complement of a dozen (Chapter 5). We are far from the 400 men of the familia under Trajan!

The organization chart of the personnel of an imperial quarry

201

The number of soldiers was naturally subject to daily variations as a result of their movements: if the infantry soldiers are 29 (plus Fannius) in inv. 1538, there were only 18 in the schedule inv. 2981 (the number of cavalrymen is lost in the latter ostracon). It will be evident that these soldiers were not used as manpower in the quarries, not even as technical staff. The ostraca of Mons Claudianus are not very informative about their function in the metallon: the only certainty is that they stood guard in routine fashion and that the cavalrymen (who are remarkably few52) supplied escorts. The centurion and decurion controlled logistics, in particular supervising the organization of water transport; they were also the natural recourse in cases of disobedience. Just the presence of this handful of soldiers must have helped maintain discipline and guarantee the constancy at work of this large body of workmen. In the well-known article of Roy Davies, “The daily life of the Roman Soldier,”53 we read the following statements: “Troops were frequently used in mines and quarries, mostly as labour force, but also to guard convicts sentenced there, or to provide supervision and technical expertise. (…) The duty-roster of legio tertia Cyrenaica assigns men to work in quarries, mining and burning lime and possibly collecting sand.” Davies cites numerous written sources in support of these assertions. Still, even if they bear witness to the presence of soldiers in the quarries or mines, they rarely specify just what they were doing there. Those historians who have taken an interest in this question always mention a metrical rock inscription incised at the entrance to one of the quarries of Jabal Tukh, opposite Ptolemais in middle Egypt. In it, a certain Isidoros son of Menippos (an architect?) thanks Pan and the Nymphs for the discovery of this quarry.54 Historians have cited it as proof of the existence of military manpower in the quarries ever since J. Zingerle proposed the following correction: ἡνίκα καστρήσιοι.55 This conjecture allows us to understand the text in the following way: “When the soldiers, on the orders of Mettius Rufus, worked the quarries for a platform for our homeland.” The idea is ingenious, but this correction–to a difficult, faulty inscription of which there is no photograph–is far from secure; castrensis is to my knowledge not attested as a substantive, and certainly not as a periphrasis to mean simply “soldier.”56 The proposed correction was, moreover, rejected by the last editor of the inscription, É. Bernand (I.métr. 116), who returned to an older conjecture that sees the problematic ατρησιοι as a faulty form of the poetic epithet ἀτρηστοίο, with reference to Μεττίου Ῥούφου; he translates, “when, thanks to the orders of the intrepid Mettius Rufus, our homeland cut some stones for a base.” In reality, there is only one certain attestation of a soldier working as a quarryman in Egypt: this is, in the quarries of the bekhen stone of Wadi al-Hammamat, the graffito of Gaius Aurelius Demos, who calls himself στρατιώτης σκληρουργὸς ὑδρευμάτων, “soldier quarryman of wells” (I.KoKo. 60). Given the extreme rarity of his cognomen, there is every likelihood that this is the same man as the soldier Aurelius Demos who made, along with his comrades, a dedication in the praesidium of Didymoi, not far from Wadi al-Hammamat (I.Did. 6.5); the inscription, cut on a base (of a statue?), is later than the rebuilding of this praesidium, which had been damaged, according to another inscription from the site, 52. We might expect that in this desert, with rather great distances between the various installations (metalla, road stations), the detached soldiers would mainly be cavalry, or even camel riders. But this is not at all the case: thus, in the praesidia of the Myos Hormos road, the cavalrymen are again less numerous than infantry (cf. O.Krok. I, pp. 2 f.). 53. ANRW II.1: 328. 54. Πανὶ ὁμοῦ Νύμφαι Ἰσιδώρῳ τάσδε ἔδωκαν | λατομίας εὑρεῖν τῷ Μενιππ⟨ο⟩ίο γόνῳ, | ἡνίκα ατρησιοι κελεύσμασι Μεττίου Ῥούφου | πάτρη ἡμετέρη κρηπίδι λαοτόμουν. The inscription dates from the prefecture of Mettius Rufus (89–91/2). It recalls another epigraphic thanksgiving, that of Gaius Cominius Leugas, who thanks Pan, under Tiberius, for the discovery of Porphyrites (AE 1995, 1615). 55. APF 9 (1930) 5–10. 56. The dictionaries cite only one possible, but questionable, attestation (Frontinus, Str. 2.5.30): omnibus castrensibus can also, of course, be neuter and vaguely designate “everything located in the camp.”

202

Rome in Egypt’s Eastern Desert

following the collapse of its hydreuma (I.Did. 3). There is nothing new about soldiers being employed in extracting material for the construction of military installations. But that was not the purpose of the monoliths extracted at Mons Claudianus. Fannius: strator? Mentioned among the soldiers in line 7, Fannius is the only one among them to be listed by his name, which is followed by the abbreviated word στρατ( ). My first instinct was to resolve this as στρατ(ιώτῃ), thinking that the case of Fannius was analogous to that of the ergodotes Sansnos, the only one of the ergodotai to be listed by name to receive a larger ration than that of his comrades. R. Haensch, however, wondered if instead one should not resolve the abbreviation as στράτ(ορι), thinking of two inscriptions of Adraha, in Arabia, relating to the construction of urban fortifications (OGIS 614 and SEG XVI 810, both from the second half of the third century: see the commentary of H.-G. Pflaum, Syria 29 [1952] 313–5); the strator of the provincial governor in that case provides the direction of works (ἐφεστώς) and is assisted by an architect (ὑφηγήσει Φλ. Οὐήρου ἀρχιτέκτονος). A strator at Mons Claudianus? I had asked myself the same question with respect to the soldier Iulius Arianus, who is perhaps ἱπποκόμος, “groom” (O.Claud. III 545.4n.). The stratores were in the service of superior officers (from the prefect of a cohort up to the praetorian prefect), of governors, and of procuratores provinciarum: their primary function, it seems, was to take care of the horses of these personages and of their headquarters, but they could also, like other officiales, be entrusted with tasks outside their specialty, as the inscriptions of Adraha show.57 The fact that Fannius reappears in l. 14 as responsible for a laundry task, however, is not favorable to connecting him with the stratores who built at Adraha. Nonetheless, one detail, however tenuous, might support the resolution στράτ(ωρ): in line 8, the scribe curiously felt the need to specify that the 24 keramia intended for the six horses of the cavalrymen were εἰς ποτισμόν, “for drinking.” I wonder if this specification, seemingly otiose, is not given in contrast to ἰς πλύσιμον, which would mean “for washing” the horses (or perhaps only one horse, seeing that Fannius, or rather Spinther, has only 13 liters of water for this task); horses are bathed when in sweat, but above all one cools with water the leg tendons of a horse who has run.58 Besides the schedules, the name Fannius is attested in O.Claud. only in inv. 2598, a Latin dipinto on an amphora in which can be read P̣ • Fannio (centuriae/centurioni) [. This document comes from an archaeological layer which could be contemporaneous with the schedules; there is a good likelihood that the person is the same. 4. Various occupations After the soldiers, the third section lists a certain number of individuals referred to by their profession or their office. The common element among them is that they belong to very small professional groups (they are sometimes the only representatives of their activity) and that for most of them their work does not take them into the quarries: their role is to ensure the good functioning of daily life in the desert. Their status is not specified; in some cases, one wonders if they are not soldiers (the surveyor of marbles, because he receives the same water ration as the soldiers, and the dromadarius, because of his occupation), but they could also belong to the familia, of which they represent the elite, as their rations show, for they are the same as those of the pagani, higher than that of the familia and lower (apart from the surveyor of marbles) than that of the soldiers. 57. I am here citing R. Haensch, whom I thank for having kindly communicated to me the pages concerning the strator from his unpublished study on the officiales. See also the useful picture provided by Le Bohec 1996: 315–17. 58. I owe this information to A. Bülow-Jacobsen, who is well acquainted with horses.

The organization chart of the personnel of an imperial quarry

203

The transportation corps of the centurion and decurion The decurion, who in principle had three horses, brought only one of them to Mons Claudianus. We have seen that the centurion had five donkeys at his disposal. These must have been kept near his hospitium, as these are the only donkeys, along with the decurion’s, to be included in the schedule. Five donkeys is quite a few, compared to the sole “Tragtier” that Petrikovits assigns to a centurion in his estimate of the transportation corps of a legion.59 It seems that these donkeys, unlike the mounts of the cavalrymen, were not the private property of the centurion and decurion: thus, instead of ὄνῳ δεκουρίονος, “for the decurion’s donkey,” as we have ἵππῳ δεκουρίονος, the scribe wrote ὄνῳ ὑπὸ δεκυρίονα, “for the donkey under the command of the decurion.” Did the administration put some donkeys at the disposal of the centurion and decurion for specific needs of their service in the desert? Tesserarius The tesserarius and the subtesserarius (add. lex.) must be reckoned among the management personnel; they are referred to as τεσ⟨σ⟩εραρίο(ις) β in the schedule inv. 2918. In the army, a tesserarius is part of the lower officers, who receive salary and a half, a fact that explains the alternative title sesquiplicarius; third in command in a century or a turma, he is in theory responsible for transmitting orders and draws his title from the tesserae on which these are written. As I have already noted earlier, it does not seem that the tesserarii of our schedules were soldiers; the other occurrences of tesserarius in O.Claud. confirm this impression. In another Trajanic dossier of lists indicating essentially a number of individuals belonging to various occupations and offices, the tesserarius is mentioned four times along with the doctor or doctors; in two of these texts we find the same sequence, tesserarius, doctor(s), cobbler(s), barber, as in inv. 1538.60 The letter inv. 1158 (year 13 or 14 of Trajan) informs its recipient that “our friend Phoinix is in charge of the provisions (kibaria), while Magius has become tesserarius”; this is thus the announcement of a double appointment: Phoinix has become the manager of provisions (kibariates) and Magius, who is known from other texts as a kibariates, has fulfilled, whether before or after having occupied this position, the functions of tesserarius, which must have been similar. The correspondence addressed to Athenodoros nearly fifty years later (around 152) confirms this impression: Athenodoros, tabularius of the procurator (scil. metallorum) Ulpius Himeros (an imperial freedman, see Chapter 2), resides at Claudianus, where he receives letters from the small satellite metallon of Tiberiane; his two correspondents are Nepheros, the curator of Tiberiane, and thus a soldier, and Kallistratos, a tesserarius. The letters of Kallistratos are deferential; in the opening address, the name of the addressee Athenodoros precedes that of the writer Kallistratos, and several times Athenodoros is referred to as τῷ δεσπότῃ, “my master”; Nepheros, in contrast, never puts the name of his correspondent in first place except when he addresses his centurion. The tone of their letters is also very different: Nepheros submits to the tabularius his chronic provisioning problems with water and food, so that he can solve them for him, while the tesserarius limits himself to sending reports on the comings and goings of men and animals and the distribution of foodstuffs; he is thus in some sense Athenodoros’s eyes at Tiberiane. It is clear that the soldier and the imperial servant collaborate on an equal footing, while the tesserarius belongs to the office of Athenodoros, whose subordinate he is. The name of this military rank was borrowed by civilian society: thus, in CIL V.2 5272 (Comum), in which 59. von Petrikovits 1975: 58. This number five makes one think, however, of the decree of Sextus Sotidius Strabo, legatus pro praetore of Tiberius in Galatia, who set at six the maximum number of donkeys that a travelling centurion had the right to requisition when he was crossing the territory of Sagalassos (SEG XXVI 1392; I thank the editorial committee of Chiron for having pointed out this connection to me). 60. O.Claud. IV 713 and 714. The two other ostraca referred to are O.Claud. IV 722 and inv. 8633.

204

Rome in Egypt’s Eastern Desert

a funerary cult is established, to be provided by a college that includes tesserarii. There are rare examples of slave tesserarii, in the imperial household: a slave of Augustus, Pederos Aug(usti) tes(serarius) (CIL VI 9081) and one of Caligula, Symphoro tesserario ser(uo) Caesaris de domo Gelotiana (CIL VI 8663);61 also CIL VI 9915 (ILS 1708), where, at the head of a group of eighteen tabellarii, all of them slaves or freedmen, are two optiones and a tesserarius, all three freedmen; for G. Boulvert, these tesserarii of the imperial household, like their military namesakes, were in charge of transmitting orders of the emperor to their subordinates.62 The civilian tesserarius was thus, just like his military model, a channel of transmission; this is, clearly, the connection between the civilian tabularius Athenodoros and his tesserarius Kallistratos, who was in fact his representative at Tiberiane, and who not only transmitted his orders but above all, as the ostraca show us, reported on their proper execution. Dromadarius Unlike the cavalrymen’s horses, the mount of the dromadarius is not mentioned, but dromedaries do not need to drink every day.63 In another schedule in the dossier, there are not one but two dromadarii, who are also not mentioned among the soldiers, but in the group of guards and artisans, which raises doubts about whether they were part of the army, particularly since their water ration is not the same as that of the soldiers; it is probably not by chance that the dromadarius is named right beside the dekanos of the camels. Dromadarius is a Latin word derived from the Greek δρομάς, “running camel”;64 it appears for the first time under Trajan, in the name of a military unit, the ala I Ulpia dromadariorum Palmyrenorum:65 these dromadarii are soldiers mounted on dromedaries. This is the only named unit of dromadarii under the principate; otherwise, some units of the same period are known to have included a small number of dromadarii. Rom. Mil. Rec. 64 (156), a pridianum of the cohors I Lusitanorum equitata, gives the following numbers of personnel for a particular day: 114 cavalrymen, 19 dromadarii, 363 infantry; in P.Brooklyn 24 (c. 215), there are 13 dromadarii for a numerus purus of 457 men; the proportions are roughly similar at Dura-Europos. In other texts from the Eastern Desert, the dromadarius is identified with his mount, called δρομάς: thus, O.KaLa. inv. 819, already mentioned (comm. ad 15), is a letter written under Domitian or Trajan by a certain Antonas to the architect Hieronymos, who was working in the quarries of Umm Balad, satellites of Porphyrites: Antonas, who is situated in a nearby but unidentified post, asks him to lend the dromas of Porphyrites and to send it to him so that he can go in search of a missing person, for, he says, “ours is at Claudianus.” It thus seems that these places disposed of no more than one dromedary. The ostraca of Umm Balad, which date from the reigns of Domitian and Trajan, give the names of several dromadarii, whose names, which are vernacular or Arab (which is significant), can scarcely at this period be those of soldiers: Pakoibis, Sansnos, and Abdaios. We also know the name, Egyptian, of one dromadarius in Claudianus: Karouris.66 The occupational term dromadarius seems thus to be applied both to soldiers and to non-soldiers employed by the army. 61. J. Ramminger points out to me that the tesserarius of this inscription is understood as a mosaicist (tesselarius) by H. Blümner, Die römischen Privataltertümer (Munich 1911) 97, n. 6, just like the tesseraria of CIL V.2 7044. 62. Boulvert 1970: 34 and 87. 63. A camel drinks every 10 to 15 days in winter, every 3 days in summer (oral communication of a Beduin of the Eastern Desert). 64. We should remember that the dromas is a representative of the same species (Camelus dromedarius) as the animal called κάμηλος; the dromas belongs to a running breed, the kamelos to a transport or hauling breed. 65. The earliest attestation of this ala has recently been pushed back to 141/2, thanks to a new military diploma: Weiß and Speidel 2004, at 256 ff. (I thank the editorial committee for having alerted me to this article). But the name Ulpia shows that the creation of the unit goes back to Trajan. 66. O.Claud. inv. 8452.

The organization chart of the personnel of an imperial quarry

205

The surveyor of marbles (ἀναμετρητὴς μαρμάρων) Ἀναμετρητής is a rare occupational term, attested only in Egypt in individual declarations of cultivated land dating to 302 and 303 and drawn up in response to an order of the rationalis Valerius Euethios: these documents are addressed to ἀναμετρηταί who, unlike the γεωμέτραι, men who measured parcels on the spot along with the cultivators, were white collar officials who belonged to the bouleutic class and operated at the level of the toparchy.67 The noun ἀναμέτρησις, in contrast, is common; it applies normally to surveying; the editors of P.Köln 53 (263), however, restore in l. 1 ἀναμ̣[έτρησις ξύλων; this account indicates indeed the dimensions and volume of several pieces of building wood. The technique and units of measure were common to wood and building stone, as we can see from the title of a treatise by a certain Didymus of Alexandria, “The measurement of marbles and various woods” (Διδύμου Ἀλεξανδρέως μέτρα μαρμάρων καὶ παντοίων ξύλων).68 We encounter the verb ἀναμετρεῖν in connection with the quarries in the Ptolemaic papyrus P.Petrie Kleon 51, which has already been mentioned (n. 31): the frustrated quarrymen ask the architect to measure the material which they have been forced to quarry beyond what had been agreed on (ἀναμετρήσας ἡμῖν ὧι π̣[λ]είω τετμήκαμ̣ες τὴν στερεὰν πέτρα[ν), in order to determine an equitable division of work with the other team. In PSI IV 423.13 (third century BC), it is the verb ἐγμετρεῖν that is used, probably to refer to the same type of work: at stake is the measurement of the volume of material extracted by the quarrymen, who are prisoners, each of whom is required to produce 1 aoilion69 each day. From ἐκμετρεῖν derives the occupation ἐκμετρητής, attested in P.Petrie Kleon 39, where it is a matter, according to the most recent editor of this text, of measuring the work accomplished not in mines, as Fitzler thought,70 but on dikes located near the mines. In PSI IV 423, the labor force does piece-work: it is above all the amount of work got through that counts; nothing to do, then, with the imperial orders for architectonic elements carried out to measure by the skilled sklerourgoi of Mons Claudianus. Here the “measurer of marbles” (whose water ration places him hierarchically on a par with the soldiers) is perhaps rather to be seen as a “controller of marbles.” This office is well known in quarries, where it is usually entrusted to imperial freedmen. It consists in verifying that the blocks conform to the orders before letting them be dispatched; an essential role, given the time and cost required by ancient transportation. The Latin term for this controller is probator; in Greek, this post seems to be known by the term δοκιμαστής.71 Platearius In O.Claud., the platearius is always single, except in O.Claud. IV 722, where there are two of them. This term, derived from platea (“street”), itself a direct borrowing from the Greek πλατεῖα (scil. ὁδός), could, in the view of A. Bülow-Jacobsen, characterize workers employed in maintaining the slipways constructed in the mountain for bringing down the blocks; but the name of these slipways does not appear in the O.Claud. Πλατεῖα never in this period refers to anything except a large and broad street, an 67. P. Cornell, p. 112. 68. Mentioned in P.Köln I 53, introduction. 69. A unit for measuring volumes. 70. Fitzler 1910: 67. 71. Hirschfeld 1905: 164 and n. 3; Hirschfeld was mistaken to cite in this context ὁ τὸν λίθον διακρίνων τεχνίτης (Diod. Sic. 3.12.5): that was a mining specialist who indicated to the miners where they should dig. In the cipollino quarries, in Euboea, this inspector was, under Hadrian, a certain Crescens, an imperial freedman (cf. Bruzza no. 1, reproduced in H. Dodge and B. Ward-Perkins (eds.), Marble in Antiquity. Collected Papers of J. B. Ward-Perkins [London 1992] 27).

206

Rome in Egypt’s Eastern Desert

“avenue”;72 it is thus remarkable that platearius (if the proposed interpretation is right) should have been formed in reference to the slipways, which are rather narrow: probably the sense of “levelness” present in some uses of the adjective πλατύς will have driven the choice of this term;73 the slipways must have been rather flat, and each movement of a block must have damaged them. It is then difficult to see why there are so few platearii: were these specialists directing teams, like the ὁδοποιοί in the Constitution of Athens (Arist. Ath. 54.1), who were five agents put in charge of building roads, with a workforce of public slaves at their disposition? In that case we are presumably dealing with the streets of Athens, but certainly ὁδοποιία, the provision of routes for the transportation of blocks, was also a frequent expense line in the accounts of the quarries of the Greek world.74 The word platearius was hitherto unknown in the ancient world, but it is not entirely an addendum lexicis: it is attested in Medieval Latin, in several texts from Sicily, where it refers to an employee of the customs station, a “toll collector.”75 Cobbler and barber Our schedules attest the presence of two cobblers (σκυτεῖς) and one barber (κουρεύς), listed in this order both in inv. 1538 and in another dossier of lists of personnel separate from the schedules; by coincidence, in the Table Vispasca I (end of the first or beginning of the second century),76 which regulates the exercise of various professions on mining territory, the chapter “cobbler” immediately precedes the chapter “barber.” Nothing in the O.Claud., however, suggests that these three artisans were private citizens who had contracted for the right of exercising their occupation at Mons Claudianus (conductores) or, a possibility foreseen in Vipasca I, employees of such conductores. In general, reading Vipasca I gives one the impression that the population of the mining territory was fairly similar to the population of any other settlement; families lived there, as the mention of women as customers of the bathing establishment and of schoolmasters shows. There is nothing like this at Mons Claudianus. The divergent natures of the methods of exploitation was probably not unrelated to this difference in the social profile of these two contemporary metalla: the shafts at Vipasca were worked by independent entrepreneurs who took them on lease (indirect management), while the quarries of Claudianus were operated by salaried employees of the imperial administration (direct management). But the main factor in play surely was the problem of the food and water supply, which was a decisive obstacle to bringing useless mouths to the place. The epimeletes of the Paneion Egyptian documentation provides only three late examples (third and fourth centuries) of epimeletai of sanctuaries: these were liturgists drawn from the bouleutic class, who administered the possessions of these temples. On this subject, a key reference is P.Oxy. XXXI 2563.5–8n.; this text, dated to 170, is a 72. The meaning of “place” or “square” is late. 73. Cf. du Bouchet 2004: 216: “une πλατεῖα n’est pas seulement une voie large, mais aussi, comme le montre bien le règlement des astynomes du Pirée [IG II2 380, 19–20], une voie dont la surface est aménagée et entretenue, c’est-à-dire aussi égale qu’il peut être utile à la circulation des piétons et des véhicules. L’égalité de la surface d’une πλατεῖα est aussi importante que sa largeur…” 74. Raepsaet 1987: 41. 75. Cf. J. F. Niermeyer et al., Mediae Latinitatis Lexicon minus (Leiden 20022), s.v. I thank Johann Ramminger for having drawn my attention to this reinvention of platearius. 76. Domergue 1983: 180.

The organization chart of the personnel of an imperial quarry

207

request written by an ἐπιμελητὴς Σαραπείου φυλῶν καὶ τῶν ἄλλων ἱερῶν τῆς πόλεως,77 an unparalleled title. Our epimeletes of the Paneion evidently has nothing to do with these municipal notables. Rather, we must be dealing with the sacristan of a modest sanctuary, and his title could be a rendering of curator fani or templi,78 a position perhaps identical to that of aedituus (terms both of which represent extremely varied functions and social statuses).79 It is not hard to see why this isolated semantic calque would appear at Mons Claudianus: this sanctuary of Pan was built in a territory that was directly under Roman administration and free of all preexisting cultural influence: the Paneia of the quarries of Semna, Persou (Wadi al-Hammamat), and Porphyrites were Roman initiatives corresponding to the founding moments of launching the exploitation of these metalla or, in the case of Persou, of the resumption of work under a new regime;80 the same must have been true at Claudianus. In another context, several of the positions that we have reviewed here could have been filled by soldiers, if we are to believe the list of the immunes, exempted from corvée, because of their specialty or their technical competence, drawn up by Tarruntenus Paternus;81 besides the soldier lapidarii (a specialty represented by our sklerourgoi pagani), it includes several occupations or positions close to those that at Mons Claudianus were discharged by non-military personnel: mensor, medicus, veterinarius, architectus, and horreorum librarius, who is reminiscent of our manager of the barley granary.

5. Is the list, despite everything, incomplete? O.Claud. inv. 1538 registers 917 persons present at Mons Claudianus on a specific day. That is a lot of people, and yet this list may not be complete. We may leave aside some occupational designations mentioned in other ostraca of Mons Claudianus (such as σφυροκόπος, φαλαγγάριος, and ἀκουάριος), which refer to activities carried out by the imperial personnel; they do not appear in the schedule because they were not permanent occupations: Parthenopaios, an employee of the familia, on one day transports goatskins, on another day collects wood. These men are thus taken into account, but they are cloaked in the anonymity of the term familia: actually, the person drawing up the list does not note the profession of the recipients except when it affects the amount of the ration. On the other hand, there is silence about the herds of donkeys and camels who served the shuttle between the well, the camp, the quarries, and the satellite establishments, as well as about the personnel who managed and cared for them (the cameleers are well represented by the “imperial cameleers” and by the “dekanos of the camels,” but there is no donkey-driver mentioned except that of the centurion). Men and beasts must have watered themselves at the wells where they went to fetch water. The absence of any mention in the dossier of a wheat granary (although there is a barley granary), or of bakers and a bakery, must also be pointed out. Perhaps at this period all of the bread, and not only that of the pagani, was routinely imported from the valley (cf. O.Claud. I 7 and 8). But this was not always the case: under Antoninus, the pagani continued to have their bread baked by their family in the 77. “Epimeletes of the [priestly] tribes of the Serapeum and the other sanctuaries of the metropolis.” 78. For the equivalence ἐπιμελητής - curator, cf. Magie 1905: 14–15. On the curators of temples: RE IV.2 1798 and 1803; Liebenam 1900: 343 ff. 79. Liebenam 1900: 344: “augenscheinlich sind aber die Funktionen und demnach Rang wie Ansehen der aeditui sehr mannigfaltige gewesen, je nachdem ob sie als Tempelwächter niedere Dienste thaten oder die Verwaltung der geheiligten Stätten überhaupt wahrzunehmen hatten.” 80. The first two were dedicated by the procurator of the mines and quarries of Egypt (respectively I.Pan 51 and I.KoKo. 41); the Paneion of Porphyrites by the discoverer of this metallon, Gaius Cominius Leugas, who does not indicate his title (Van Rengen 1995 = AE 1995, 1615). 81. End of the second century. This list is preserved in Dig. 50.6.7.

208

Rome in Egypt’s Eastern Desert

valley, but the imperial staff had it delivered by a baker (inv. 5389); O.Claud. II 287–288 attest that in some cases, not only was the bread baked in the desert, but the wheat was also milled there. In the south dump, thus under Trajan or Hadrian, a dump of carbonized bread was found. Finally, another dossier coming from a chronologically ambiguous layer82 of the west dump notes day after day the dispatch of two men (probably belonging to the familia) to the bakery (ἀρτοκόπιον). There is no trace of women, either, although there were certainly some on the site, wives, concubines, and prostitutes. They must have been supplied outside official channels.

V. The measure of water 1. Metrology In sections 1 and 2 of the ostracon, water is reckoned in goatskins (ἀσκοί) and in amphoras (κεράμια), while in the following sections only amphoras are recorded. This difference in the units of measurement used is likely to reflect a difference in the packaging of the water; if this supposition is correct, the water, apart from small quantities, was packed in goatskins when it was transported to workplaces, that is, the quarries and the adjacent smithies; on the other hand, it is distributed in keramia to individuals. It was easier to transport goatskins than the heavy and awkward keramia across the mountain; as well, the work areas brought together sizable groups of humans whom it was more sensible to supply through large-capacity containers, but these groups dispersed in the evening when the men returned to their kellai. The quantities expressed in goatskins and keramia show that the keramion functioned as a subdivision of the goatskin. But how many keramia per askos? Ironically, we have reasons to support three different ratios: 1:4, 1:5, and 1:6. We know the capacity of the keramion, which will serve as the point of departure for our analysis: indeed, this vessel, omnipresent in O.Claud., can hardly be anything but the biconical Egyptian amphora AE3, which represents the vast majority of the ceramic material gathered in the dumps of the metallon and, more generally, of the Roman sites of the first and second centuries in the Eastern Desert. The capacity of this amphora has been measured: it is 6.5 liters (the equivalent–by chance?–of 2 Roman congii).83 (a) 1 goatskin = 4 keramia. Because in none of the nine surviving cases is the number of keramia greater than 3, it is possible that the capacity of a goatskin was equal to four keramia; this would agree nicely with the conventional capacity of the goatskins used today in the Sahara, which is 20 to 25 liters (see infra); at the period of the Napoleonic Expédition d’Égypte, the city of Qusayr was supplied with water by camel caravans transporting goatskins of about 20 kg.84 If 1 goatskin = 4 keramia, the capacity of the goatskin would have been 26 liters, which is precisely the capacity of the goatskins in which tax oil was delivered at Carthage in 373.85 This relationship appears confirmed by O.Claud. inv. 1999, a text that does not belong to the 82. This context has produced both an ostracon dated to 116/7 and an entole of 136–138. 83. This capacity was measured both at Mons Claudianus by Roberta Tomber and at Maximianon, on the Myos Hormos road, by J.-P. Brun. 84. Du Bois-Aymé, “Mémoire sur la ville de Qoçeyr et ses environs et sur les peuples nomades qui habitent cette partie de l’ancienne Troglodytique,” Description de l’Égypte. État moderne, I (Paris 1809) 193. 85. Peña 1998: 171.

The organization chart of the personnel of an imperial quarry

209

dossier of the schedules. But the numerical information preserved in the first and second sections of O.Claud. inv. 1538 suggests, although not without some internal contradiction, a larger capacity. (b) 1 goatskin = 6 keramia. The first section has preserved for us the quantity assigned individually to a blacksmith (for the tempering of the metal), which is 3 keramia; besides that, 2 blacksmiths receive 1 goatskin, and 11 blacksmiths receive 5 goatskins and 3 keramia. The conversion of the keramia into goatskins comes out exactly in both cases if 1 goatskin = 6 keramia (or 39 liters). (c) 1 goatskin = 5 keramia. The second section, however, supposes an askos of 5 keramia. The quantities of water distributed to the various quarries seem actually to have been calculated on the following bases: natives and imperial employees received respectively 1/2 and 1/3 of a keramion (which is also their ration in the fourth section) and 1 goatskin is, as we have just seen, equivalent to 5 keramia. In five of eight cases, the result obtained is the same as that reached by the scribe; in the cases where he departs from this, the number of askoi and of keramia is identical, but the scribe has allowed the fractions of a keramion drop out. A goatskin of 5 keramia would have a capacity of 32.5 liters. What we can discover of the number of goatskins transported respectively by donkeys and camels does not help us very much, given the imprecision of the information at our disposition on the carrying capacity of these animals, particularly donkeys. We know only, from several ostraca, that at Mons Claudianus four goatskins were fastened on the back of a camel used for carrying water, while a donkey carried only two of them. The skins used in the Roman sites of the Eastern Desert were made with the skins of small ruminants (sheep or goat).86 Might the capacity of the skins used in Africa today allow us to decide in favor of one hypothesis over the others? Alas not, because it is in fact highly variable, to judge from the following testimony:87  “La guerba, l’outre en peau de bouc (elle est le plus souvent d’ailleurs en peau de chèvre), souple, épousant les flancs du chameau porteur, solide, imperméable, conservant à l’eau une certaine fraîcheur, à défaut d’un goût agréable, est faite d’une peau entière taillée et retournée.88 Les pattes sont ficelées et permettent l’arrimage. Le col du récipient est formé par le cou de l’animal. La capacité moyenne est de 20 à 25 litres, jusqu’à 50. Les Nemadi chasseurs de Mauritanie, qui n’ont pas de chèvres, en font en peau d’antilope qui ont jusqu’à 80 ou 100 litres de capacité. L’arrimage normal des guerba se fait par paires, une de chaque côté de la monture, pour raisons d’équilibre.” The equation 1 goatskin = 5 keramia is close to the indications given by H. Lhôte with respect to practice in the Sahara: a camel can carry four goatskins containing 30 liters each, for a total of 120 liters per animal.89

86. S. Winterbottom, in Maxfield and Peacock 2001a: 330–32; Leguilloux 2004: 141 f. 87. de Planhol and Rognon 1970: 87–88. 88. This would already be the case for the skins in which the pharaonic expeditions to the quarries of Wadi al-Hammamat transported their water: “l’outre est appelée šd, mot qui semble désigner la peau entière plutôt qu’un récipient de cuir” (A. Gasse, in Menu [ed.] 1994: 172). 89. Lhôte 1967: 60.

210

Rome in Egypt’s Eastern Desert

load capacity of 1 goatskin: 4 keramia (= 26 liters)? 5 keramia (= 32.5 liters)? 6 keramia (= 39 liters)?

1 donkey (= 2 goatskins) 50 kg,90 65.5 kg,91 70–90 kg92 52 liters (kg)95 65 liters (kg) 78 liters (kg)

1 camel (= 4 goatskins) 187 kg,93 196 kg94 104 liters (kg) 130 liters (kg) 156 liters (kg)

2. The level of rations It appears clearly that the level of the rations did not depend only on the physical needs of individuals: considerations of hierarchy also played a role, such that the amounts of water allocated to individuals are an indication of the social prestige of the men and the occupations they practiced in the microsociety of the metallon. In O.Claud. inv. 1538, the quantities assigned to the officers and the architect were left blank; this is not the case in other schedules, which show that they were not fixed and these persons received water ad libitum, measured in goatskins. On the other hand, when we descend in the hierarchy, the individual rations are strictly defined, according to a scale (for humans) that is composed of four levels: – 1 keramion (6.5 liters): this is the maximum ration; only two individuals, one soldier and one native, enjoy this level, both listed by name, while their colleagues are not: Fannius and the ergodotes Sansnos. – 5/6 keramion (5.4 liters): ordinary soldiers, both infantry and cavalry, but excluding recruits; the surveyor of marbles, whose juridical status is unknown, and who alone of the recipients listed by the name of their occupation receives the military ration; in fact, his appearance comes regularly at the divide between the list of soldiers and their mounts and that of tradespeople and guards: in inv. 1538, he is listed between the horses of the cavalry and the tesserarii, in inv. 2918 directly after Fannius, in inv. 2853 directly before the decurion’s horse. I wonder if the ration of 5/6 should not be broken into 1/2 + 1/3, 1/2 being the 90. A maximum load recognized in 1923 by the British War Ministry (Nibbi 1979: 155). The donkeys of the British army were better treated than those of antiquity; when we consider that some customs taxes were calculated not on the basis of the weight of goods carried, but on the number of transport animals used (P.Lond. III, p. 40; Wallace 1938: 269), we can hardly be surprised if the Egyptians were not about to avoid overloading their animals. In the Mediterranean, donkeys were loaded up with burdens of 75 to 100 kg (Vigneron 1968: 135). On the other hand, modern armies use their transports half-loaded in order to be able to cross difficult terrain more easily (I owe this observation to A. Bülow-Jacobsen); what is true for trucks ought also to have been for donkeys. 91. The load supposed by the Edict of maximum prices 14.11 for a donkey transporting wood; note that the corresponding load for a donkey in the same chapter is only 400 pounds (131 kg), although an ordinary camel is capable of carrying as much as 600 pounds (196.5 kg) in Chapter 17.4 (where the load of a donkey has been omitted): this heavy load may have been intended only for short journeys, as Jean Bingen suggested to me. 92. White 1984: 129 (without citation of sources). 93. 6 artabas is a classic load for a camel (P.Wisc. II 47, BGU XI 2109): using the standard artaba of 39 liters, that makes 187 kg. On camel loads, see also O.Bu Njem, p. 102, where, however, the reference to the Berlin papyri that supposedly attest camel loads of 150 kg are to be eliminated: these are actually receipts for the 3 percent tax (ρ⸍ καὶ ν⸍). Other testimonies, ancient and modern, are collected by P.-L. Gatier, “Le chameau de transport dans le Proche-Orient antique,” in B. Redon and D. Agut-Labordère (eds.), Les vaisseaux du désert et des steppes: les camélidés dans l’Antiquité (Camelus dromedarius et Camelus bactrianus) (Lyon 2020) 227–55, at 234. 94. Camel load according to the Edict of maximum prices 17.4. 95. Without, to be sure, counting the several kilos that we must add for the weight of the containers themselves.

The organization chart of the personnel of an imperial quarry

211

personal ration of the soldier, 1/3 that of his personal slave: it is in fact known that soldiers often had a slave to wait on them,96 and 1/3 is precisely the individual ration of the familia. In this case, the personal ration of the soldiers would be the same as that of the tirones, but the latter would not have a slave (or would not have been authorized to bring a slave with them). – 1/2 keramion (3.25 liters): the recruits (tirones), the pagani workers, whether quarrymen or blacksmiths (but excluding the bellows operators), the small number of tradesmen who were not working in the quarries (as well as the two doctors and the veterinarian, along with the two cobblers, the two butchers, and the barber), the guards of various living quarters or general purpose; usually, the status of these people is not indicated, but we observe that in this group Ummidius, a member of the familia, receives 1/2 keramion, just as does a paganus carter. – 1/3 keramion  (2.16 liters): the anonymous members of the familia, providing an additional marker of the privileged conditions accorded to the workers who were pagani in comparison to their imperial employee comrades; the physetai, who operated the bellows of the smithies (a hard task that was still at the beginning of the twentieth century in France left to young apprentices: cf. O.Claud. II 217.2n.); besides these two disadvantaged groups, only a messenger who had come from Raïma receives this lower ration. It should therefore represent a minimum for life, a fact that allows us to establish that the schedules providing for rations of 1/3 keramion were not drawn up during the summer heat (2 liters constitute the daily requirement of an inactive man who stays in the shade at 20°C or 68°F). A confirmation is provided by O.Claud. inv. 3666, the only schedule in the series which preserves a date, Hathyr 20 (= 16 or 17 November). As for the animals, meaning the cavalry horses and some donkeys, the rations are 4 keramia (26 l.) for a horse and 2 (= 13 l.) for a donkey. We must admit that if the schedule has indeed been drawn up in the cold season, 13 l. is a generous ration, given that a donkey drinks about 10 liters of water each day in hot weather. On the other hand, 26 liters for the horses is a modest ration: today, in temperate countries, they get 30 liters of water a day (a bucket in the morning, one at noon, and one in the evening), an inadequate ration if the animal, which sweats a lot (unlike the donkey, which is better adapted to desert conditions), works or is hot. The 26 liters here must represent the ration for a horse who is staying in camp on a winter’s day. All of the rations discussed above are those distributed in camp. There remains the question of the water rations distributed in the quarries (lines 3–6). We have observed already that the individual rations of the pagani and the familia were in line with those set out in the fourth section. A doubt arises in this respect: supposing we are dealing with the same distribution of water? I do not think so: we have seen that not all of the familiares worked in the areas. It therefore seems that the “non-sedentary” personnel, who worked outside, received two water rations. This is not surprising: a quarry worker who labors eight to ten hours in full sun becomes dehydrated more than the guard of a storeroom or a doorkeeper. In fact, even at the same temperature the needs of an individual who is working in the sun are twice those of an inactive person in the shade.97 It is nonetheless striking that even in the work areas where physical needs should have come first, the amount of water allocated continued to depend on considerations of social status (but we do not know if the apportionment of the water brought in bulk to the quarry took account of the fact that the familiares were disfavored in principle with respect to the pagani). 96. Rouland 1977: 29: under the principate, ordinary soldiers had only a single slave in their service, while the lower officer ranks generally had two; Speidel 1992: every cavalryman had a valet; auxiliary infantry often had one, but we do not know if this was universally the case. 97. de Planhol and Rognon 1970: 228. Measures reproduced by A. Gasse, in Menu (ed.) 1994: 171.

212

Rome in Egypt’s Eastern Desert

The ceramologists have not found a container that could represent a half-keramion or a third of a keramion. How then was the distribution of individual rations carried out when they did not correspond to 1 keramion? I think that the recipients received a keramion to be shared among two or three individuals (should we suppose that the recipients of 5/6 keramion received 5 keramia for six people?): such a scenario may be behind O.Claud. inv. 1, a ticket on which two men with Egyptian names are mentioned (thus pagani), followed by the indication “1 keramion.”98 The lists of two, three, four, five, or six men that are so numerous in the Claudianus corpus could correspond to such collective distributions. *** We must finally ask at what stage of the procedure of the water distribution O.Claud. inv. 1538 is located. There are two conceivable options, and it does not seem possible to decide between them: (1) It is a list of the day’s needs, drawn up on the basis of the number of persons present in the metallon and their division between the workplaces; it could serve to show the maximum authorized consumption. (2) It is a statement of the quantities of water actually distributed on that day. In this case, I would be tempted to connect it with a series of ostraca which are lists of work areas, with the name of the area followed in each case by a more or less long series of figures, which are mainly 1 and 2 (i.e., either α or β). The names of these areas coincide in part with those that figure in the schedules (we find in them Traiane, Mese, Hera, and the stomoterion). These might be lists in which the scribe of the office in charge of distribution of goatskins of water reported each dispatch as it went out; at the end of the day, he would add up the amounts dispatched and write the totals in a schedule of this type. Indeed, those in charge of the distribution of a product would not proceed without receiving instructions from a hierarchical superior, nor without making a report subsequently to the person who gave the order. This administrative principle is well summarized in a letter found at Mons Claudianus and already referred to: “You will distribute 1 measure of kibarion to nine men of the familia who are arriving from Egypt with the twelvewheeled wagon, and you will submit an accounting to me” (ἀνοίσεις τὸν λόγον).99 Whatever may have been the intention of the scribe who prepared and filled out O.Claud. inv. 1538, this exceptional document, which is in a sense a snapshot of the Claudian microsociety, reveals, with the help of figures, the structure of the workforce and the main lines of the division of work in a prestigious quarry of the principate: under the control of a handful of soldiers worked side by side a community of local artisans, almost all quarriers and stonecutters, who were, we know, well paid and enjoyed a certain autonomy, and a similar number of imperial workers, the familia, mostly not specialized, who seem above all to have been employed in the logistics of desert life and in the transportation of the great monoliths.

98. We cannot, however, exclude the possibility that this is a question of a keramion of wine. The keramia used to measure water were, after all, only recycled wine amphoras. 99. O.Claud. IV 871.

12 A dedication to Zeus Helios Great Sarapis in honor of a desector on an ostracon from Mons Claudianus O.Claud. inv. 7363 Gate, US 6 Fig. 50

28.5 × 24 cm

222–235

A large, shield-shaped sherd of light ceramic with fifteen lines in red ink which is rather faded. The ostracon was found in a surface layer outside the gate of the main fortified village of Mons Claudianus.1 Lines 1–11 are a dedication in epigraphic style, carefully written in round majuscules, also called “Roman uncial,” especially popular under the Antonines. Our ostracon, which dates from Severus Alexander, would thus be a late example of this style of writing, which is well suited for epigraphy and was imitated by stonecutters in the second century.2 From line 12 on, the nature of the document changes, and what little can be read is reminiscent of acclamations. The style of writing also changes: the book-hand is replaced by a tall and narrow chancery hand which the writer tries out in several modules, perhaps with the intention of filling the whole surface of the sherd. The contrast is most clearly seen between line 12 and the following. Is this a model for the engraving of an inscription? Or is it a draft which became a writing exercise? I am inclined to think that it is a cheap substitute for a stone-inscription, and my reasons are the following: – The choice of a nicely symmetrical ostracon and the careful writing. The drafts that one gave to the stonecutters were not calligraphic, a reason for the errors they made.3 The ordinatio and the tracing of the letters were the job of the stonecutter and were done directly on the stone. P.Oxy. XLI 2950 (286–305), a dedication to Diocletian and Maximinian by a vexillatio of the legio V 1. Maxfield and Peacock 2001a: 18. 2. Cavallo 2005: 157 (not beyond the third cent.), 158 (epigraphic imitation). 3. There is a list of drafts, models, and copies of inscriptions on papyrus in P.Oxy. LXXIX, p. 127.

213

214

Rome in Egypt’s Eastern Desert Macedonica, is a parallel: it is so perfectly executed that the editor thinks it might be a “carton,” meant to be copied directly on the stone. This would be unique. As an alternative he suggests the hypothesis that it might be a banner, which seems plausible. – The color of the ink: In papyrological texts, the use of red ink is exceptional and reserved for very specific uses.4 Even more rarely red was used when nothing else was available.5 On the other hand, inscriptions were often painted, and the red ink here is a marker that gives the ostracon a status as an inscription. – Contrary to those stationed in the praesidia along the road to Berenike, e.g., Dios or Didymoi, the soldiers at Mons Claudianus did not have access to a soft stone such as steatite or sandstone, which could be incised with the point of a knife.6 To execute an inscription like this on the local granite would have demanded time, technical skill, and considerable expense, all of which were perhaps in excess of the intentions.

Our dedication is the last precisely dated evidence of official activity on Mons Claudianus. This late activity was probably insignificant compared to what went on during the reign of Trajan, when a workforce of some 850 men and 60 soldiers were active there.7 Claudianus ostraca later than Antoninus Pius are sporadic, and the call for help from the vice-curator Rufus Aristoteles in year 29 of Commodus (AD 189) gives an impression of dereliction.8 The detachment of workers and soldiers of whom Rufus was in charge was limited, to judge from the number of slaves (two), donkeys (one), and water-skins (eight old ones) that were at the disposal of the vexillum. Besides, no important construction containing Claudianus granite is known from the period of Severus Alexander.9 It must have been a small command for stone which needed only a limited expeditionary force. The ostracon was not found in a sanctuary but in front of the fortified village. Was it transported there from the Serapeum on the mountainside, or was it, as Peacock suggests, suspended at the gate where it was found?10 We may try to imagine the little group of soldiers making a farewell ceremony at the gate of the praesidium, happy to have accomplished their task and soon to go back to the more animated environment of the Nile valley.

4. Red ink was first of all used to mark out extracts from official registers written at a later date than the original registration. See most recently an updated list of papyri written with red ink in P. Schubert, “BGU I, 361 et P.Gen. inv. 69 : retour sur l’encre rouge,” APF 51.2 (2005) 249–52. 5. Apart from some tituli on amphora stoppers of plaster or on ceramic containers which were probably not inscribed locally, the only uses of red ink on Mons Claudianus are the inscription on the end of the 18 m long column abandoned at the head of the “Pillar Wadi” (Chapter 1, pp. 38 f.) and some ostraca that all seem to have been written in the quarries. They are lists of personnel present in a given quarry on a given day. They are often written in charcoal or in red ink, the only writing materials at the disposal of the person taking the count at the work-place (see O.Claud. IV, p. 11 and nos. 636, 644, 659, 660, 679, 681, 686, 740, and 818). 6. Chapter 31, pp. 490–99. 7. Chapter 11. 8. Chapter 4. 9. At the most, the Historia Augusta attributes to him the paving of the courtyard of the imperial palace with opus Alexandrinum, which means a mixture of the Egyptian red porphyry and the green porphyry of Sparta. This innovation is also attributed to Elagabalus (Anton. Elag. 24.6). 10. Maxfield and Peacock 2001: 18 and 21.

A dedication to Zeus Helios Great Sarapis in honor of a desector

Figure 50. O.Claud. inv. 7363. © A. Bülow-Jacobsen

4

8

12 m.2

1 διϊ

10 ϊτουραιων

Διὶ Ἡλίῳ Μεγάλῳ Σαράπιδι καὶ τοῖς συννάοις θεοῖς ὑπὲρ διαμονῆς τοῦ κυρίου ἡμῶν Αὐτοκράτορος Μάρκου Αὐρηλίου Σε̣ο̣υ̣ήρου̣ Ἀλεξάνδρου Εὐσεβοῦς Σεβαστοῦ καὶ ὑπὲ̣ρ̣ εὐχαριστίας τοῦ ἐναρέτου κυρίου δησέκτορος κρ ν̣ο̣δ̣ος ἐ̣τ̣ε̣λ̣έσαμεν τ̣ὸ̣ν̣ κλῆρον ἡμῶν ἀμέμπτως χώρτης δευτέρας Ἰτουραίων [2–4?] (ἑκατονταρχίας?) Ἀμ̣μωνιανο̣ῦ̣ vac. ε Καῖσαρ Ἀλέξαν[δρε?] c. 12  Καῖσαρ ἈλέξαTraces of 2 lines broken to the left.

215

216

Rome in Egypt’s Eastern Desert

“To Zeus Helios Great Sarapis and to the gods who share his temple, for the continuance of our Lord the Emperor Marcus Aurelius Severus Alexander Pius Augustus, and as a token of our gratitude towards our valorous master the desector … We, of the Second Cohort of Itureans, century(?) of Ammonianus, have irreproachably accomplished the duty assigned to us. “Lord Caesar Alexander! … Lord Caesar Alexander!” 1.

For this god, the real tutelary deity of the desert from the reign of Trajan, see Chapter 31, pp. 523–26, and Tallet 2011.

3–5.

This titulature corresponds to the 22nd for this emperor in the list of Bureth, except that Εὐσεβοῦς has been omitted.

6.

ὑπὲρ εὐχαριστίας τοῦ κτλ. This expression is first of all found in Egyptian inscriptions. Perhaps it is influenced by the ὑπέρ-formula of allegiance to the Emperor, which is probably of Alexandrian origin. Elsewhere, to express gratitude, εὐχαριστίας ἕνεκεν is used. The formula normally expresses the gratitude of the dedicant towards the deity, a gratitude that is physically expressed by the inscription itself. It is very rare that a genitive depends on ὑπὲρ εὐχαριστίας, and I have found only one occurrence in Searchable Greek Inscriptions from Packard Humanities Institute, namely the two versions found in Axum of the inscription of Ezana, Emperor of Ethiopia (fourth–fifth cent.) republished as RIÉth 270 and 270bis. There we read, in lines 28–29 (= 26–27): ὑπὲρ δὲ εὐχαριστίας τοῦ μαι (l. με) γεννήσαντος ἀνικήτου Ἄρεως ἀνέθηκα αὐτῷ κτλ., “to witness my gratitude to my genitor the invincible Ares, I have put up this (inscription) to him…” In the ostracon, the gratitude of the dedicant is directed, not to the god, but to a man, of whom we only know that he was desector (n. ad 7). This detail places our document in the category of honorific inscriptions where one thanks people, using the verb εὐχαριστῶ/ εὐχαριστοῦσι, a formula that fixes on stone an actual acclamation: cf. L. Robert, Hellenica X, 46–62, especially 55 and 61. Robert analyzes one of these honorific inscriptions, TAM IV.1 39, dedicated by two soldiers and some mule-drivers (probably employed in the cursus publicus) to an ἐπιμελητὴς κτηνῶν Καίσαρος. Contrary to our ostracon, this inscription is not a dedication to a god, but the text is accompanied by a relief of Zeus, in whose sanctuary the dedicators probably placed their inscription and where the acclamation took place.

6–7.

ἐναρέτου. This adjective, used as an honorific epithet, is not found in Egypt before the fourth century (for its use in protobyzantine papyri, see J. Diethart, Tyche 10 [1995] 241). On the other hand, it is found in inscriptions of the principate from Asia Minor, where it qualifies the words ἀνήρ, γυνή, and πρόγονοι, as well as abstract words for moral qualities (Chr. Habicht, Die Inschriften des Asklepieions, Altertümer von Pergamon VIII.3 [Berlin 1969] 161).

7.

δησέκτορος. This Latin loan-word is attested only in two papyri from the fourth century, neither of them providing a helpful context: – SPP XX 75, i.22 (Hermopolite): φορέδρου (l. φορέτρου) ὄνου ἐγκαρίας (l. ἀγγαρείας) τῷ στρατιώτου (l. -τῃ) ἕνεκεν τῶν δισεκτόρων ὄνος α (τάλαντα) κ, “price of requisition of a donkey for the soldier for the sake of the disectores: 1 donkey, 20 talents.” The beginning of this document is lost, but it appears to be a list of payments, some of them

A dedication to Zeus Helios Great Sarapis in honor of a desector

217

expressly to military personnel, and ends in the form of a letter. Among the payments there are salaries to several workers (three or four) at the κρηπίς (the context does not allow us to decide what kind of construction this might be), e.g., ii.11: κρεπῖδαν (l. κρη-) ὁμοί[ω]ς ἐργάτ(αι) γ ἐπὶ μῆναν (τάλαντα) ξϛ (δραχμαὶ) Δ. – P.Erl.Diosp. 1.148 is also from the fourth century and is a list of cash payments. In lines 3 and 4 the beneficiaries are respectively “those from Maximianopolis” (i.e., Kaine, where the roads to Mons Claudianus and Porphyrites began), and “Psenthous, the Coptite.” In lines 7–8 we see τοῖς ὀφ(φικιαλίοις) τοῦ δοῦκ(ος) προφάσει τοῦ δησέκτορος Ἠλ(ίας) ἀπ( ) (τάλαντα) μ, “for the officiales of the dux concerning the desector Elias apaitetes(?) 40 talents.” The Lex.lat.Lehn. (s.v. δισέκτωρ) stresses that the two spellings are contradictory: δη- would refer to a Latin word in dē- (*dēsector < dēsecare), while di- suggests that the corresponding word begins with dĕ- or di. The Lex.lat.Lehn. discards *dēsector, preferring *disiector, a verbal noun that would derive from disicere. The disiectores would thus be an advance guard meant to disperse the advance guard of the enemy. I, myself, think that the existence of two verbs of similar meaning, dēsecare and dissecare, could explain the two spellings in Greek. Fr. Mitthof (P.Erl.Diosp. 1.148n.) proposes another hypothesis based on the occurrence of the word δισεκτορία in Passio S. Eusignii (P. Devos, “Une recension nouvelle de la Passion grecque BHG 639 de saint Eusignios,” Analecta Bollandiana 100 [1982] 215). In the two manuscripts which appear to him closest to the archetype, the phrase runs as follows: Ἑξήκοντα γὰρ ἔτεσιν ἐστρατευσάμην ἐπὶ Κωνσταντίνου τοῦ μεγάλου βασιλέως καὶ ἐπὶ Κωνσταντίου τοῦ πατρὸς αὐτοῦ, καὶ οὐχ εὑρέθην ποτὲ βακαντῖβος ἢ δισεκτορίαν πράξας, “I have served as a soldier for sixty years under the great Emperor Constantine and under his father Constantius, and I have never been found cowardly nor have I committed disectoria.” Mitthof interprets δισεκτορίαν as “desertion” on the basis of a scholion on this passage that is found in one of the manuscripts: φυγὰς ἢ λιξουράρις γενόμενος ἐν τῷ στράτῳ, “deserter or luxurious in the army.” Δισέκτωρ/δησέκτωρ would be a deformation of desertor. This meaning is obviously excluded in the ostracon, and I rather think it may be a craft related to the extraction of stone (cf. TLL s.v. desecare II.A seiungere partem aliquam): “the extractor.” But there is only one attestation and a poetic one of desecare in this context, Stat. Silvae 2.2.85: hic Grais penitus desecta metallis saxa. What is more, this attestation is based on a conjecture, and though it has found its way into the Latin Dictionary of Lewis & Short, it has generally not been favored by editors who have mostly preferred the delecta of the manuscripts. Nevertheless, the recent edition by E. Courtney (Oxford, 1990) returns to desecta in analogy with Silv. 1.5.36–37: sola nitent flavis Nomadum decisa metallis marmora. I would add that the word penitus is often associated in the literature with the concept of extraction of materials, see especially Cic., 2Verr. 5.68 (about the quarries: totum est e saxo in mirandam altitudinem depresso et multorum operis penitus exciso). Could the desector have been an engineer specialized in extraction? Perhaps the meaning “cut” could also be accepted in the case of disectoria, from which Eusignios tells us that he has always abstained? It is known that in late antiquity young people cut off a finger in order to avoid conscription. The author of P.Herm. 7 had tried this trick around 381, according to the explanation of C. Zuckerman 1995, esp. 183–8; cf. Chapter 32; see also F. Dunand and R. Lichtenberg, “Des réfractaires à l’enrôlement? Plusieurs cas d’automutilation dans une nécropole

218

Rome in Egypt’s Eastern Desert égyptienne,” BIFAO 119 [2019] 115–23). If this interpretation is correct, our case would be a little different, since Eusignios would not have been trying to avoid conscription, but to be on the sick-list before a battle.

7–8.

In the illegible space between δησέκτορος and ἐ̣τ̣ε̣λ̣έσαμεν we expect the name of the desector. It would be strange for this to have been left out. The subject of ἐ̣τ̣ε̣λ̣έσαμεν may also have been there, but could have been omitted. κρ. After κ one might read α or λ. καί is possible and would introduce the name of another craft or function which would have been followed by the name of the desector. After this we should put a full stop, since a new sentence would have begun with ἐ̣τ̣ε̣λ̣έσαμεν.

8.

ἐ̣τ̣ε̣λ̣έσαμεν. Instead of λ one might read δ, but τελέω is more attractive because of the object κλῆρον, which here means “task completed” (see the following note). κλῆρον is not elsewhere particularly associated with this verb. I find only SEG XV 698.234 and passim (ἐπιτελεῖν κλῆρον “organize a lottery”) and, closer to our meaning here, “accomplish your allotted destiny” (M. Sayar, Perinthos-Herakleia, no. 219).

9.

κλῆρον. In some ostraca from Mons Claudianus κλῆρος is employed in connection with soldiers, but mostly in the sense of assignment to a given place (e.g., O.Claud. II 379). In the present case, the verb τελέω encourages us to understand κλῆρον like ἔργον, the task being to execute a command. In PSI VII 822 (second cent.) we also find κλῆρος in the context of exploitation of quarries (of alabaster), but there the word is applied to the zones of extraction to which the teams were assigned. In any case, these uses of κλῆρος are derived from the general sense quod sorte obvenit (Th.Gr.L. 1635 A). There is also a mention of κλῆρος in a collective dedication by soldiers found in the chapel of the praesidium of Didymoi (I.Did. 5), but the meaning is unclear.

9–10.

χώρτης κτλ. How should we explain the genitive? There are no parallels for ἀμέμπτως + genitive (one might have thought of a meaning like “in such a way as not to put a blemish on the cohort”), and there is a single example from Aeschylus (Pers. 692) where a genitive depends on ἀμέμπτως. I prefer to take it as the identification of the unit of dedicants, which goes well with the absence of a definite article before χώρτης. Grammatically the genitive depends on ἡμῶν. The Notitia dignitatum mentions the cohors II Ituraeorum, which had never left Egypt, and which was, at least in 138, based at Pselchis in Lower Nubia. On this unit, see Daris 1988b: 759.

11.

(ἑκατονταρχίας?). Only the top of the χ and the superscript ρ are visible. If the conjecture is correct, the soldiers detailed to Mons Claudianus thus all came from the same century.

12.

m.2. Properly speaking, there is no change of scribe, only of writing-style. κ̣ύρ̣ι̣ε̣ Καῖσαρ is tempting and corresponds better to the traces than χαῖρε Καῖσαρ. This formula of address is found in acts of martyrs and in P.Fouad 8.11 (which, according to O. Montevecchi, should be a fragment of ἄκτα τῶν τιμῶν a “report on honors” for the use of imperial propaganda: “Vespasiano acclamato dagli Alessandrini,” Aegyptus 61 [1981] 155–70). On the other hand, the presence of a cognomen is unexpected. We do not know whether what followed still concerned the Emperor. It seems that after Ἀλέξαν[δρε] the line of writing slopes

A dedication to Zeus Helios Great Sarapis in honor of a desector

219

slightly upwards. The emperor is again called on in the following line. It is possible that the formula is ritually repeated three times. The text in chancery hand would thus report the triple acclamation of the group of dedicators during the religious ceremony by which it celebrated the end of their mission on Claudianus. Cf. Weinreich 1916–1919: 134 (= Ausgewählte Schriften 1: 465 ff.): “Man denke vor allem auch an die unendlich häufige Dreimaligkeit von Kult- und Zauberbräuchen, in mythische Triaden, an die kunstmässig oder rituell oder volkstümlich verwendete dreimalige Wiederholung von einzelnen Worten, ganzen Sätzen, (…).” Some inscriptions, where letters from the Emperor or an important person are copied, reproduce, as would a report of proceedings, the acclamation of the public who were present at the reading (L. Robert, Hellenica, X, 61 f.; Corbier 2006: 26, n. 40). I know of no examples from religious dedications. This acclamation replaces the “acclamation-wish” (L. Robert) ἐπ’ ἀγαθῷ that ends so many dedications from Egypt.

13 An unpublished ostracon from the Eastern Desert and the provenance of O.Amst. 9* The ostracon1 published below was found in January 1991 in the dump which extends in front of the door of the small fort of Al-Zarqaʾ (Maximianon): it had been disengaged by the heavy rainfalls of 31 December 1990, which had re-dug the holes made by illicit excavators. It was entrusted to the Council of Antiquities. Al-Zarqaʾ is the best preserved of the eight stations that bordered the ancient caravan route linking Qusayr (Myos Hormos), on the Red Sea, to the river port Koptos.2 This small fort, with its square ground plan (60 × 60 m) and its corner towers, is built from shale and surrounds a collapsed well. The document is complete. It is a list of 13 names, each preceded by a number, from 7 to 19 in ascending order: these 13 men were needed for a service which they were to complete one by one on the 7th to 19th days of an unspecified month. The text would be trivial if it weren’t part of the same file as O.Amst. 9, a list of differently organized rounds of service, which brings together more or less the same staff. The fact that we find Νεκρῦς and Φαρικός, both practically hapax legomena, in both lists, leaves no doubt. W. Clarysse has shown, from a prosopographical study, that the military ostraca in Amsterdam (notably lists of guards, O.Amst. 8–15 and 82) belonged to an important file dispersed across the world, for a large part published (cf. especially the O.Amst. and the O.Florida), and originating from the same place.3 But where? He himself suggested, based on toponymic criteria, the Theban left bank.4 The previous editors of these ostraca, following the contradictory declarations of the antiquarian in Luxor who was selling them in the 1970s, assigned Contrapollonos polis Magna (which faces Edfu) and Laton polis * Translated by Mathilde Bru. 1. Now SB XXII 15541. 2. Recent descriptions: Reddé and Golvin 1986: 182–85; Reddé and Golvin 1987: 14–15; Zitterkopf and Sidebotham 1989: 175–76. 3. Clarysse 1984. 4. Clarysse and Sijpesteijn 1988; on provenance, 89–92.

223

224

Rome in Egypt’s Eastern Desert

(Esna) as their findspot.5 When I presented the ostracon from Al-Zarqaʾ at the Copenhagen congress, I thought that I had settled this enigma once and for all: I announced that the O.Amst. and the O.Florida had come from the dump of this site, the surface of which was moreover riddled with small craters. The two excavation campaigns at Al-Zarqaʾ refuted this appealing hypothesis. Although the names of Nekrys, Pharikos, and their comrades often resurfaced under our trowels, none of the personal or place-names that W. Clarysse had invoked in his demonstration appeared in our texts, which number more than 1500: O.Amst. 9 does indeed originate from Al-Zarqaʾ, but it is an intruder in the series of military ostraca from Amsterdam. In 2020, the provenance of the military ostraca from Amsterdam, Florida, etc. remains unknown. Most ostraca of “Nekrys’s dossier” (as it is dubbed in my database) are fragments of duty rosters of the same type as O.Amst. 9: names are arranged by four for each day and preceded by a Latin number from I to IV, referring to tours of duty (the presentation of our text is different). All these ostraca belong to phase B5 of the dump and thus date to the end of the second century or the beginning of the third.6 The names are those of soldiers who belonged to one of the last garrisons which manned Maximianon. In the first version of this chapter, I could, thanks to the ostracon, correct several names in O.Amst. 9. Another name that often occurs in the dossier, Παπυρᾶς̣ is to be recognized in O.Amst. 9.8 instead of Πχηυρο̣ς̣.

Figure 51. SB XXII 15541. © A. Bülow-Jacobsen 5. In 2003, G. Nachtergael challenged Clarysse’s demonstration and gave credit back to the former hypotheses of Laton polis and Contrapollonos polis, with a preference for the latter (P.Hombert II, p. 12). 6. Cuvigny (ed.) 2006: 124.

An unpublished ostracon from the Eastern Desert

225

SB XXII 15541. Fig. 51. 10.5 × 13 cm. Fragment of AE3 amphora. Against the right edge, faint traces of the beginnings of three lines belonging to a second column.

4

8

12

ζ̅ η̅ θ ι̅ ια̅ ιβ̅ ιγ̅ ιδ̅ ιε̅ ιϛ ιζ ιη̅ ιθ̅

Ἀπολινάρις ι̅ Νεκρῦς Τιθοῆς Γερμανός Μάξιμος Ὡρίων Ἁρποχρατ(ίων) Ψεντάησι(ς) Ἄρης Ἑρμίας Βῆκις Φαρικός Οὔλπις

1.

I do not express an opinion concerning the iota topped by a horizontal line, identical to the number on line 4, which follows the name (due to the oblique upper border it is written between lines 1 and 2): (δεκανός), (δεκανία) are sometimes represented in this way, but more typically by an elongated vertical stroke topped by a “hat.” Anyway, since these men are soldiers, this resolution must be rejected.

2.

Νεκρῦς: to be read instead of Νερ̣ρης in O.Amst. 9.13. I have, in vain, asked several Egyptologists to help me get rid of this dismal name. Michel Pezin, who kindly analysed it, eventually rejected the idea that it is one of those Egyptian names in nḫt (“powerful is”) transcribed in Greek Νεκθor Νεχθ-, or even Νεκ- (cf. Νεχφερως, Νεκσουχος). With this hypothesis set aside, it must be a new Greek hypocoristic form in -ῦς (cf. O. Masson, OGS I, 93, n.6 and E. Lhôte 2007). Eric Lhôte suggests (email March 20, 2020) that Νεκρῦς could belong to the category of apotropaic names. It could also be a descriptive nickname (such as Lemort in French). Thanks to O.Max. inv. 227, we know that it is actually the second name (or nickname) of an Ἀπολλῶς, obviously to distinguish him from a homonym present in the same list, Ἀπολλῶς Σα[. While names in -ῦς are rather rare, the duty-rosters of Nekrys’s dossier offer two other examples: Κα(ε)ινῦς (inv. 11 and 1306) and Μαγ̣νῦς (inv. 99, unless Μαι̣νῦς should be read). Μαγ̣νῦς appears, as in the case of Ἀπολλῶς Νεκρῦς, as a second name: ]μος Μαγ̣νῦς). This garrison seems to have a penchant for the suffix -ῦς.

4.

Γερμανός. To be read instead of Πρ μ ̣ ̣νος in O.Amst. 9.10.

6.

Ὡρίων. I am very tempted to restore this name in O. Amst. 9.14 (ed. [ι̅β̅ Ι ]ν). It seems clear to me that there are three and not two letters in the lacuna, and that the ν is preceded by an ω which is inclined to the left like those of Ὡρίων here: Ὡρί]ων.

226

Rome in Egypt’s Eastern Desert

7.

Ἁρποχρατ(ίων). This is the Ἁρποχρᾶς in O.Amst. 9.4 but the photo on the plate gives rather the impression that at least αρποχρατι is written. In the other ostraca of Nekrys’s dossier, the name is abbreviated Ἁρποχρατ( ), -τι( ), -τιω( ) and never occurs in the form Ἁρποχρᾶς.

8.

Ψενταησι. Final ι elongated to the bottom, which indicates that the name is abbreviated; the scribe has written Ψενταήσιος in O.Amst. 9.

9.

Ἄρης. In Nekrys’s dossier, the name is spelled Ἄρης three times, and ῎Αρειος twelve times, as well as in O.Amst. 9.12 (BL IX, 377).

12.

Φαρικός. One of those names suffixed in -ικός from the Imperial Period, probably derived from the Pharos. To my knowledge, this name only appears elsewhere in the scholium 398 of Nicander’s Alexipharmaka: there we find that, according to Praxagoras, the poison called φαρικόν derives its name from its inventor, a Cretan named Φαρικός.

14 Kaine, a new town: an experiment in familial reunification in the second century ad* In the quarries of Mons Claudianus, the largest part of the specialized workforce (quarrymen and stone carvers combined under the name of σκληρουργοί on the one hand and blacksmiths on the other) was made up of Egyptian artisans and formed a community referred to by the administration as pagani (“natives”). The pagani community itself was largely divided into two main groups, the Syenites and the Alexandrians. On a Trajanic ostracon, datable from the stratigraphy and certain prosopographical overlaps to the years 110–115, the Syenite artisans numbered 130, and the Alexandrians 210.1 Work in the metallon was in full swing, as it was at the time of the large commissions of pillars for Trajan’s Forum. Under Antoninus, the extraction continued, albeit perhaps on a smaller scale. At the same time,2 novelties in administrative practice were introduced, thanks to which we know what the salary of the pagani and its distribution method were at the time. Each month they earned a salary, appealing by the standard of the time, in cash payment, as well as a ration of unprocessed wheat (Chapter 9). This money and wheat were not, however, allocated to them directly at the metallon by the imperial administration. Each month, the workers delegated a few among them to “go down” (καταβαίνειν) to the Nile valley to collect the wages and grain rations. These representatives assumed the title of κιβαριάται (“food-supply stewards,” from the Latin cibaria, food) for the occasion. Yet the role of the kibariatai did not stop there: their colleagues would charge them, through lists of instructions inscribed on ostraca and titled ἐντολή (“instructions”), to proceed with multiple transactions (loans, debt reimbursements, purchases…), which greatly reduced their cash credit and even often exceeded it. As for the wheat, the kibariates rarely brought it back as it was to Mons Claudianus: more often than not, he had the task of delivering it to the mother or the wife of the worker, who ground and baked it into bread. The kibariates would bring back this bread, placed in sacks on which was attached a label with the name of the worker,3 to the metallon, * Translated by Mathilde Bru. 1. Chapter 11, pp. 186, 195. 2. To be precise, the new system was put in place under Hadrian (Chapter 9, n. 5). 3. Chapter 7, p. 138.

227

228

Rome in Egypt’s Eastern Desert

along with other food supplies which the worker had ordered: almost always dates, lentils, or onions bought in the market or else homemade goods (vinegar, meat, or fish in brine). Should we then understand that the kibariatai would make their way to Alexandria or Syene each month in order to make contact with their mandators’ families? An organization of such a kind, needlessly time-consuming and costly, is implausible. A small number of Antonine ostraca – letters and entolai – give us the answer to this question. The names of Syene and Alexandria are never mentioned in texts emanating from pagani: the families’ places of residence, in the rare occasions when they are specified, are always Καινή and Ἀπόλλωνος:

Καινή O.Claud. inv. 8598: entole (1. 2) (ἀρτάβη) ἰς Καινὴ⟨ν⟩ δι[ὰ] Ἴ̣σιτος “(carry) the artaba to Kaine, to be made into bread by Isis.”4

O.Claud. inv. 4312: entole of Tithoes son of Herakleides, “Syenite” (l1. 5–8) τὰς λυπὰς (sc. δραχμάς) ἕως καταβῶ κατενήνκις5 μοι εἰς Καινὴν οἴνου κ(εράμια) γ. l. λοιπάς, κατενέγκεις

“you will pay me my credit balance at Kaine when6 I come down. Three amphoras of wine” or “(keep) my credit balance until I come down. You will bring to Kaine 3 amphoras of wine.”

Parallels do not allow us to decide between the two interpretations. Either way, the worker Tithoes has in mind “to come down” shortly from Mons Claudianus, and it is at Kaine that he demands to receive his salary or his wine. O. Claud. inv. 7863: entole (l1. 9–12) γυναικὶ (δραχμὰς) ιδ καὶ λ[αμ]βάνις ἐπιστολὴν ἢ [ἀ]π̣[οχ]ὴν ἐν Καινῇ σεῖ̣το⟨ν⟩ τ̣ῇ̣ γ̣υ̣ναικί, ἄρτοι ἰς ὄρος. “To my wife 14 drachmas, ensure that you are given a letter or a receipt(?) at Kaine, wheat to my wife (or: “ensure that you are given a letter or a receipt(?), wheat at Kaine to my wife”), the bread to the desert.”

4. In the entole O.Claud. inv. 6607, this Isis makes the bread of another worker. 5. There is some difficulty surrounding κατενήνκις: this verb suggests that the kibariates “will bring down” the money or the wine to Kaine. But from where? From Mons Claudianus? From a town on the banks of the Nile upstream from Kaine? Unless, however, one should interpret καταφέρειν in the sense of “to put down, discharge,” which the verb sometimes has. 6. ἕως καταβῶ here stands for ὅταν καταβῶ.

Kaine, a new town

229

O.Claud. inv. 4674: a letter of which the content suggests that its author, Psenpnouthis, is a kibariates; in that case Horion, the recipient, would be an artisan paganus (l1. 4–7) ] ἡ̣μ̣ι̣αρτάβιν σείτου δοῦναι τῇ γυναικί σου ἰς Κενήν l. Καινήν

“[you have written to me (?) to] give a half-artaba of wheat to your wife at Kaine.”

O.Claud. inv. 6027: letter from Soterichos to Serenus7 (l1. 6–7) ἐλθὸν δὲ εἰς Κενὴ⟨ν⟩ εὗρεν τὴ̣ν̣ μητέραν σου l. ἐλθών, Καινήν, ηὗρον

“Having arrived at Kaine, I found your mother…”

Ἀπόλλωνος O. Claud. inv. 5283: entole of Isidoros, son of Apion, “Syenite” (l1. 5–7) τὸν σεῖτον ἰς Ἀπόλλωνος Σενπαχούμι καὶ ὃ ἄν σε αἰτήσῃ κοπ⟨τ⟩ούρας “(hand over) to Senpachoumis, at Apollonos, the wheat and whatever she should ask you for the milling”

Known from other entolai, Isidoros always uses a baker foreign to his family and therefore has to pay her. O.Claud. inv. 4339: entole (l1. 2–3) τὴν ἀρτ[άβην - - -] ἰς Ἀπόλ[λωνος “my artaba … to Apollonos.”

The identity of the baker is in lacuna. O.Claud. inv. 8455: entole of Didymos son of Apollonios, “Alexandrian” (1. 3) ὁ σεῖτος ἰς Ἀπόλλωνος Σεραπιάδι ε̣  ι̣ς̣, ἄρτοι ἰς ὄρος “(hand over) my wheat to Serapias … to Apollonos, bring the bread to the desert.”

7. This name is that of two authors of entolai.

230

Rome in Egypt’s Eastern Desert Ο.Claud. inv. 5123: entole of the pais8 Isidoros, “Alexandrian” (1. 6) ὑπόδημ(α) ἀπὸ Ἀπόλ(λωνος)

“(bring back) a pair of shoes from Apollonos”.

O.Claud. inv. 2568: entole (1. 2) μητ(ρὶ) ἰς Ἀπόλλω̣ν̣[ “to my mother at Apollonos”

The identification of Καινή leaves no doubt: the name, which means “new (town)”, survives to this day in the form Qena. Καινή/Qena is at the origin point of the Roman roads leading towards Mons Claudianus and Porphyrites (and a few other areas of lesser importance). We do not know of any ancient remains there, 9 and it is not attested in sources written before the Roman period: it is highly probable that its foundation on the right bank of the Nile, facing the ancient city of Tentyra (just as Villeneuve-lèsAvignon faced the city of the popes), is contemporary with the opening of the two metalla10 and their adjacent settlements. Kaine was not only the port in which ships that were specially built to transport monoliths docked, and where caravans heading to the quarries were formed: we have seen how the workers, or at least some of them, had a family there. Unlike in the mines of Vispaca, women, children, and elderly parents were precluded from being admitted into the metalla in the Eastern Desert, given the difficulties of procuring water and the need to import all food; families, generally from Alexandria and Syene but also possibly from elsewhere,11 were therefore settled in the Nile valley, as close as possible to where the men worked. A letter, of which the author, Tertius, is possibly a soldier,12 gives a brief view of urbanism at Kaine in the second century: Tertius asks his addressee to take with him various objects to Kaine to give to his wife, whose address he indicates: εἵνα ἀποδοῖς εἰς Καινὴν Ἡρακλοῦτι πρὸς τὸ κουρίῳ ἰς τὴν συνοικίαν ἐχόνομα τῆς φόσσης l. τῷ κουρείῳ

“In order that you may hand them over to Heraklous, near the barber’s shop, in the building next to the canal, at Kaine” (O.Claud. inv. 8208).

8. Probably not a slave, but an apprentice. 9. Porter and Moss 1937: 122 do not indicate any architectural remains, only a small number of movable objects (stelai, tables of offerings), a few of which date from the pharaonic Period. Some of these objects are known to have been bought in Qena, but this does not guarantee their origin. 10. Mons Claudianus was supposedly opened by Claudius, in accordance with its name (the most ancient dated text from Mons Claudianus dates to the time of Nero); Porphyrites was “discovered” in Tiberius’s second year (Van Rengen 1995). 11. The onomastics of the “Syenites” is actually more characteristic of Thebes than of Syene. 12. His Latin name, the fact that he demands fresh vegetables, a luxury that, in the desert, was reserved for soldiers, and the use of a Latin word when referring to a canal.

Kaine, a new town

231

Συνοικίαι, investment rental properties, were up to now only attested in Egypt in Alexandria and in several towns of the Fayyum;13 this single mention does not, of course, allow us to know whether this type of housing was widespread in Kaine and provided an urban response to the influx of migrants living in the periphery of the metalla. In the O.Claud., Apollonos is nevertheless as much attested as Kaine as the terminus of an itinerary regularly completed by men and their goods. Among the six towns of that name recorded in Calderini’s Dizionario, three are located in Middle or Upper Egypt; these are, from the north to the south: – Ἀπόλλωνος πόλις Ἑπτακωμία (Kum Isfaht, c. 195 km to the north of Qena)14 – Ἀπόλλωνος πόλις μικρά (Qus, c. 30 km south of Qena) – Ἀπόλλωνος πόλις μεγάλη (Edfou c. 170 km south of Qena) If our Apollonos is not the name of a hitherto unknown site, Qus seems to be the most satisfactory possibility: (1) The town is close to Qena; (2) it was possible, when departing from Qus, to join the road of Koptos to Myos Hormos (during the Islamic period Qus replaces Quft as terminus of the Red Sea road); (3) the Roman road of Koptos to Myos Hormos, the ὁδὸς Μυσορμιτική, is not only a road for the caravans of eastern trade, but it also passes through a zone of mines and quarries, which explains why families of quarrymen and miners may have lived at Qus;15 several ostraca from Claudianus seem to suggest that transfers of manpower sometimes occurred between metalla: families did not necessarily follow the workers, and it must have been the case that some of the men of a family would work in metalla on the road of Myos Hormos and others in metalla on the ὁδὸς Κλαυδιανοῦ or the ὁδὸς Πορφυρίτου which converged on Kaine. Qus, unlike Qena as I view it, was not established by the Romans, but already existed in the Old Kingdom. Our documentation does not, however, explain why families of quarrymen would have settled in Apollonos polis Parva and not in Koptos, a more natural terminus of the road, and closer to Kaine. But perhaps the quarrymen whose families lived in Qus were actually from this town, not far from which are limestone quarries (Chapter 31, p. 481). The displacement of the population is not an exceptional phenomenon in this context of mining; the most famous case is that of the gold mines of Dacia where, in the larger context of the colonialization of a whole province, which had been disrupted and depopulated by war, a workforce experienced in working in mines and originating from Dalmato-Illyrian tribes (Pirustae and Baridustae are mainly mentioned) was settled. This workforce formed groups by ethnic origin in agglomerations called vici or castella,16 whence the name of castellani which appears to designate the inhabitants.17 Sources tell us 13. Husson 1983: 271–75. 14. The only link found between this Apollonos polis and the metalla is P.Giss. 69 (AD 118), an urgent demand addressed to the strategos of the Apollonopolite to send to Kaine all the barley available in his nome in order to feed the animals gathered to transport a 50-foot column. This text does not suppose a special relationship between Kaine and Apollonos polis Heptakomia: the food intended for fueling the quarries could come from any part of Egypt, for example from the Fayyum (SB XIV 12169 [AD 96]). 15. In Itin. Anton. 165.5, Qus is called Vico Apollonos. Vici was the name of certain agglomerations where Dalmatians who had immigrated to Dacia were assembled (other agglomerations were castelli); this is no doubt a coincidence, and the vico of Itin. Anton. must simply correspond to the Greek κώμη. It is noteworthy that the ostraca found in the sites on the Myos Hormos road never mention Apollonos, but only a tiny number of them originate from a milieu of quarrymen. 16. Cf., e.g., the vicus Pirustarum, the name of a district of Alburnus Maior. On this subject and for other examples of migration of mining workforce, see Andreau 1990: 90–92. 17. The noun appears only in CIL III 7821, where it is written out in full, which allowed us to restore k(astellum), always shortened as K, in other inscriptions (Wollmann 1989: 112).

232

Rome in Egypt’s Eastern Desert

nothing about the methods of this migration: was it forced, spontaneous, or encouraged? The problem is clearly laid out by Noeske, who favors the idea that the Dalmatians of all strata of their ancestral society spontaneously emigrated – not only manual workers but also entrepreneurs and businessmen.18 Moreover, we do not know whether the Illyrian miners came with their families, although we can nevertheless suppose that this was the case. A few inscriptions and wax tablets from Alburnus Maior mention women of Illyrian origin, wives and mothers, but we suspect that they belonged to the more privileged strata of society, which could afford tombstones or to buy, like Andueia daughter of Bato, half a house (CIL III, p. 944, TC VIII). Even if an authoritative transfer of populations did not occur, they were not lacking incentives: the rather appealing salary of the workers, which, as I have shown in Chapter 9, seemed to be on the same scale as at Mons Claudianus, must have played a part in these incentives. In the jargon of Mons Claudianus, the use of a word borrowed from the Latin (κιβαριάτης) to refer to the backbone of the system of food supply and transport suggests that imperial administration intervened to help the community of the pagani to establish this self-directed structure which included women. The migration of these wives, mothers, daughters, and sisters, and the organization of their lives in a new environment where the head of the household was simultaneously nearby and absent, could not have happened without the help of the State. Both parties had something to gain. The emperor saved money in administration, only providing each month the payroll and the wheat; workers profited from a personalized service thanks to which they received their salary à la carte, so to speak, and kept in close contact with their families: we do not know how regularly they were permitted to go down to visit their families, but they at least had the satisfaction of eating homemade bread every day (which provided the further advantage of limiting the importation of foodstuffs to the metallon). The Claudianus-Kaine track, on which kibariatai who had been entrusted with instructions, letters to family members, and homemade bread went up and down, represented a true umbilical cord for separated families. This psychological aspect should not be neglected: from Pharaonic times to nowadays, Egyptians are famous for their repugnance for the desert. This is but a syndrome of depression and anxiety well-known in small isolated communities in difficult environments: the “outpost troubles.” States and businesses should not skimp on the ways to help individuals who are sent to the desert, on oil rigs or even into space, to overcome the stress of solitude and homesickness. The imperial orders of monoliths, exclusively intended for prestigious buildings of an ideological purpose, suffered no delays or flaws: the system established at Mons Claudianus shows that emperors were conscious of the need to handle the qualified workforce in a way that they could work at their best expeditiously.

18. Nœske 1977: 342.

15 The road system of the Eastern Desert of Egypt under the Early Empire in the light of the excavated ostraca In 1942, Octave Guéraud published a group of military ostraca dating to the Early Empire, which the Egyptian Mining and Prospecting Company had just found in Wadi al-Fawakhir.1 This was a revelation: for once, ostraca were not off-putting tax receipts, but for the most part Greek and Latin letters exchanged between soldiers. The O.Fawakhir long remained one of the principal documentary sources for describing the daily life of the Roman soldier. Oddly, and despite the fact that the Eastern Desert had drawn the attention of travelers curious about antiquity and epigraphy since the nineteenth century, no one thought to pull on this loose thread, perhaps because the logistical difficulties of creating and maintaining a camp for an excavation team in the desert seemed insurmountable. Or should we find the explanation in a hesitation by papyrologists to excavate for texts, as they did in the era when Grenfell and Hunt discovered the vast papyrus collection of Oxyrhynchos? Are papyrologists of today paralyzed by the progress and the demands of archaeology? It was only in 1987 that an international team, directed by Jean Bingen, and of which I was a member, undertook to study another Roman site in the Eastern Desert, Mons Claudianus, carrying out a project that consisted of excavations guided by a purely archaeological agenda2 and the systematic exploration of the dumps: for it is almost entirely from the dumps that the ostraca of the Eastern Desert come.

1. Guéraud 1942. These ostraca come from a praesidium built in Wadi al-Fawakhir that has now completely disappeared; its existence is deduced from the ostraca of Krokodilo, thanks to which we know that it was called Persou (Cuvigny [ed.] 2006: 284 f.). 2. This included detailed description and interpretation of the remains, and the establishment, through observation and targeted trenches, of at least a relative chronology of the various stages of construction. At Mons Claudianus, this work was carried out under the direction of David Peacock (University of Southampton) and Valerie Maxfield (University of Exeter).

233

234

Rome in Egypt’s Eastern Desert

We worked at Mons Claudianus from 1987 to 1993. More than nine thousand Greek and Latin ostraca were registered, overwhelmingly Greek.3 Hardly any papyrus has been found in the Eastern Desert: some references in the letters suggest it must not have been a high-priority item among the supplies brought by the caravans that supplied the desert posts; soldiers and civilians thus had adopted the habit of writing on potsherds, with a marked preference for fragments of Egyptian wine amphoras made from alluvial clay, the surface of which was sufficiently smooth and somewhat porous, thus receiving the ink well without making the pen run. As the soldiers drank a lot of local wine, to judge from the quantity of fragments of amphoras, writing materials were not in short supply. After Mons Claudianus, the team dispersed. The British part carried out four years of exploration of “Mons Porphyrites,”4 then moved to the port of Qusayr al-Qadim (the Myos Hormos of Agatharchides, Strabo, and Pliny the Elder, as we now know). The papyrological documents coming from these two sites are still unpublished.5 Adam Bülow-Jacobsen and I proposed to some French archaeologists to apply the twin approach tested at Mons Claudianus to other Roman installations of the Eastern Desert, beginning with Al-Zarqaʾ (Maximianon), a fortlet on the road from Koptos to Qusayr, where the well-preserved dump showed holes where clandestine digging had taken place. Michel Reddé, who had already drawn up, along with J.-Cl. Golvin, the plans of the forts spread along this road, agreed to take part in the project, and Claude Blanc, who had worked at Mons Claudianus, put us in contact with JeanPierre Brun. Nicolas Grimal, director of IFAO at the time, encouraged our project and helped us find the financial means. And thus, we set foot in another administrative zone of the Eastern Desert of the imperial period: the Mons Berenices. Whether one is dealing with fortified villages guarding quarries or stations along the roads, the fortlets of the Eastern Deserts that we have excavated were always called, beginning in the Flavian period, praesidia, both in inscriptions and in ostraca.6 The garrison that they housed was, except at some periods in the metalla, under the command of a curator praesidii, whose rank is known only in the case of Rufus, signifer and curator of Didymoi (O.Did. 187). It was long thought that the function of curator praesidii was peculiar to the Egyptian Eastern Desert, but this is not the case: it appears in an inscription of Mauretania Caesariana that M. Christol dates to AD 40 or a bit later, on the basis of a correction not registered in Année Épigraphique.7

I. The administrative divisions In December 2009, our excavation of the praesidium of Xeron Pelagos was delayed some days because the antiquities inspector accompanying our mission did not know that Xeron was located in a different military zone from that of the praesidia where we had worked earlier; the road from Edfu (Apollonos polis Magna) to Marsa ʿAlam, which lies on the 25th parallel, marks the frontier between the military zone dependent on Hurghada (to the north) and that dependent on Assyut (to the south). Under the 3. Four volumes have appeared so far. Two important texts have been published separately, in Chapters 5 and 11. 4. As I have shown in Chapter 1, the expression Mons Porphyrites is a modern creation. In antiquity, this metallon was called simply Porphyrites. 5. According to the preliminary reports or publications in press, around 700 ostraca were registered at Porphyrites, about 20 in the road station at Badiya, and c. 850 at Myos Hormos. 6. On the other hand, the Greek dedication of the via Hadriana scrupulously avoided this Latinism, which it replaced with three Greek terms describing the principal functions of a praesidium of the Eastern Desert: ὕδρευμα (well), σταθμός (lodging), and φρούριον (guard post). 7. E. Albertini, in Bulletin archéologique du Comité des travaux historiques 1927: 74–76. I thank Michel Christol for having brought this inscription to my attention. It is discussed in Bernard and Christol 2008–2010, esp. 2215–20.

The road system of the Eastern Desert of Egypt

235

Early Empire, the desert was equally a military zone and seems also to have been divided into two large regions: (1) in the more northern, two important imperial quarries were located, that of Porphyrites and that of Mons Claudianus; (2) the more southern, although hardly lacking in mines and quarries, had above all a role driven by commercial and customs activities. It is crossed by two tracks connecting Koptos on the Nile with the ports of Myos Hormos and Berenike on the Red Sea. Berenike, the more important of these ports, gave its name to the entire region: Mons Berenices or Berenicidis.8 In this toponym, Mons does not refer to a mountainous massif, as in Mons Claudianus: it is the semantic calque of the Egyptian ḏw, translated into Egyptian Greek as oros, which designates the high land of the desert regions flanking the Nile on both sides, no matter how actually mountainous.

1. The zone Claudianus-Porphyrites and the founding of Kaine Mons Claudianus and Porphyrites were served by tracks that converge at Qena, today a large town built on the right (east) bank of the river, nearly opposite the archaeological zone of ancient Tentyra (modern Dendera), the metropolis of the Tentyrite nome (Chapter 14). Qena gets its name from the Greek Kaine (scil. polis), “new city.” Unlike Tentyra, Qena shows no traces of pharaonic or even Ptolemaic antiquities. All of the evidence suggests that this “New City” often mentioned in the ostraca from Mons Claudianus was built in the Roman period – whence its name – to provide a Nile terminus for the roads of Claudianus and Porphyrites. It must have had a suitable harbor for loading the enormous monoliths extracted from the granite and porphyry quarries and for unloading the provisions and perhaps also the pack and draft animals requisitioned for serving the metalla, if these came from a distance, like the camels borrowed from cultivators of Soknopaiou Nesos, in the Fayyum, in 163, because that year a porphyry column had to be transported.9 One may equally imagine at Kaine warehouses to hold provisions and a public bank (δημοσία τράπεζα) from which one could withdraw the salaries for the labor force. The ostraca from Mons Claudianus also allow us to see that the families of the quarrymen and smiths who had come from Alexandria or from Syene lived in Kaine: under Hadrian and Antoninus, these workmen would designate one of them (endowed with the title of kibariates for the occasion) to go down to the valley to collect the salaries of his comrades and draw their rations of free wheat. Often enough, the kibariates passed the artaba of wheat on to the spouses, sisters, or mothers of the workmen for them to turn it into bread.10 The O.Claud. several times mention an ὁδὸς Κλαυδιανή or Κλαυδιανοῦ, the Latin equivalent of which would have been via Claudiana. This very likely refers to the track that connected Mons Claudianus with Kaine. Its twin track serving the Porphyrites must have been the ὁδὸς Πορφυρίτου mentioned in several passes.11 Older maps of the Eastern Desert in the Roman period, notably that of Meredith (1958), show these two roads joining at the Red Sea, at Myos Hormos, which these maps locate at Abu Shaʿr, a Late Antique fortlet, which (it is now recognized) cannot be identified with that Ptolemaic and Roman port.12 This supposed extension of the roads of Claudianus and Porphyrites to the sea is also now 8. Mons Berenices: CIL IX 3083 = ILS 2699 (praef. praesidiorum et montis Beronices); Berenicidis: CIL X 1129 = ILS 2698 (praefecto Bernicidis) and ILS 2700 (praef. montis Berenicidis). 9. BGU III 762 and P.Lond. II 328: two declarations of camels made by two brothers. The village of Soknopaiou Nesos specialized in camel-raising. 10. On this familial connection, authorized and perhaps even organized by the imperial administration for the material and psychological comfort of the skilled Egyptian workforce, see Chapter 14. 11. Cf. O.Claud. I, index. 12. It was James Burton who, in 1822, first identified Abu Shaʿr with Myos Hormos (Cuvigny [ed.] 2006: 25).

236

Rome in Egypt’s Eastern Desert

to be discarded, given the correct identification of Myos Hormos with the modern Qusayr al-Qadim.13 Even if it is likely that Porphyrites communicated with the sea, which is not far away, by an unimproved track (if only to get supplies of fresh fish, of which the soldiers were fond14), the two metalla have to be seen as the other ends of the two roads that converge at Kaine, which were named after them, just as Myos Hormos and Berenike gave their names to the two major desert routes ending at those ports. Unlike the Desert of Berenike, which constituted a well-defined entity, with a name and a commander-in-chief who appears constantly in ostraca of this region (the prefect of the Desert of Berenike), the identity of the northern part of the Eastern Desert, which was dominated by the two great imperial quarries, is more difficult to make out. The official designation of the employees of the imperial familia may offer elements of a solution.15 In the receipts for advances of rations and salaries issued by this category of personnel, which was present at Mons Claudianus under Hadrian and Antoninus Pius, the declarants identify themselves by their membership in a numerus and in an arithmos, the arithmos referring (somewhat paradoxically) to a subdivision of a numerus. Almost all belong to the numerus of Porphyrites and to the arithmos of Claudianus;16 some belong to the same numerus, but to the arithmos of Tiberiane, a small satellite metallon of Claudianus, which we have securely identified with the Roman site of Wadi Barud. The entire region of the Eastern Desert crossed by the roads of Porphyrites and Claudianus could thus take its name from Porphyrites and be called, by metonymy, Porphyrites. This hypothesis is, however, not so far confirmed by any document. The dominant position of Porphyrites may be explained either by the prestige and beauty of the material, or simply by its precedence: The exploitation of Mons Claudianus, if one is to rely on its name (and no documents contradict this indication), did not begin before Claudius, while an inscription found in 1995 at Porphyrites shows that the porphyry quarries had been discovered under Tiberius.17 This preeminence emerges in any event from an inscribed lintel found at Hermou polis, which refers to hosp(itium) tabula(riorum) Porphyr(itae) et aliorum metallorum,18 “residence of the accountants of Porphyrites and the other mines and quarries.” One might rather have expected to find such a hospitium at Kaine or Tentyra. But perhaps “the other metalla” include the marble and alabaster quarries located along the east bank of the Nile between Akoris and Ptolemais.19 It is also possible that there were other hospitia, dispersed across the province, for the tabularii working in the administration of the imperial metalla directed by a procurator metallorum. Two of the beneficiaries of advances of rations belong not to the numerus of Porphyrites, but to that of Alabastron20 (the arithmos to which they were attached does not appear), which could be the name of another administrative zone of quarries, comprising the line of alabaster and marble quarries mentioned in the previous paragraph. The administrative divisions of this imperial work force thus did not prevent occasional transfers from zone to zone without requiring a change of unit designation. 13. Peacock 1993; Bülow-Jacobsen, Cuvigny, and Fournet 2006: 24–27. The identification was fully confirmed by the excavations of the University of Southampton at Qusayr al-Qadim (Peacock and Blue 2006). 14. O.Claud. II 241–242; SB XXII 15452. This track toward the sea thus crosses the via Hadriana, which parallels the coast. It was not impossible, although less direct, to reach the sea from Mons Claudianus. The tracks between the metalla and the sea were taken above all by the “Fish-eating Arabs” who came to sell fish (O.Claud. inv. 529). 15. In O.Claud., familia is the collective designation for the imperial workforce, mainly employed in logistics, in opposition to the indigenous specialists in the extraction and cutting of stone. 16. E.g., O.Claud. III 482: Ἁρποχρατίων Εἰσαροῦτος ἐγ νουμέρου Πορφυρίτου, ἀριθμοῦ Κλαυδιανοῦ. 17. Van Rengen 1995 = AE 1995, 1615. 18. Cockle 1996. 19. The presence of the Roman army in the quarries of Akoris and Ptolemaïs is attested by inscriptions. 20. O.Claud. III, pp. 37 f. The name of this numerus is to be connected with that of the village of Alabastrine (Alabastron polis in the Ptolemaic period) and with Ptol., Geog. 4.5.12, who locates the “mountain of alabaster” immediately to the north of the “mountain of porphyry.”

The road system of the Eastern Desert of Egypt

237

The structure of the administrative hierarchy in the zone Claudianus-Porphyrites is not clear. Some elements of information are provided by the ostraca from Mons Claudianus, by the great Trajanic and Hadrianic inscriptions from Claudianus and Porphyrites, and finally by the ostraca of Umm Balad, a small metallon opened under Domitian in the south side of Porphyrites, where we excavated in 2002 and 2003. The material there dates essentially to the reigns of Domitian and Trajan. The site was originally called Kaine Latomia Domitiane, then only Kaine Latomia (“New Quarry”). The praesidium was placed at the mouth of a deep and narrow valley, in which only two quarries were opened; the material, a dark granodiorite, turned out to be unsuitable for the extraction of monoliths, and work was abandoned under Trajan. A handful of Antonine ostraca suggest a short-lived attempt to resume exploitation of the quarry. A connection links this site to the road of Porphyrites. I know some ostraca of Porphyrites only from the unpublished preliminary reports by W. Van Rengen, who has been entrusted with their study. They hardly seem to add any information on the administration of the metallon and the region. In contrast, the O.Claud. form a body of material stretching over a sufficiently long period for it to be possible to detect changes in this administration. The outer dates attested by these ostraca are year 14 of Nero (68) and the reign of Severus Alexander (222–235). Occupation was not continuous: it depended on imperial orders for stone. The ostracon dated in 68 comes from a dump near the fortlet wrongly called the “Hydreuma”21 by the German team that worked at Mons Claudianus in 1961 and 1964 to distinguish it from the large fortified village of the Wadi Umm Hussayn. The latter was clearly built under Domitian (like that of Umm Balad): the shattered pieces of his dedication, found in the road leading to the gate, bear witness to this;22 on the other hand, no dump or Domitianic material has been found. There were two peaks of activity: (1) at the time of the construction of the forum of Trajan (begun in 107, inaugurated in 112); (2) under Hadrian, at the time of the construction of the Pantheon, then of the temple of Venus and Rome. From these two reigns date the inscriptions that reveal the command structure in the zone: I.Pan 38 is a dedication offered under the reign of Trajan to Zeus Helios Great Sarapis by the Alexandrian architect Apollonios son of Ammonios, “under the procurator (sc. metallorum) Enkolpios and the centurion Q. Accius Optatus.” Also, under Trajan, another centurion, Annius Rufus, detached from a non-Egyptian legion, the XV Apollinaris, dedicated an altar on which he describes himself as praepositus ab optimo imp(eratore) Traiano operi marmorum Monti Claudiano (I.Pan 39). The absence of a territorial equestrian prefect in I.Pan 38 is noteworthy, as he would have to have been named if this function existed: the only individual whose competence goes beyond Mons Claudianus is the freedman procurator Enkolpios. After the great Jewish revolt of 115–117, the metalla of Claudianus and Porphyrites were reopened at the start of Hadrian’s reign. The emperor entered into a business agreement with one of his slaves, Epaphroditos, who dedicated, at Claudianus and Porphyrites respectively, two temples to Zeus Helios Great Sarapis.23 Epaphroditos calls himself μισθωτὴς τῶν μετάλλων24 in both dedications. These are the only attestations of leasing the working of imperial quarries in the Eastern Desert. As administrative authorities, Epaphroditos cites the procurator (metallorum) Ulpius Chresimos and two centurions, one for each metallon (Avitus at Claudianus, Proculeianus at Porphyrites). The start of Hadrian’s reign appears to be poorly represented in the ostraca found at Mons Claudianus, but that is probably just the result of the absence of any precisely dated ostracon from that period. 21. There is in fact no hydreuma, in the sense of “well,” at this location. The fortlet was called that because of the cistern by its side. The meaning of hydreuma in the Eastern Desert was imperfectly understood at the time the name was given (cf. Cuvigny [ed.] 2006: 353 f.). The excavation of the dumps of this fortlet was carried out by J.-M. Carrié. 22. O.Claud. I, pp. 11–13 (AE 2001, 2044). This dedication dates from 85/6. 23. I.Pan 42 (118) and 21 (probably the same date). 24. Translation of conductor metallorum. Cf. O.Claud. III, pp. 15–21.

238

Rome in Egypt’s Eastern Desert

We tend to suppose that the south dump25 was formed under Trajan because the material that came from this dump has yielded several dossiers dated to the reign of that emperor, but two tituli on amphoras with the name of the centurion Avitus26 suggest the presence of Hadrianic texts, which we have no means to identify as such. Some other tituli picti in Avitus’s name come from the dump accumulated around the well. The only ostraca dated by a regnal year of Hadrian come from other dumps and date to the last years of his reign (136–138). Along with the Trajanic corpus, the texts from the reign of Antoninus Pius form the other large body of ostraca found at Mons Claudianus. This period does not correspond to any known imperial order for stone, but the reason for its existence might be the completion of the temple of Venus and Rome.27 The size of this corpus is the result of a new procedure28 for the distribution of the pay of the native skilled workforce: each quarryman or smith drew up each month instructions (entolai) concerning his salary in cash and his rations. These monthly slips make up one of the most common documentary types in the O.Claud. (the database of the entolai includes more than 1200 items). Texts of this type have also been found at Umm Balad and Porphyrites. They generally have no date apart from the month, but they come from the same stratigraphic units as the receipts for advances of rations and salaries issued by the other category of workers, the imperial familia, which in contrast include a mention of the regnal year. A single entole from Mons Claudianus is dated. It comes from the West Sebakh (i.e., the west dump), which was especially rich in documents from the end of Hadrian’s reign, and dates specifically to year 21 of that emperor.29 It therefore seems that when the new system of managing rations and salaries was put into place, under Hadrian, the entolai were customarily dated by regnal year. The only dated entole found at Porphyrites confirms this: it is older and dates from year 13 of Hadrian (128). In contrast, the middle of Hadrian’s reign has not left any visible traces at Mons Claudianus: perhaps because there were no orders for granodiorite at that time, or perhaps because the entolai of this period were discarded in a dump that has not been excavated or that has simply disappeared. Antoninus’s successors are represented in O.Claud., but only sporadically, and ostraca with year dates are very rare: after all, ostraca were not normally meant to be archived or kept for later production as legal evidence; it was thus useless to date them precisely. The ostraca later than Antoninus have often come from trenches inside the praesidium that have also yielded Antonine texts. Sometimes they are even inextricably mixed up with these, when we are dealing with secondary dumps: the garbage was swept up in an undifferentiated mass at the time of a new occupation of the praesidium and thrown into abandoned rooms. It was under Antoninus Pius, it seems, that the function of curator Claudiani first appears. The earliest known of these curators is Iulius Alexandros, known from four letters that he sent to two different centurions. The existence of this function suggests that there was no longer a centurion stationed at Mons Claudianus, at least on a permanent basis: these letters are probably drafts. This hypothesis is supported by the discovery in layers 1 to 5 of trench FW1, room 1, of a lot of ostraca originating in the office of this curator: among them are several letters addressed to subsequent curators (designated as κουράτωρ Κλαυδιανοῦ/μετάλλου Κλαυδιανοῦ/πραισιδίου Κλαυδιανοῦ) by their colleague of the praesidium of Raïma. This praesidium, to which the Romans gave an apparently Semitic name,30 must necessarily be near the metallon at Claudianus, for it is the provenance most often mentioned in the 25. The South Sebakh of the publications. 26. O.Claud. inv. 53 and 4152. 27. Bülow-Jacobsen 1996: 723. 28. Introduced in fact under Hadrian, whose reign is poorly represented in O.Claud. 29. O.Claud. inv. 8496. 30. The toponym Rayma is particularly common in Yemen (Chapter 1, p. 54).

The road system of the Eastern Desert of Egypt

239

letters found there. The most plausible candidate is Abu Zawal.31 The messages on ostraca of the curators of Raïma to their counterparts at Claudianus are almost all cover letters accompanying official correspondence: the hour of departure from Raïma and the name of the messenger are always indicated. Amid this correspondence arriving from the valley, we find two letters “of the centurion” that the curator of Raïma, Ulpius Dios, forwards to the curator of Claudianus, Alexas Amans. One of the cover letters of Ulpius Dios, unfortunately damaged, grammatically connects “the prefect and the centurion,”32 who must therefore both have been in the valley. A draft or archival copy of two letters found in another dump, both written on the same potsherd and dated to 189, allows us to place chronologically one of the curatores Claudiani of the group of trench FW1, Iulius Silvanus, who is mentioned in it, and thus indirectly the entire dossier of curatores Claudiani from this trench. The author of the two letters is the temporary replacement of Iulius Silvanus, the vice-curator (ἀντικουράτωρ) Rufus Aristoteles. They are addressed respectively to Vibius Alexandros, prefect (ἔπαρχος), and to Tertullus, procurator (sc. metallorum).33 The hierarchical distance between the sender and the recipients explains the diplomatic form of these two letters, which are formally memoranda, hypomnemata,34 while the letters written by the curators to a centurion have a properly epistolary form, which was viewed by the ancients as more familiar, even if the name of the addressee is respectfully put at the start before that of the writer.35 The existence of a curator of Claudianus would seem to exclude the presence on site of a centurion, as had certainly been the case under Trajan. All the same, the presence of centurions at Claudianus is attested by inscriptions on jars with their names and by letters addressed to them. But a letter can be only a draft, or a copy kept by the sender; this is what one suspects particularly in the case of four letters addressed to the two different centurions Annius Decmus and Claudius Barbula, by the curator Claudiani Iulius Alexandros, at the end of Hadrian’s reign or the start of that of Antoninus Pius. In contrast, a letter addressed to a centurion from a neighboring praesidium is a reliable witness of the presence of this centurion at Claudianus. But was that presence permanent? Let us look at the case of Claudius Barbula. There is one titulus with his name on an amphora, and he is the recipient of two letters from the curator Claudiani, as well as of a third, of which we have only a draft, written by the hunter (κυνηγέτης) Chrysanthos.36 Unless we are to invent an excessively complicated scenario, this draft should have been written on the spot: Chrysanthos was at Claudianus. The letter begins, “I am sending you word through Bassus, the duplicarius.” This opening already suggests that Chrysanthos and Barbula were not in the same place. Then Chrysanthos comes to the point: “Since you went down (κατέβης) from Porphyrites, my three aides have been taken away, so that I am now alone.” The verb καταβαίνειν is regularly used in the ostraca of the Eastern Desert to express movement toward the valley. Let us suppose that movement from Porphyrites toward Claudianus was similarly expressed by καταβαίνειν (we do not have direct evidence of this). In any event, Barbula was not at Claudianus when Chrysanthos composed his petition. It may therefore have been the case that at this moment (between 136 and 153, according to the dated ostraca found in the layers that provided the attestations of Barbula) there was only one centurion 31. Chapter 1: 26° 40’ 18” N/33° 14’ 26” E. Abu Zawal lies 32 km from the fortified village of Mons Claudianus. Its well supplied not only people and passing animals (cf. infra, n. 92), but also Mons Claudianus. The praesidium of Abu Zawal is built on an old site of extraction and crushing of gold ore (Meredith 1952: 101 f.). This is also the case with Persou and Kompasi in the desert of Berenike. Just like these two sites, where water was perhaps more abundant than elsewhere, Raïma was also a marketgarden center, which supplied Mons Claudianus with vegetables (cf. O.Claud. II 370). 32. ] ἐπάρχῳ καὶ τῷ (ἑκατοντάρχῃ) (O.Claud. inv. 7027). 33. Chapter 5. 34. The prescript of the hypomnema is τῷ δεῖνι, παρὰ τοῦ δεῖνος (“To so and so, from so and so”). 35. Prescript τῷ δεῖνι ὁ δεῖνα χαίρειν (“To so and so, [from] so and so, greetings”). 36. O.Claud. inv. 8574. The draft character of this petition emerges not only from the writing material, but also from the erasures and the numerous abbreviations.

240

Rome in Egypt’s Eastern Desert

for the two metalla, to which this officer made periodic visits. It is perhaps in this light that we must interpret O.Claud. IV 868, a draft of a letter addressed by a quarryman to a beneficiarius of the prefect of Egypt, asking him to forward a petition (λίβελλον) to the prefect; the petition and this cover letter were transmitted to the beneficiarius by a centurion. It is tempting to think that when the centurion came to inspect the metalla, the workers took advantage of his visit to hand him petitions. Under Commodus, a new figure in the administrative hierarchy of the region of imperial quarries makes his appearance: the prefect (ἔπαρχος, without further elaboration). We know of two of these, Vibius Alexandros, who has already been mentioned, in 189, and Antonius Flavianus, who may have been his predecessor. Flavianus receives the same type of collective letters from workers as the procurator metallorum Probus, who was also the recipient of a letter from Pomponius Faustianus, the prefect of Egypt in 185–187.37 These prefects in fact coexisted with the procurator metallorum. Thus at least from the reign of Commodus an officer of equestrian rank was in charge of the region, as had for long (at least since AD 11) been the case for the desert of Berenike. We do not know what the definition of the prefecture of these two officers was: were they only territorial prefects, or did they combine a prefecture of the region with that of a military unit? Both models are known for the prefects of Berenike.38 Vibius Alexandros is known as epistrategos of the Heptanomia39 from P.Lips. II 146, dated between 18 January and 5 April 189 (Chapter 5). The letter of Rufus Aristoteles is dated 1 March 189. Since we do not know any other territorial prefects in Egypt besides those of Berenike, it seems very likely to me that Vibius Alexandros combined the prefecture of a military unit with the epistrategia of the Heptanomia (we do not know if the seat of this epistrategia was at Hermou polis or Antinoou polis). The most similar case known to us would then be that of [- - -] Marcus, prefect of the ala Herculiana and interim epistrategos of the Thebaid.40 If the prefecture of the desert of Berenike or that of the ala stationed at Koptos maintained standing connections with the epistrategia of the Thebaid, it would seem that the upper levels of the administration of the metalla of the zone Porphyrites-Claudianus were based in the Heptanomia. This may explain the presence of a hospitium tabulariorum Porphyr(itae) et aliorum metallorum at Hermou polis (the border between the Heptanomia and the Thebaid was the southern boundary of the Hermopolite nome). It is then paradoxical that the road serving these metalla left the valley at Kaine, fully in the Thebaid. Another possible link between the Porphyrites and the epistrategia of the Heptanomia is the mention, in an unpublished ostracon from Porphyrites, of a dekanos of the Hermopolitans. This ostracon belongs to a dossier concerning the distribution of bread to the workers, dated to the first half of the third century on paleographic grounds;41 in this dossier, the bread is distributed to workers (called ergatai) by their dekanoi, whose usual title is dekanos of the ergatai. The Hermopolitans were thus a body of workers (but of what status, we do not know) who came from Hermou polis or its nome. In this period, the system of kibariatai described earlier apparently no longer existed: the bread dossier of Porphyrites reveals a different organization of work and recruitment (Hermou polis is never mentioned as the source of workers at Mons Claudianus, apart from the exceptional reference to imperial workmen belonging to the numerus of Alabastron). Antonius Flavianus presents a different problem: he reappears in an undated but late ostracon from Dios as the recipient of a letter from the curator of that praesidium.42 His title is lost, but the context 37. P.Bagnall 8. This ostracon is a translation into Greek of the original letter in Latin. 38. Chapter 3, pp. 102 f. 39. The Heptanomia was the epistrategia located between the Thebaid and the Delta. It corresponds roughly to middle Egypt. 40. P.Brooklyn 24 iii.13–14 (213–16). 41. Van Rengen in Peacock and Maxfield 2007: 409. 42. O.Dios inv. 514. Its provenance, a dump inside the fort, implies a date between the end of the second century and the first half of the third.

The road system of the Eastern Desert of Egypt

241

of the document is a small series of letters on ostraca (thus copies of drafts) addressed by curatores praesidiorum to the governor of the desert of Berenike.43 This personage bears various titles: prefect of the desert (of Berenike), prefect of some ala and of the desert, procurator Augusti or even procurator Augusti and prefect of the ala of the Vocontii.44 Two possibilities offer themselves: (1) Antonius Flavianus served in the Heptanomia and in the Thebaid at different times in his career; or (2) he was prefect of the desert of Berenike with authority as well over the imperial metalla located north of his zone; this distinction would be tied to the origins of the prefecture of Berenike. Indeed, the first known prefect of Berenike, P.  Iuventius Rufus, attested from 11 to 18, exercised authority at the same time over the mineral resources of the province, which gave him the supplementary title of μεταλλάρχης/ ἀρχιμεταλλάρχης);45 he was supported in this function by a procurator metallorum, who was not yet at this time an imperial freedman, but his own freedman. The two inscriptions that enumerate the titles of Iuventius Rufus are located in the granite quarry of Samna (called Ophiates) and in that of bekhen stone in Wadi al-Hammamat. These two materials, however, do not appear in the titulature of Iuventius Rufus, who names only the most prestigious – all located in the desert of Berenike: he is “(archi)metallarch of the Emerald, the Topaz, the Pearl, and all the mineral resources of Egypt.”46 It was during his time in office that Porphyrites was discovered, in 18, but the discoverer, a certain Gaius Cominius Leugas, does not mention the archimetallarches in his act of thanksgiving to Pan and Sarapis.47 This discovery must have marked a decisive turn in the exploitation of the geological resources of the Eastern Desert. The titulature of the archimetallarches has an aura still of the Ptolemaic period, when precious stones were considered the main mineral wealth of the region, as appears from the description of the mission entrusted in 130 BC, by the strategos of the Thebaid, to Soterichos son of Ikadion (I.Pan 86): “in charge of the extraction of precious stones, of naval expeditions, and of the provision of security for those who bring back from the Desert of Koptos the cargos of incense and other gifts.”48

2. The main roads of the Desert of Berenike Myos Hormos and Berenike were Ptolemaic foundations. Berenike, created under Ptolemy II Philadelphos for importing war elephants, was the older: before, there had been attempts to bring the elephants in through Heroon polis, at the head of the Gulf of Suez, and by the canal of Nechao, reopened by Philadelphos, but the difficulties of navigation in the Red Sea forced the Ptolemies to find a port of disembarkation as far south as possible. Berenike was connected to the Nile valley by a road to Edfu (Apollonos polis Magna). Its activity halted at the time of the revolt and secession of the Thebaid (207/206–c. 186 BC), under the reign of Ptolemy IV Philopator (222–205), the last Ptolemy to have imported elephants. The first known ancient author to mention Myos Hormos is Agatharchides (second century BC), who describes it as a “large port” (λιμὴν μέγας); in contrast, he does not even mention Berenike,49 but only two geographical entities in the region: the Topaz Island and the Foul Gulf. When Strabo visited 43. O.Krok. 10; 14; O.Did. 38; 40. 44. See the prosopography appended to Chapter 3. 45. I.Pan 51 and I.KoKo. 41. 46. The names of the materials exploited serve here as toponyms, as I have shown in Chapter 1. 47. Van Rengen 1995. 48. And not “exotic products.” In the language of administrative jargon of the period, the imported merchandise is presented as diplomatic presents (cf. A. Łajtar, JJP 29 [1999] 62–65). 49. At least in the version of his text preserved by Photius (250.81) and Diodorus Siculus (3.39.3). A third version, transmitted by Artemidoros, himself known through Strabo, mentions Berenike, described as a city (polis) on the Foul Gulf (16.4.5). Strabo, again, speaks in 17.1.45 of “the city of Berenike, which has no port,” πόλιν Βερενίκην ἀλίμενον.

242

Rome in Egypt’s Eastern Desert

Upper Egypt in 27 or 26 BC, not long after the Roman conquest, Myos Hormos was experiencing an unprecedented level of activity, to which he refers. Mixing what he knew from hearsay and his recollections from reading, and trying to provide a coherent compromise out of this information, the geographer provides a confused idea of the topography of this part of the Eastern Desert: he imagines that Myos Hormos and Berenike are near one another, that Myos Hormos serves as a port for Berenike,50 and that a single road serves the two places, beginning in Koptos. In reality, if (and this is by no means certain) Berenike had already recovered some level of activity at that time, it remained connected to the valley by the Ptolemaic road to Edfu. It was only around 4 BC that the Roman authorities steered to Koptos the commercial traffic that would from this point on pass mainly through Berenike.51 Pliny the Elder describes – with great precision, thanks to the quality of his informants, some negotiatores – the caravan itinerary between Koptos and Berenike, but makes no reference to Myos Hormos.52 The silence of Pliny does not, however, mean that the road of Myos Hormos had been abandoned: the excavations of the University of Southampton at Qusayr al-Qadim have shown that the Roman port remained in use until the middle of the third century.53 The same is true for some of the praesidia along that road.54 Myos Hormos, even if it was harder to reach by sea for the larger vessels, had a number of advantages: its port was well sheltered; it is more or less the point on the Red Sea coast closest to the Nile valley; and it occupies a central position on that coast, along which ran, from the reign of Hadrian on, the newly-built via Hadriana, finished in 137. The praesidia of this new road, at least those relatively close to Myos Hormos on its north and south, must have been supplied from Koptos.55 Wagons and wagon-drivers are mentioned fairly often in the ostraca from the road of Myos Hormos (especially in those from Krokodilo, which date to Trajan’s reign and the beginning of Hadrian’s), while they are totally absent from the ostraca of the praesidia that we have excavated on the road to Berenike (Didymoi, Dios, and Xeron56). Wagons are poorly suited to the often-sandy terrain of the wadis of the Eastern Desert, and this form of transport was used only in case of absolute necessity: this was the situation with the metalla of the Porphyrites-Claudianus zone, where they were used to transport blocks extracted from these quarries. For transporting water, food, and merchandise, donkeys and camels were used. A circular from the prefect of Berenike gives an explanation for the presence of wagons on the Myos Hormos road and sheds some interesting light on the reason for the continued existence of Myos Hormos after the rebirth of Berenike, which have been elucidated by A. Bülow-Jacobsen.57 In this circular, the prefect orders the curatores praesidiorum to prevent the wagon-drivers from letting wood fall at the side of the road (this wood “fallen from the trucks” must have been the object of illicit trade): ξύλα τὰ περὶ τῆς χρήσεως τῶν πλοίων φερόμενα εἰς Μύσ[ορ]μον εἰώθα̣σ̣ιν ο̣ἱ ἁμαξηλάτα̣[ι] ἐν τῆ ὁδῶι 50. Οὐκ ἄπωθεν δὲ τῆς Βερενίκης ἐστὶ Μυὸς Ὅρμος, πόλις ἔχουσα τὸ ναύσταθμον τῶν πλοϊζομένων, “Not far from Berenike [280 km, actually!] is Myos Hormos, a city where the terminal port for navigators is located” (17.1.45). This passage of Strabo is deciphered in Cuvigny (ed.) 2006: 3–10. 51. The date of 4 BC is that of the oldest dated graffito at the Paneion of Wadi Minayh, a rock shelter where commercial travelers stopped before the Koptos–Berenike road was fully equipped with fortified wells under Vespasian. It was F. De Romanis who had the idea of connecting the graffiti of this Paneion with the date of the opening of the new road. He also formulated the hypothesis that a growth in the tonnage of the vessels would have spurred the renaissance of Berenike (De Romanis 1996: 183). 52. Nat. 6.102–103. 53. On the basis of the pottery: Peacock and Blue 2006: 174. 54. Cuvigny (ed.) 2006: 201 f. 55. Only one has been excavated so far, Abu Shaʿr al-Qibli (Sidebotham 1993: 263–67). It yielded only four ostraca (Bagnall and Sheridan 1994: 116–19). Few fortlets have been found on the via Hadriana (Sidebotham, per litt.), whether because it was less well equipped than the dedication of Antinoou polis would lead one to think or because they have disappeared. 56. Respectively Khasm al-Minayh, Abu Qurayya (Iovis in the Antonine Itinerary), and Jirf. 57. Bülow-Jacobsen 2013: 567.

The road system of the Eastern Desert of Egypt

243

κατ̣α̣λεινπάνε[ι]ν̣, “the wagon-drivers are accustomed to abandon on the road the wood transported to Myos Hormos for the boats.”58 The excavations of David Peacock at Qusayr al-Qadim have in fact allowed discovery of the location of the zone of the arsenal.59 The masts for which the inscription known as the “Koptos Tariff ” specifies a toll charge were probably destined for Myos Hormos.60 It is there that the ships of the Red Sea commerce must have been built and later returned, empty no doubt, to be repaired in dry dock. When Pliny the Elder was gathering information from the negotiatores, the new road from Koptos to Berenike still had only a few outfitted stopping-places. In year 9 of Vespasian (76/7), a prefect of Egypt, Iulius Ursus, traveled to Berenike and decided, during the return journey, the locations at which the prefect of Berenike was going to have praesidia built, supplied with wells and cisterns. This scenario is known to us from a dedication found, in more or less identical form, at three of these fortlets (Didymoi, Aphrodites Orous, and Sikayt).61 This is the oldest attestation of the use of the term praesidium in the desert of Berenike. Until then, even those wells where there was a garrison stationed (which was probably the case at all of them) seem to have been called hydreuma: this is the case at Apollonos Hydreuma, where, in Tiberius’s reign, the infantryman Gaius Iulius Longinus drew up a receipt for the delivery of wheat in favor of the camel transportation firm of Nikanor.62 These first-generation garrisons probably looked very different from those of the small quadrangular forts so characteristic of the Roman roads of the Eastern Desert. The constructions in front of the Paneion of Wadi al-Hammamat which were occupied by soldiers, according to the ostraca that have been found there, may give some idea of the irregular architecture of these first garrisons.63 From c. 4 BC to 76/7, the caravans traveling between Koptos and Berenike had to make do with four stopping places provided with wells, between which they depended on water that they carried with them. These stations were, according to the accurate distances given by Pliny, Phoinikon, Kompasi, Apollonos Hydreuma, and Kainon Hydreuma. The distances between these wells are, respectively, 98, 132, and 88 km. The first and last stretches of the road were the shortest: Phoinikon is 32 km from Koptos, Berenike 25 km from Kainon Hydreuma. The other sections must have been covered in two or three days each. The outfitting of the road to Berenike, decided under Vespasian, allowed the caravans to reduce substantially the number of animals carrying waterskins. This economy may not have been the main reason that the Roman authorities strengthened the military presence along the road; that was instead brigandage and smuggling. The Eastern Desert was by no means empty of inhabitants. Strabo congratulates himself on the peaceful disposition of the Beduins; he mentions that they exploited the emerald mines. In his time, then, those had not yet come under the control of the prefect of Berenike. On the other hand, from the reign of Trajan until about the middle of the second century, a number of ostraca mention attacks on travelers and even on praesidia, by Barbaroi. It is easy to imagine that the penetration of the Roman administration into the Eastern Desert upset the balance of the desert populations: they 58. O.Krok. 41.20–23. I owe the reading κατ̣α̣λεινπάνε[ι]ν to John R. Rea (per litt. Oct. 2009). The correct form would be καταλιμπάνειν. 59. Peacock and Blue 2006: 59, 80–81, and 87. 60. OGIS II 674 = I.Portes 67 (AD 90). Mast: line 29. 61. Bagnall, Bülow-Jacobsen, and Cuvigny 2001. Sikayt is a praesidium 7 km northwest of Berenike; it was one of the wells dug surrounding Berenike to supply it with water (Pintozzi 2007). 62. O.Petrie 245 = O.Petr.Mus. 149. 63. The ostraca were published by Kayser 1993. A prosopographical connection with a dated graffito suggests that they date to the reign of Tiberius. The plan of the buildings is reproduced on p. 156 of Kayser’s article (see also Cuvigny [ed.] 2006: 95 and fig. 72). The place was called Persou, as was later the now vanished praesidium which was located at Biʾr Fawakhir.

244

Rome in Egypt’s Eastern Desert

became accustomed to dealing with soldiers and merchants, but at the same time the Roman takeover of the water and the emerald mines must have provoked resentment among them, while the passage of caravans loaded with valuable products through what they considered their territory would have provoked their greed. To the temptation for robbery was added that of smuggling. The ostraca hardly mention this, but it is an inevitable occurrence in this desert region, which the nomads must have known infinitely better than the soldiers did.64 Smuggling also was no doubt prevalent at the level of the Red Sea, and not only in the desert of Berenike. An ostracon from Maximianon (published in Chapter 26) opens unexpected horizons. It is a modest exit pass for cavalrymen carrying an official letter concerning the Kinaidokolpitai, a people which Ptolemy locates on the southwest coast of the Arabian Peninsula and whose name suggests that they were smugglers.65 They must have been eminently capable of creating a nuisance for the Roman administration to have communicated a circular on their subject to the garrisons of the road to Myos Hormos, which were fairly far to their north. It is also possible that this ethnic was used to refer to all of the tribes of the coast of the Arabian Peninsula south of Leuke Kome, the southernmost port of the province of Arabia (the location of this port remains debated66). The recent discovery of the dedication of a Roman fort, dated to the reign of Antoninus Pius, in one of the Farasan islands (the archipelago is located off the south coast of the Arabian peninsula, thus near the Kinaidokolpitai, as Ptolemy situates them) testifies to Rome’s efforts, with the empire not hesitating to project a garrison a bit less than a thousand kilometers beyond the boundaries of the Empire, to protect maritime commerce in the Red Sea.67 In my view, it is in this context that we must understand the creation of the via Hadriana. It was certainly not a caravan road: the ancients always preferred water transport, which was less expensive than land transport. The via Hadriana must have served above all to make it easier to keep the coast under surveillance in order to prevent unwanted or clandestine landings, and to assure the service and supply of the coastal praesidia. Its connection with Antinoou polis remains a riddle; there are no architectural remains on the leg that joins the city to the coast.68 Was it just a matter of prestige? Or had the authorities considered developing it into an alternative route between middle Egypt and the area of the great imperial metalla (cf. the hospitium of the tabularii of Porphyrites at Hermou polis, which is opposite Antinoou polis)? The dedication of the via Hadriana found at Antinoou polis69 is dated one year before the death of Hadrian: it is possible that the road was inaugurated before being completely outfitted and that, after the death of the emperor, it was not thought necessary to construct the stations and dig the wells on the leg between Antinoou polis and the sea, which was not of much use.70 While the praesidia on the road to Myos Hormos were abandoned, as was the harbor, in the first quarter of the third century,71 the army pulled back later from the road to Berenike. Since the first version of this chapter, the discovery in the latest layers at Xeron of an archive dating to the 11th year, probably of Gallienus, led us to push forward the abandonment of the three praesidia that we have fully 64. A Beduin driver told one of our inspectors that he was perfectly capable of crossing the Eastern Desert with his pickup from the Sudanese border to the Mediterranean without encountering anyone. 65. Chapter 26, pp. 413 ff. 66. Cuvigny (ed.) 2006: 28–30. See also the installations proposed on the map on p. 15 of the article of Villeneuve 2007. 67. AE 2007, 1659. 68. S. E. Sidebotham has carried out an exhaustive survey on the via Hadriana. Until the full publication appears, one may consult Sidebotham, Zitterkopf, and Helms 2000. For the Antinoou polis–coast leg, see Sidebotham and Zitterkopf 1997: 228. 69. I.Pan 80: ὁδὸν καινὴν Ἁδριανὴν ἀπὸ Βερενίκης εἰς Ἀντινόου διὰ τόπων ἀσφαλῶν καὶ ὁμαλῶν παρὰ τὴν Ἐρυθρὰν θάλασσαν ὑδρεύμασιν ἀφθόνοις καὶ σταθμοῖς καὶ φρουρίοις διειλημμένην ἔτεμεν. The text gives the impression that construction began at Berenike in order to be completed with great ceremony at Antinoou polis. 70. On the via Hadriana, see also Rathbone 2002. 71. Except for the late praesidium of Qusur al-Banat, perhaps built in 202 (Brun 2018: §20).

The road system of the Eastern Desert of Egypt

245

excavated on this road (Didymoi, Dios, Xeron): no longer 250, but 270, presumably with the Palmyrene invasion of Egypt.72 These three forts show major changes starting in the last third of the second century or perhaps, rather, in the beginning of the third. The management of garbage shows an astonishing relaxation of discipline: instead of carrying it regularly to dumps outside the forts, it was allowed to accumulate in the buildings, where the floor levels rose, or else it was thrown into abandoned rooms. However, the chapels continued to be maintained. Some walls were cobbled up, and small spaces, often with troughs and small silos, proliferated. The proportions between the bones of pigs, which had been in the majority in the previous period, and those of sheep and goats, were reversed.73 J.-P. Brun associates these changes with the presence of eastern, notably Palmyrene, soldiers in the garrisons from Caracalla on.74 The contours of the bodies of ostraca found in them also change; the most remarkable novelty is the reduction in the number of private letters, quantified in the following table: Table 15.1. Proportion of private letters in earlier and later middens Dios Didymoi Xeron

Inventoried Ostraca

Private Letters

Percentage of Private Letters

1,180

474

40.16 %

fort

331

49

14.8 %

dump

592

277

46.79 %

fort

340

37

10.88 %

dump

785

198

25.22 %

fort

423

28

6.6 %

dump

Because of this gap in the supply of letters, the historian loses an important source of information. Although they are never dated and rarely indicate the status of the persons involved, and even though they reflect the narrow interests and limited horizons of the correspondents, the private letters allow many connections to be made when they are numerous enough.75 Most often, they are requests for the dispatch of foodstuffs or objects of daily life, letters accompanying the goods requested, and notices of receipt of the goods. They are highly standardized. The object of the message is usually wrapped in conventional formulas: wishes for the good health at the start, and greetings sent or to be sent at the end (“So and so greets you, greet so and so”). But even this formal structure can be useful for the historian: first, it allows clusters of sociability to be discerned, even if the exact connections between individuals forming them usually escape our understanding. And at the end of Trajan’s reign, the letter-writers of the desert start to begin their letters by telling the recipient that they offer prayers on his behalf before the tutelary divinity of the place from which they are writing;76 this is the epistolary proskynema. This conventional formula (provided that the writer does not simply refer vaguely to “the gods here”) has the advantage of locating the author of the letter, for each praesidium normally had a different genius loci (in the desert of Berenike, however, Athena looked out for both Persou and Xeron, but since these stations were out of 72. Brun 2018: §27. See Chapter 27, p. 416. 73. Leguilloux 2018. 74. Brun 2018: §25. 75. Cf. the network analysis made by Bérangère Redon in O.Krok. II on the basis of three dossiers of letters and the graphic reproduced in Chapter 41, p. 612. 76. For example: τὸ προσκύνημά σου ποιῶ παρὰ τῷ κυρίῳ Διί, “I offer a prayer in your name to our lord Zeus.” Cf. BülowJacobsen, “Toponyms and Proskynemata,” in Cuvigny (ed.) 2006: 51–59. See also Chapter 18, pp. 301–4.

246

Rome in Egypt’s Eastern Desert

range with respect to epistolary communication with one another, this was not a source of confusion for ancient readers or for us). The proskynemata have shown that letters on ostraca were not normally sent farther than the two nearest praesidia: in the absence of a postal service available to private individuals, the sender could not be confident that his letter or parcel would be entrusted to reliable persons. The parcels that generated the most intense exchanges of correspondence (letters of dispatch or of receipt) are kitchen-garden foodstuffs (green vegetables, salad greens, and aromatic herbs). Thanks to the proskynemata, it has been possible to locate the kitchen gardens and to observe that they supplied the immediately neighboring praesidia: greens from Persou were consumed at Maximianon and Krokodilo, those of Kompasi at Didymoi and probably also Aphrodites.

II. The praesidia and their occupants 1. Civilians and soldiers Soldiers detached from military units in Egypt were present in the quarries as well as in the stations along the roads. Except for some legionary centurions, they were always drawn from the auxiliary units. The Roman sites of the Eastern Desert accommodated also non-military personnel. In the quarries, this means above all the workforce, of which two major groups appear distinctly under Trajan (see Chapter 11 for details): (1) the παγανοί (a transcription of the Latin pagani), of which the most appropriate translation is “indigenous” or “native” (i.e., Egyptian) staff: these were the specialists of local origin involved in extracting and cutting the stone, along with the smiths; (2) the familia, which is in charge of logistical tasks: drawing and transportation of water on human backs, and the removal and transportation of blocks. Although this is never directly said, this familia probably belonged to the emperor, but the names of some of its members show that it was not all of slave status.77 Imperial slaves of a status higher than that of the members of the familia78 took charge of managing the provision of food and other supplies: for example, the oikonomos (= dispensator) Geta, from whom the curator of Raïma requested equipment for drawing water. Among the civilians working in the quarries, we must also include the dekanoi,79 who managed the donkeys and camels used for local transportation of water and in hauling the large stone blocks to the valley. We are fortunate enough to have found at Mons Claudianus a register of distributions of water, which offers us a snapshot of the population present at the metallon on a given day in the reign of Trajan.80 The pagani number 421, the employees of the familia 400; there are in addition another 35 civilians (guards, artisans, etc.), whom we cannot assign clearly to either of the two larger groups. The only civilian staffer mentioned is an engineer (architekton). The military personnel number sixty, including a centurion and a decurion, but only six cavalrymen. The praesidia of the desert of Berenike housed also non-soldiers. Since they did not work for the emperor, they were never the object of an administrative census, and their number is thus impossible to determine. Their presence is revealed almost entirely by the private letters exchanged between praesidia. It is, moreover, often impossible to discern if an individual is a soldier or civilian: individuals are not regularly identified in the letters by more than a single name, which is the one that they went by in daily life, and these names in general belong to the same onomastic repertory as those of the civilian and 77. This still obscure question is discussed in O.Claud. III, pp. 25–29. 78. They belong to the category of “domestics and functionaries.” 79. I return to this title below, pp. 257 f. 80. Chapter 11.

The road system of the Eastern Desert of Egypt

247

military personnel, at least for those recruited locally. An Apollonios or a Maximus may equally well be civilian or soldier, whereas the duo nomina or a Thracian or Dacian name (e.g., Eptakinthos, Kaigiza) certainly belongs to a soldier. It is natural to suspect that there were economic and family links between the two communities, but in the absence of documents recording the identity data for individuals, such as we routinely find in the papyrological documentation from the Nile valley, we cannot provide direct support for this hypothesis. The civilians seem to work for the most part as suppliers, but nothing forbids us to suppose that soldiers also engaged in this activity. To whom, for example, did the chickens and pigs raised on the dumps belong?81 This commerce, at least as it appears in the letters, is small scale and never concerns more than small quantities: some baskets of vegetables, a pair of piglets, an amphora of wine … Demand was greater than supply: people often had to ask a comrade in the next praesidium if he could buy in his fortlet what was lacking locally. It is often impossible to know if we are dealing with retail commerce or exchanges between friends. It is again prostitution that allows the ramifications of an organized and vigorous commercial network to emerge: the way in which it was regulated gave rise to exchanges of correspondence between the pimps, their agents, and their clients. The men of a garrison subscribed to hire a prostitute for a month; they agreed in advance with the pimp on the cost of hire, and there were negotiations about costs connected with the transaction, such as transportation and the quintana, the tax charged on prostitution.82 As the exchange of correspondence between praesidia that were not immediate neighbors was unpredictable, the pimps used their contacts at more distant stations to deal with the clientele there. The most verbose of these is a certain Philokles, whose letters have been found at Didymoi and Krokodilo. At the latter fort he had chickens and a pig that he entrusted to his friend Kapparis. One of the most striking traits of the circle of Philokles is that the suppliers and pimps worked as couples; their female partners are constantly associated with them in the letters that they wrote and received, and in the greetings to friends that generally conclude private letters: Kapparis and Didyme, Menandros and Demetrous, Ischyras and Zosime; Philokles, for his part, seems to have formed for some time a ménage à trois with Sknips and the young Hegemonis.83 In the Krokodilo material, the correspondence of the circle of Philokles constitutes 36 per cent of the entire body of private letters. The lack of private letters that we have observed in the later bodies of material found inside the forts might be explained by the absence of a similar network of suppliers: with smaller and less prosperous garrisons, such as we seem to have in the later period of the praesidia, the desert may have become less attractive for merchants.

2. Cavalry and infantry in the Desert of Berenike. underuse of the dromedary In the desert of Berenike, as at Mons Claudianus, the numbers of soldiers are modest. We can estimate them on the basis of two types of documents: the service lists showing the guard duty rotation that have been found at Maximianon, Dios, and Krokodilo, and the postal logs from Krokodilo, particularly O.Krok. 1, which is nearly complete and covers a month and a half. This method rests on two assumptions: (1) that the cavalrymen did not take turns on guard duty and had no other duties than those visible in the postal logs; and (2) that no infantryman in a garrison escaped taking a turn on guard duty. No site has yielded a postal journal that is certainly contemporary to a duty roster, but the essential characteristics seem to remain constant. During the month and a half covered by O.Krok. 1, there were 81. Pigsties have been excavated on the outside dumps at Krokodilo, Didymoi, Dios, and Xeron. 82. Chapter 24. 83. On the circle of Philokles, see Cuvigny (ed.) 2006: 376–88; O.Did. 376–399. On couples of procurers, see Chapter 41, pp. 612 f.

248

Rome in Egypt’s Eastern Desert

only three cavalrymen at Krokodilo to provide postal service and escorts. The horse was not well suited to desert conditions, so a scarcity of cavalry is in no way surprising. Several letters by cavalrymen found in the Desert of Berenike, moreover, testify to the fragility of these animals and the worries that they caused their owners. It is astonishing that the Roman army did not think of making greater use of riding dromedaries, which would have been ideal for the postal service. Their presence is indeed attested in our textual corpus, but it remains exceptional, as was, probably, their type of business. This breed is called ἡ δρομάς (feminine), in contrast to ὁ κάμηλος, the pack or haulage camel. We know that there was – probably under Trajan – a dromas attached to Porphyrites and one to Claudianus.84 In contrast, no garrison of the desert of Berenike seems to have owned a dromas; the only attestation in this region is to be found in O.Krok. 47.37–40, a copy of a circular from a decurion of the ala of the Vocontii, ordering the curators of the road from Myos Hormos to Persou to have a dromas sent to the prefect of Berenike (who was, one gathers, on inspection tour at Persou), escorted by cavalrymen. The general impression is that the use of riding camels was still at this period experimental (in fact, the term dromadarius, to designate someone who rides a dromas, does not appear until the reign of Trajan). It is, moreover, striking that the ostraca more readily refer to the mount than to the person who rides it. And two further indications suggest that the dromadarii of the Eastern Desert were not soldiers: their names (Egyptian, Arabic), and the dromadarius who appears in the great organization chart of Mons Claudianus receives a water ration lower than that of the soldiers.85 The infantry (called στρατιῶται, as against ἱππεῖς for the cavalry, in the ostraca) varied in number according to the duty roster: the minimum number was eight, the maximum twenty. The ostraca produced in the praesidia assign them no activity other than their turns on guard duty. The garrison was made up of soldiers drawn from various units in Upper Egypt. We do not know how many months they remained on detachment (although we do know that a Dacian cavalryman stayed at least five months at Krokodilo86). They were commanded by a curator praesidii who reported directly to the prefect of Berenike, as numerous drafts and copies of correspondence between them show. The prefect of Berenike often combined this prefecture with a military command, generally the prefecture of the ala currently stationed at Koptos. In some periods, he bore at the same time the title procurator Augusti (Chapter 6).

3. The internal organization A. Wells and cisterns The praesidia along the roads were almost always organized around a well; indeed, these were essentially fortified wells. In the quarries, in contrast, the wells were not inside the forts: at Porphyrites, there are two wells not far from the fort, but, as we know at least for Claudianus, the local well was insufficient for the needs of a population that, toward 110, reached the level of a thousand. The metallon also derived its water from several wells, not all of which we can locate, but the names of which we know from the ostraca. There is no well near the fort of Umm Balad: it got its water from caravans of camels carrying water skins, who filled the cistern with water brought from several nearby hydreumata that we have not been able to identify on the ground. 84. O.KaLa. inv. 819, commented in Chapter 11, p. 204. 85. Chapter 11, p. 204. 86. AE 1996, 1647.

The road system of the Eastern Desert of Egypt

249

The wells, or at least those which were integrated with a praesidium, were associated with one or several cisterns. Drawing water must indeed have been a rather slow process, particularly because one had to allow time for the water table in the well to return to its previous level; it was therefore useful to keep cisterns filled in advance to guard against the unexpected arrival of a caravan. The logistical importance of the cisterns (lacci, λάκκοι) is underlined by the famous inscription ILS 2483 (I.Portes 56), which commemorates the construction of a camp (at Koptos, in my opinion) and of lacci at Myos Hormos, Berenike, Kompasi, and Apollonos Hydreuma, by a vexillation made up of about 1,400 soldiers drawn from all of the units present in Egypt at the time.87 We have only a limited sense of the amount of water available to be drawn from the wells in the praesidia of the Eastern Desert. Digging out the wells, which are generally sanded up and collapsed, might give us some information, but it would require technical means that have not been available to us. Several praesidia of the desert of Berenike present, on axis with the gate, a building standing on the edge of the crater that marks the original location of the well:88 excavating these buildings has never brought to light the slightest indication of their function: an administrative structure, or a technical building housing a water-lifting mechanism? At Claudianus, at Porphyrites, and in the desert of Berenike, the ceramicists have identified some fragments of scooping cups from a water-lifting machine, and these are also mentioned in a small number of ostraca with the term kadoi.89 It is, however, difficult to imagine a water wheel in the cramped spaces of the praesidia. It could not in any event have been a saqiya turned by an animal, the mechanism of which would have taken up too much space (two cog wheels, of which the horizontal would have been surrounded by a circular track). It is easier to imagine a vertical wheel with steps, with an axle for a chain of pots, suspended perpendicular to the well. Propulsion could have been supplied by a man treading the steps or a saqiya mechanism operated by human force.90 O.Did. 415 is a vehement letter, the author of which (his name is lost) writes “you imagine that I am still drawing water” (ἐν τῇ ἀντλείᾳ). This expression may be connected with a warning dream reported by Artemidoros of Daldis: a man who dreamed that he was walking without moving forward was condemned to antlia. Artemidoros explains that this is the fate of those who draw water (τοῖς ἀντλοῦσι), to cover long distances while remaining fixed in place.91 A letter of the curator of Raïma, a praesidium near Mons Claudianus that was provided with a well, includes a request for ropes “for the shadufs.”92 This is the only precise technical term in our ostraca that specifies a lifting machine (we also find once 87. CIL III 6627 = ILS 2483 = I.Portes 56. 88. Maximianon, Al-Hamraʾ, Krokodilo, Dios, and Xeron. 89. Tomber in Maxfield and Peacock 2006: 95 f., Peacock and Maxfield 2001: 245–47,  and Peacock and Maxfield 2007: 191–93. R. Tomber observes that scoops are attested in small numbers at Mons Claudianus beginning with Trajan, but that they are particularly characteristic of the Antonine and Severan periods. Pottery scoops in the praesidia of the desert of Berenike have been found in stratigraphic contexts of the third century. Scoops made of metal were very likely used in more prosperous periods; see also the next note. 90. These are the two alternatives for moving a chain of pots (Oleson 1984: 350). The second possibility is illustrated by a drawing of Claude Perrault in his edition of Vitruvius, conveniently reproduced in A. Bouet (ed.), Aquam in altum exprimere. Les machines élévatrices d’eau dans l’antiquité (Bordeaux 2005) 13. It should be noted that the scoops on a chain, exposed to frequent shocks, were necessarily of bronze. 91. Artemidoros of Daldis 1.48; Oleson 1984: 30 f. If, at least in Artemidoros, antlia is associated with a certain drive mechanism, we do not know of what machine he is thinking. There are iconographic testimonies to the Archimedean screw driven by an individual walking above (reproduced in Oleson 1984, figs. 71, 86, and 101), but the use of this procedure for a chain of pots is only a deduction. 92. Σχυνία εἰς το[ὺς κ]ήλωνας (O.Claud. inv. 2238). Leo Tregenza thought in fact that he recognized at Abu Zawal (Raïma in my view) some built remains connected to the presence of a shaduf (Meredith and Tregenza 1949: 21). See the reconstruction of a way station equipped with a shaduf (Talʿat al-Zarqaʾ, on the via Claudiana) in Sidebotham et al. 2008: 320, fig. 13.8.

250

Rome in Egypt’s Eastern Desert

organon,93 which, like mechane, can be applied to any type of lifting machine). It is thus possible that there were several types of machines, varying according to local conditions and different periods.94 B. Other installations inside and around the praesidia I shall deal here only with the praesidia of the roads of Myos Hormos and Berenike, leaving to one side the fortified villages of the metalla, which are complex structures with their own problematics, as well as the praesidia along the roads of Claudianus and Porphyrites. Only one, Badiya, has been the object of excavation;95 as in all of the stations of these roads,96 to the praesidium is added a large enclosure for beasts of burden. Otherwise, the excavation produced few ostraca, and these unfortunately provide no information on the functioning of the station or on the reason for the existence of these enclosures.97 The major buildings inside the praesidia of the Desert of Berenike are the praetorium, the granary, the baths, and the chapel. A series of duty rosters found at Maximianon identifies the four corners of the fort, called coxae, where the soldiers took turns standing guard, by the building that stood there: these were the corner of the horreum, that of the praetorium, that of the small canal (potamion), and finally that of the bath. As the bath has been excavated at Maximianon, each of these corners could be identified. That of the small canal indeed offers a channel that crosses the wall at a certain height, the purpose of which we do not know.98 The praetorium was probably the apartment of the curator praesidii. Duty rosters of the same type found at Dios also mention an ἄνγλον πρετωρίου (= angulum praetorii), which we have probably excavated, for the northeast corner of this praesidium included two rooms with mosaic floors, a modest luxury found otherwise only in the chapel. In the duty rosters of Dios, which are later than those of Maximianon, the corner of the praetorium is the only one where a permanent guard was posted. In contrast, the presence of a guard in the skopelos is always anticipated in the duty rosters from all of the sites of the Eastern Desert, including at Mons Claudianus.99 Skopelos is a common word in Upper Egypt under the principate: the inhabitants of Syene, Edfu, and the Theban region paid in the second century a tax to pay guards stationed in the skopeloi. But the ostraca of the Eastern Desert also mention skopeloi, both in the metalla and in the roads of the Desert of Berenike. It is not clear if the skopeloi were built structures or natural observation points on the mountains: the Greek word allows both senses. Is the skopelos, associated in the duty rosters with the corners and often also with the gate of the fort, an architectural element of the praesidium? It is hard to see what that would be: the praesidia do not have any tower higher than the others, which are almost always found on the four corners of the walls. The presence of an impressive cover of potsherds on a hill 400 m south of the praesidium of Xeron suggests that it is there that we should locate the skopelos: the position there is ideal for watching the circulation in the surrounding valley.

93. O.Claud. inv. 2302. 94. This is what Sidebotham, Hense, and Nouwens 2008: 320–23 think (with proposed graphic reconstructions). 95. Peacock and Maxfield 2007: 25–81. Since then, one campaign of the MAFDO could take place in Dayr al-Atrash in January 2020, the excavators being Joachim Le Bomin and Julie Marchand. 96. Including those that are not, properly speaking, praesidia, but more modest stations (Bab al-Mukhaniq, Talʿat al-Zarqaʾ): see the map at the beginning of this volume. For a good photo of Bab al-Mukhaniq: Meredith 1952: pl. XVI, 2. 97. They are briefly evoked in Peacock and Maxfield 2007: 411. 98. Cuvigny (ed.) 2006: 104 and fig. 88. 99. O.Claud. II 304 (cf. Cuvigny [ed.] 2006: 218).

The road system of the Eastern Desert of Egypt

251

We excavated bathing establishments in four praesidia of the Desert of Berenike (Maximianon, Didymoi, Dios, and Xeron):100 these fortlets seem in general to have had baths. The ostraca say almost nothing about their functioning and use. We know in any case that they were not reserved for the soldiers: O.Krok. II 293 witnesses that two women had been together in the bath at Krokodilo.101 The praesidia were also provided with sanctuaries, the location of which varied. The chapel of the fort is mentioned only four times in the Desert of Berenike: that is the chapel of Aphrodites, for which palms are ordered to decorate it.102 It is pompously called τὰ πρινκίπια τῶν κυρίων (= principia dominorum), which suggests that in it a cult was offered to the emperors. That scarcely emerges, however, from the excavation of the two best-preserved chapels that we have uncovered, at Dios and Didymoi: the inscribed and iconographic finds in both cases indicate instead a Serapeum. In fact, from Trajan’s time on, the syncretic god Zeus Helios Great Sarapis, who seems to have been created in Alexandria under this emperor, dominates the sanctuaries of the Roman installations in the Eastern Desert. At Dios, he had the Fortune (Tyche) of the praesidium sharing the chapel. That chapel has produced a body of inscribed material that hardly suggests a temple devoted to military cults. It was open to passers-by, like the Alexandrian shipper who inscribed a modest dedication there in the third century, and it was even, for some time, the seat of an oracle where travelers could calm their worries. The fill of the statue pedestal in fact contained the membra disiecta of a book of oracular responses on ostraca and tablet, which their reassuring style and tenor link to the knucklebone oracles of Asia Minor.103 Most of the barracks were living spaces. Those in which the soldiers lived were called contubernium, sometimes perhaps also cella, a term certainly used also for the living space of non-soldiers. At Mons Claudianus, under Trajan, the centurion’s lodging was called ξενίαι (curiously in the plural).104 The barracks excavated in the praesidia are small and seem intended for a single occupant, a fact confirmed by the references to living conditions that can be gleaned from the letters.105 The most intelligible, because the fortlet was not much occupied and underwent no renovations, are those of Dawwi, where we observe two rows of barracks. These are rooms of about 3.6 by 3.4 m, with a brick bench to hold bedding.106 Several show troughs against the wall on the outside, probably for a horse fed there. We do not know how civilians and soldiers were distributed inside the praesidia: did they have distinct zones? Was there a room specially intended for prostitution, or did the prostitute of the month go to men’s rooms?

III. Circulation on the roads In the northern network, the roads served the wagon traffic of materials extracted from the quarries and supply for the metalla and intermediate stops. In the Desert of Berenike, the two major cross-desert roads existed for the transshipment of the products of the Red Sea commerce between the Nile and the Red Sea, with Myos Hormos also specializing in shipbuilding. So far, most of the written documentation from the Desert of Berenike has come from the praesidia along these roads, while the Red Sea commerce has left hardly any written or even archaeological traces in these praesidia, apart from an 100. Reddé 2009. 101. Chapter 41, p. 613. 102. O.Did. 31. 103. Chapter 31. 104. The Latin equivalent of ξενία is hospitium (cf. Chapter 11, p. 193). 105. The letter O.Did. 325, sent from Aphrodites Orous, however, indicates that two soldiers shared the same room: Sertorius informs his correspondent in this letter that since he “went up” (sc. into the desert), he has been living with Longinus (μετὰ Λονγίνου συνσκηνῶ). 106. Cuvigny (ed.) 2006: 134 f. and figs. 168 and 171.

252

Rome in Egypt’s Eastern Desert

Indian coin found at Dios,107 the inscription malabathron on a pot found at Maximianon,108 occasional mentions of merchants (emporoi) but without any useful context, and the dedication made by an Alexandrian shipper in the aedes of Dios. Paradoxically, the only documents showing control of circulation by the garrisons watching the roads are the passes given to insignificant groups, the reason for whose movements and presence in the desert, whether administrative or professional, is completely unknown.

1. The control of circulation: passes and safe-conducts A. Passes from Mons Claudianus and the question of the stationarii Brief safe-conducts written in Greek have been found at Mons Claudianus, Didymoi, Xeron, Berenike, and Myos Hormos,109 generally characterized by the use of the imperative πάρες/πάρετε (“let pass”). The passes from Mons Claudianus110 are chirographs,111 the issuer of which is almost always a centurion (once a decurion), while the recipients are usually designated as “the stationarii of such-and-such a road,” generally that of Claudianus or that of Porphyrites, cf., e.g., O.Claud. I 68 (Trajan): Ἀντωνεῖνος  στατιωναρίοις ὁδοῦ Κλαυδιανοῦ χα(ίρειν). πάρετε ἄνδρας δύο. Χοιακ ι̅ε̅, “Antoninus, centurion, to the stationarii of the road of Claudianus, greetings. Let two men pass. Choiak 15.” The recipients are rarely identified by name, but only by number, sex, age class (men, women, children), and on occasion by the number of donkeys accompanying them. They are always very small groups. As these documents were found at Mons Claudianus, we might readily suppose that the centurion was stationed at Kaine and would deliver these passes to the small groups who requested them, paying duties as necessary. These ostraca would then correspond rather well to the pittakia mentioned in the Koptos Tariff, which would have been given to travelers at Koptos against a modest payment.112 Nevertheless, four of the eight centurions who issued them are known to have commanded at Mons Claudianus. How are we then to interpret the presence of these passes at the metallon? Are they passes that were never used, or drafts intended to be replaced by copies with a more official aspect? Or did the centurion in charge of Mons Claudianus not reside there permanently?113 And how, similarly, are we to explain the presence in this dossier of passes for travel on the road of Porphyrites? Supposing that the link that necessarily connected Claudianus with Porphyrites was also called the “road of Porphyrites” would hardly help.114 A different explanation is proposed below. The term stationarii also deserves comment. In the documentation of the Eastern Desert, occurrences of statio and its derivative stationarius are extremely rare.115 Statio in fact appears only in two ostraca116 and in the enormous graffito that the Dacian cavalryman Dida left on the rock that dominates 107. Inv. 645. S. Bhandare, H. Cuvigny, and T. Faucher, “An Indian coin in the Eastern Desert of Egypt,” in Red Sea Conference IX, Networked Spaces: The Spatiality of networks in the Red Sea and Western Indian Ocean (forthcoming). 108. O.Max. inv. 1370. 109. Van Rengen 2001. This pass, until now the only ostracon published from these excavations at Myos Hormos, is issued by a duplicarius for two women. 110. O.Claud. I 48–82 (edited by Wilfried Van Rengen). 111. Papyrologists use the term chirograph for contracts, receipts, memos, and administrative notes that use an epistolary prescript (“So and so to so and so, greetings”). 112. OGIS II 674 = I.Portes 67.23–32. 113. I owe this hypothesis to J.-P. Brun. 114. We do not know what route was used in antiquity to go from one metallon to the other. 115. Στατιωνάριος: apart from the Claudian passes, O.Claud. IV 708.14 and 715.10 (Trajan); O.Dios inv. 1519 (third century); O.Amst. 12.2 (second century) (below, p. 254). 116. O.Krok. I 51.37 (AD 109), O.Claud. inv. 7439 (mid second century). No useful context in either.

The road system of the Eastern Desert of Egypt

253

the praesidium of Krokodilo: armatus feci stationi meses quinque, “I passed five months under arms at the station.”117 But if it is true that statio can mean “place where a soldier is detached,” in contrast to the permanent camp of his unit, it might be more accurate to translate stationi with “on detachment.” Could a soldier on detachment in the Eastern Desert then be called stationarius, which in that case would mean “soldier on detachment”? This question was in fact already posed by W. Van Rengen. But actually, some of the passes suggest that stationarius does not refer in these ostraca to just any soldier detached from his unit. In fact, in three cases of the thirty-five the titles curatores praesidiorum and epiteretai replace stationarius: – O.Claud. I 48.1–7 (c. 110–114): Κουίντος Ἄκκιος Ὀπτᾶτος (ἑκατοντάρχης) κουράτορσι δ̅ πραισι⟨δί⟩ων ὁδοῦ Κλαυδιανῆς χαίρειν. πάρες Ἀσκληπιάδην, “Quintus Accius Optatus, centurion, to the curators of the 4 praesidia of the via Claudiana, greetings. Let pass Asklepiades.” – O.Claud. I 76.1–2 (Trajan): Ἰούλιος Κέλσος (ἑκατοντάρχης) ἐπιτηρηταῖς ὁδοῦ Πορφυρ̣ε̣ί̣το(υ), “Iulius Celsus, centurion, to the epiteretai of the road of Porphyrites.” This title, derived from the verb ἐπιτηρεῖν (“watch over,” “observe carefully,” “watch for”) appears only here with this meaning in Egypt, where it otherwise refers only to an administrator (“supervisor”), whether of a large estate (ousia), or, most often, of an office or a tax that was generally farmed, but for which no bidder had come forward; in the latter case, a liturgist was named who was responsible for the sum due to the treasury at the risk of his own fortune and who assumed the title of epiteretes. Here, however, the epiteretai watch over the circulation of individuals on the desert roads. Since the only two passes addressed to epiteretai are those issued by Iulius Celsus, I suppose that this manner of expressing himself is an idiosyncrasy of this centurion. It remains odd that all of the centurions who issued passes, except Accius Optatus (from whom we have a single pass), so consistently use the word stationarius, which is nearly a unique usage for the Eastern Desert, rather than the word curator, which was the ordinary title of the commanders of praesidia. The only hypothesis that occurs to me to explain this anomaly is that under Trajan, at the period of the passes, the praesidia did not yet exist, and that the posts along the roads of the region of PorphyritesClaudianus had not yet taken the form of the fortlets to which the Roman administration gave the name praesidia, but were very small installations for which the written documentation of the desert has not given us the technical term. In this case, the pass of the centurion Accius Optatus would be later than the others and would date to a time when the via Claudiana had just been outfitted with four praesidia: in fact, there was a lot of building of praesidia, wells, and sanctuaries in the Eastern Desert during Trajan’s reign. The number of four curatores118 also invites us to look into the route of the via Claudiana. Meredith’s map shows it passing by Abu Zawal, Talʿat al-Zarqaʾ, Al-Qurayya, and Al-ʿAras, but that comes to only three praesidia, Talʿat al-Zarqaʾ being only a well with animal lines, with no facilities for lodging.119 117. AE 1996, 1647. 118. If this is indeed how the number δ̅, whose place is unclear, is to be interpreted. 119. It can readily be seen on Google Earth (26° 35’ 09” N/33° 11’ 56” E). Photo in Sidebotham, Hense, and Nouwens 2008: 320, where Talʿat al-Zarqaʾ is called Wadi Abu Shuwehat (the name adopted in the Barrington Atlas of the Greek and Roman World). I wonder if Talʿat al-Zarqaʾ is the well called Akantha in O.Claud. II 362, a letter in which the curator of Raïma (Abu Zawal in my opinion) defends himself to the centurion stationed at Claudianus for having sent him fewer camels than planned: he has been obliged to send some imperial camels to get water from Akantha because there was not enough for the needs of a cart (i.e., of haulage) that had arrived unexpectedly. Abu Zawal and Talʿat al-Zarqaʾ are 11.5 km from one another if one crosses the mountain (cf. n. 126).

254

Rome in Egypt’s Eastern Desert

On the other hand, there are indeed four praesidia on the route which, from Abu Zawal, goes at an angle toward Al-Saqiya. Travelers of the beginning of the twentieth century observed the tracks,120 some of which are still present,121 which would have been left by Roman carts on the two routes. It has been supposed that the carts took either of the roads, depending on whether they were descending fully loaded or ascending empty.122 This detour provides a possible explanation for the issuance by the centurions of Claudianus of passes addressed to the stationarii of the road of Porphyrites. Another later ostracon appears to anticipate only two praesidia between Claudianus and the valley: O.Claud. II 358 (reign of Antoninus), a letter addressed by Nemonianus (centurion or curator Claudiani ?) to Antonius and Furius, curatores praesidiorum, in which he asks them to supply at each stage a tabellarius to accompany a certain Eutyches, who was going to Kaine. One may imagine that instead of going via Al-Saqiya, Eutyches, traveling light, was in a position to take the direct road from Abu Zawal to Al-Qurayya, and that once he arrived at Al-ʿAras, the last praesidium before Kaine, he no longer needed a guide. Still, even if we admit that the stationarii were not curatores praesidiorum, but were responsible for tiny posts, we have no means of knowing if this word refers to the heads of these small posts or to the entirety of the soldiers who occupied them. The only other published attestation of a stationarius in Egypt under the principate is an ostracon in Amsterdam (O.Amst. 12), which also probably comes from a desert praesidium that it has never been possible to localize (Chapter 13, pp. 223 f.). It is a list of stationarii, which suggests that these are simply detached soldiers. B. The passes found at Didymoi These are five in number.123 Three of them are similar to the type observed at Mons Claudianus, but the issuers (a decurion and a centurion) omit the names of the addressees. The pass issued by the decurion is in Latin and uses the Latin imperative equivalent to πάρετε: transmittite. O.Did. 48 and 49, in contrast, take the form of a request for passage addressed to a centurion by one of the members of the group of travelers, who identifies himself once as a donkey-driver. The request is validated by the centurion’s subscription; in O.Did. 49, this is a legionary centurion, who subscribes in Latin (transmittite supra scriptos). Since the request is formulated as for a trip to Koptos, this centurion must have been stationed at Berenike. C. The passes found at Berenike Unlike the passes described in the previous section, those from Berenike were intended to be presented at the end of the road, at the customs gate of Berenike. The status of the issuers is never indicated; they are not in any case military officers. As for the addressees, their position is sometimes mentioned: they were κουιντανήσιοι, a Greek word formed on the Latin quintanenses. This term is derived from quintana, the name of a farmed tax that seems to be due on commerce in a military milieu (including, as 120. Peacock and Maxfield 1997: 261 f. 121. See the photo in Peacock and Maxfield 1997, fig. 7.13. 122. Meredith 1952: 101 f. According to him, carts would have descended from Claudianus via Abu Zawal, Talʿat al-Zarqaʾ, and Al-Qurayya, but would have climbed via Al-Saqiya and Abu Zawal. D. Peacock, who regards use of the track between AlSaqiya and Abu Zawal as a highly probable hypothesis, is above all interested in the existence of two alternative routes linking Abu Zawal and Talʿat al-Zarqaʾ, one going around the mountain on the west, the other going through a pass that is too steep for loaded carts (Peacock and Maxfield 1997: 265–66). 123. O.Did. 47–51.

The road system of the Eastern Desert of Egypt

255

we have seen, the hire of prostitutes). When the Berenike passes present the imperative πάρες/πάρετε, the verb has as its object not persons but goods transported, which have been cleared through customs and can circulate freely. Cf., for example, O.Ber. I 26: Σωσίβ(ιος) Ἀνδούρῳ χα(ίρειν). πάρες Πακοίβι Κλίτου ἐλαίου ἡμικ(άδια) δ̅ καὶ οἴνου ἰτα(λικὰ) δ, “Sosibios to Andouros, greetings. Let pass for Pakoibis son of Kleitos 4 hemikadia of oil and 4 Italian amphoras of wine.” Andouros is known from other texts as a quintanensis, i.e., an agent of the quintana. The ostracon was probably drawn up at Koptos by Sosibios, who worked for this tax-farming contract. The ostraca of Krokodilo allude to other written authorizations that persons traveling on the roads of the Desert of Berenike may have had. The postal journal O.Krok. 1 (108 or earlier) gives us in lines 26–28 the summary of a diploma from the receiver124 Avitus, instructing the curators of the praesidia to obtain escorts from fort to fort for a certain Modestus. Unlike with the other documents registered in this journal, the mention of this diploma is not accompanied by an identification of a cavalrymancourier: it was therefore presented by Modestus himself, which is confirmed by the information that the same day the cavalryman Eial provided the escort (l. 30). This diploma giving rise to the right to a military escort recalls, even if at a more modest level, the well-known postal passports, which were also called diplomata. The book of incoming correspondence O.Krok. 87 (118) contains the copy of a letter from Arruntius Agrippinus, the prefect of Berenike or at least a high-ranking individual in this prefecture. In it, he orders the curators to furnish reliable escorts to “those who are transporting supplies from Koptos under my subscription” (l. 99), so as to give warning about possible attacks by nomads. It is obvious that the transporters were provided with a document that gives them at least the right to military protection; coming from the prefecture of Berenike, it is validated, like two of the passes found at Didymoi, by means of a subscriptio.

2. The official post The ostraca of the Desert of Berenike shed little light on the reason for the existence of the praesidia. Why did the Romans decide to construct a dense network of stations in the Desert of Berenike, when the supplying of these distant and isolated posts represented a non-trivial source of pressure on the military units of Upper Egypt, and above all a heavy logistical burden to manage? Had not the negotiatores managed for almost a century with spaced-out wells, thanks to the well-known ability of camels to do without water for long periods? The letter of Arruntius Agrippinus shows that garrisons could be called upon to protect travelers against the attacks of Beduins: but the road to Berenike had been outfitted under Vespasian. Had the nomads already become aggressive? That is possible. In any case, the first evidence of attacks by “Barbarians,” as they are called, dates to the reign of Trajan (O.Krok. 6 and 47). The area for which the ostraca best illustrate the function of the network of praesidia is not so much the struggle against brigandage and smuggling, or of supplying water to caravans, as the transport of the official mail, for which the curatores praesidiorum filled the function of postmaster. We owe this density of information to the fact that, for reasons not known to us, under Trajan and Hadrian, some curators of Krokodilo copied some postal journals, of which the official copies must have been written on papyrus, onto amphoras. For the most part we have only more or less substantial fragments of these journals, but one of the amphoras could be reconstructed, providing us for forty-eight days a listing of the traffic in documents and official parcels (O.Krok. 1, already mentioned). These objects were identified by 124. See Ast and Bagnall 2015.

256

Rome in Egypt’s Eastern Desert

means of a succinct description (letter, diploma, parrot-fish…125); when the documents came from an important personage, as was often the case, he is named and identified (“letters from Cosconius, prefect of the Desert”). Some are circulars addressed to the curators; in this case, their contents are summarized. Some of these documents are safe-conducts that had been presented to the curator by their beneficiaries (this is the case for the diploma of Modestus, discussed earlier); the others were brought by a cavalryman from one of the two neighboring garrisons (in this case, Persou and Phoinikon126), whose name is indicated along with his hour of arrival. There then follows the name of the next destination and that of the cavalryman of the garrison of Krokodilo who carried the mail. Three cavalrymen were attached to Krokodilo: the Dacian Kaigiza, the Libyan Eial, and a certain Aestivus. Each received a permanent number (1, 2, 3), inscribed as a marginal note next to each entry, which allowed the fair distribution of the burden of service among them.127 The cavalrymen not only served as orderlies: we have seen that they were sometimes called on to escort travelers to the next station, also a registered service. The postal logs made it possible to check on the courier and to identify who was responsible for any delays (a professional failing called katoche, “holding back,” in O.Did. 28). These control systems did not, all the same, deter a certain Herakles, a cavalryman from Dios, from holding back letters of the prefect of Egypt for nine hours before getting on the road, because he was enjoying female company.128 The postal logs comparable to those that I published in O.Krok. were not the only documents that the curators produced in their role as postmasters. They also copied in full the circulars that they were asked to read before forwarding them: O.Krok. 87 is the best-preserved example of these libri litterarum allatarum. These circulars, always in letter form, are sometimes memos,129 sometimes safe-conducts,130 sometimes cover letters forwarding important mail to the curators. The technical term for such dispatch notes is δίπλωμα τῆς ἐπίθεσεως, “diptych of consignment.”131 The two most interesting examples are O.Krok. 47.47–58 (109), which summarizes the duties of a postmaster (“after noting the hours of receipt, from whom you receive these letters, and to whom you are sending them on, you will dispatch them again as quickly as possible …”), and O.Did. 28 (third century). This sherd contains the end of a cover letter, the author of which (in a lacuna) specifies that the mail has been entrusted, on the stage KoptosPhoinikon, to the monomachos Chrysoplokamos. Under the copy of this cover letter, the curator of Didymoi has added: “I have forwarded the letters through Indos, monomachos, at the 9th hour of the night.” Not all the messengers of the official post were regular soldiers. In the Desert of Berenike, we see that after c. 170, official mail was regularly entrusted to monomachai (or -machoi). This term, which in the Roman period normally means “gladiator,” is, in this usage, distinctive to the Eastern Desert, including Mons Claudianus. It seems here to recover its etymological meaning of “one who fights in single combat.” Presumably, the monomachai were armed and trained to defend themselves.132 Their names point to a servile status, but also, with their glittering tones, to gladiators’ stage names: Chrysoplokamos, Gigas, Eukylistros… Could they have been discharged gladiators (who are not necessarily slaves)? These 125. The transportation of fish from the Red Sea, which is logged only for some days, is probably to be connected with the presence of the prefect of Egypt at Koptos or Dios polis (Thebes). 126. The praesidia of Qusur al-Banat and Biʾr al-Hammamat did not yet exist at this period. 127. The notation of the numbers of turns for the service of horsemen is specific to O.Krok. I 1. An entry of this journal is quoted and commented on in Chapter 18, p. 309. 128. Chapter 16. 129. O.Krok. I 41.16–26, 70–71. 130. For example, that of Modestus, or even O.Krok. I 41.47–52: an order of the prefect of Berenike to give water to three donkey-drivers, who probably were carrying this letter on their persons to produce at each praesidium. 131. The meaning of this phrase is elucidated in Chapter 16. 132. Just like the soldiers’ servants in Josephus’s description of Vespasian’s army (B.J. 3.69 f.).

The road system of the Eastern Desert of Egypt

257

monomachai are attested in the desert sites beginning with Trajan, but it is only in Mons Berenicidis that we see them, at a late date, as messengers for the official mails.133 Mons Claudianus has given us only a few cover letters and one notification of receipt, addressed respectively by Ulpius Dios and Diourdanos, curators of Raïma, to two curators of Claudianus, Alexas Amans (c. 150–170) and Archibios (c. 152), as evidence about the official mails.134 In this case, the cover letters concern only one stage, and the control of the schedule is carried out by the curators of the two immediately neighboring forts. The bearers of letters are probably slaves. In Diourdanos’ notice of receipt, the messenger is called tabellarius, while the letters forwarded by Ulpius Dios to Claudianus (which are for the most part imperial letters, kyriakai epistolai) are entrusted to a familiaris.135 Monomaches, tabellarius, and familiaris are likely to be synonyms.136

3. Supplying the praesidia If the caravans of international commerce have left virtually no traces in the ostraca, the sherds are by no means so mute about the supply and travel convoys that operated between the stations. The principal sources for this subject are several copies of circulars of the prefect Artorius Priscillus, who, under Trajan, seems to have had to contend with thefts of construction wood and misappropriations of donkeys, as well as numerous private letters: the civilians, whose business forced them to make frequent trips between the praesidia, awaited the passage of donkey or camel caravans in order to join them. They are called “the poreia” (“caravan”), “the camels,” or “the donkeys.” The definite article suggests that these means of transportation passed by regularly. Everything seems to indicate that what was called “the poreia,” both in the ostraca from Mons Claudianus and in those from the Desert of Berenike, was a supply caravan. It was made up of donkeys and camels, who, because these animals do not walk at quite the same speed, formed two separate groups. Unfortunately, these references leave many questions unresolved: – In what proportions were these animals the property of transportation enterprises137 or instead requisitioned, like the camel borrowed, on order of the prefect of Egypt, from a villager of Soknopaiou Nesos “for the imperial service of the caravans (poreiai) coming from Berenike”?138 – Did the same animals, according to circumstances, transport the supplies intended for praesidia and the Red Sea ports, and also the merchandise for the Red Sea trade? – When animals were unloaded as the supply caravan made its way through the desert, did they continue to travel with the caravan even though empty, or did they return to Koptos in small groups? – One of the most irritating enigmas of the ostraca from the Berenike road is the question of the dekanoi. It is connected to the much larger problem of the dekanoi of Upper Egypt, notably illustrated by a series of lists of day and night guards from the Theban region, drawn up for each 133. Chapter 17, pp. 274 f. 134. Letters received by Archibios are published in Chapter 28. 135. But once to an infantryman (stratiotes): O.Claud. inv. 6967. In these letters, familiaris is sometimes transcribed in Greek as φαμιλιάριος, sometimes φαμιλιαρικός. 136. But in O.Did. 44 (start of the third century), a monomachos uses a familiaris as his messenger. 137. This was probably the case for the camels of the transportation business of Nikanor and company, which had its seat at Koptos and between 6 and 62 provided transportation both for export goods and for the supply of garrisons on the roads to Myos Hormos and Berenike (Cuvigny [ed.] 2006: 274 f.). 138. P.Lond. II 328.

258

Rome in Egypt’s Eastern Desert month. These guards have Egyptian names and patronymics, and the name of one of them on each list is preceded by a symbol familiar to papyrologists, which means (according to the context) dekanos or dekania.139 Roger Bagnall, who analyzed this documentation, concluded that the local population was organized in a kind of militia to provide the surveillance of town quarters on a rotating basis.140 For his part, without knowing the ostraca from the praesidia of the road to Berenike, and based on analyzing a body of contemporary fiscal receipts, also from the Theban region and issued for the benefit of dekanoi and their partners (metochoi), Dieter Hagedorn formulated a hypothesis that the dekanoi, far from being liturgists, were transportation entrepreneurs.141

Three of the praesidia that we have excavated on the Berenike road (Didymoi, Dios, and Xeron) have produced particularly laconic documents, which on the one hand include the mention of mostly Egyptian personal names, some of which are preceded by the dekanos/dekania symbol, and on the other hand an expression for a quantity of wheat, barley, bread, or chaff.142 Connections that it would take too long to explain fully here,143 especially with the customs ostraca from Berenike, show that these dekanoi were involved in the transportation system (in my opinion only camel-drivers are in question) on the Berenike road, during the period betweeen the Flavian emperors and Hadrian. The transporters seem to be identified by their membership in a dekania named after the dekanos. Since the quantities mentioned in the ostraca never correspond to those of standard donkey or camel loads, I think that it cannot be a matter of receipts for supplies delivered to the praesidium, but rather of orders for supply against which these transporters drew their rations in the praesidia for themselves and for their animals. We do not know if these persons were professional transporters who had been hired through a public contract, or if they were compelled to work under the category of a personal liturgy, a munus personale. We have seen that some dekanoi worked also in the quarries, whether as workmen (cf. the ergatai in the third century at Porphyrites), or to take care of local transportation of water from the wells or for the hauling of blocks. The ostraca of the Eastern Desert have thus enriched and diversified the dossier without resolving its contradictions. All that we can say is that in the Thebaid under the Early Empire contingents drawn from the local population were gathered to discharge functions as varied as the maintenance of public order and transportation; but the exact functioning of this organization and its legal character completely escape us.

Conclusion Since the excavations at Mons Claudianus, papyrologists have started to retake the initiative and to direct excavations with the hope of finding written documents, renewing in this way the tradition of the great papyrological excavations of the end of the nineteenth century: Claudio Gallazzi at Tebtynis, Roger Bagnall at Amheida (Dakhla Oasis), Mario Capasso at Bacchias and Soknopaiou Nesos; at Antinoou polis, Rosario Pintaudi continues an Italian tradition of “papyrological” excavations, for the archaeological concession has belonged since 1935 to the Istituto papirologico G. Vitelli of Florence. The harvests of 139. Dekanos means literally “chief of a team of ten men,” the group being called the dekania. I wonder if this symbol, added subsequently in the margin and not necessarily to the first name on the list, may represent a choice by drawing lots. 140. Bagnall 1977. 141. Hagedorn 1995. 142. E.g., “Dekanos (or dekania of) Psenosiris son of Bassus. Peteminis son of Horos. 2 artabas of barley” (O.Dios inv. 804). 143. See the chapter “Décanies et dekanoi” in Didymoi II.

The road system of the Eastern Desert of Egypt

259

texts are, by comparison with the excavations of yesteryear, no doubt somewhat modest: already, on the eve of the first world war, the terrain was starting to be exhausted, less because of the rough and ready methods of the archaeologists than because of the collection of archaeological deposits by agricultural companies, for whom they were a prized source of fertilizer.144 In our times, the rise of the water table and the accelerating urbanization of Egypt are continuing to reduce the likelihood of finding texts. Archaeology has in the meantime become a highly technical discipline with demanding standards; papyrologists do not supervise excavation themselves, although they take part in forming excavation strategy. On sites where they are not simply guests invited to work on material found by chance, the excavation strategy results from subtle compromises. It is easy for the interests of archaeologists and papyrologists to be at odds: each wishes to bring to light that in which he is competent. Not only do written documents not commonly occur in the areas that interest the archaeologists, but the latter in addition dislike excavating the contexts most likely to produce texts, that is, dumps. Actually, it takes an experienced excavator to peel the layers of a dump properly and to discern its stratigraphy. Now this austere work is profitable above all to the specialists of the material found,145 for whom it furnishes assemblages and chronology. From the archaeologist it demands a considerable investment of time and attention, even though it is not very intellectually gratifying and, from the publication point of view, not much valued. The product of an infinite number of baskets of trash thrown away, day after day, the dumps are a precise, but difficult to decipher, reflection of the material life and the short histories of the praesidia. Diversity or homogeneity dominates, according to whether the garbage was the product of regular housekeeping or the result of major cleanings or works,146 visible in thick layers of chaff (animal bedding), ash (residue from ovens and baths), of lime, and of gravel. Large or occasional discardings have been spread out by men, disturbed by pigs, or packed down by natural processes. The layers have interpenetrated, have slid one over the other on slopes, overlap. The ostraca, coming from bedrooms or offices, punctuate this complex stratigraphy. They are precious aids to dating, especially if they include a regnal year. Where dated ostraca are lacking (as was the case at Maximianon), the forms of glassware are the most reliable guide. The repetitious nature of the ostraca and their paleographical connections allow us to reconstruct dossiers and situate them in the broad chronological phases of the layers, which without the ostraca the archaeologist would not have had any particular reason to connect: they are thus invaluable in disentangling the intricacies of the stratigraphy of the trash.147 One might suppose that the interlocking approaches of the archaeologists, papyrologists, and other specialists of the material found in the excavations would automatically reinforce one another. Experience shows that this supposed synergy is largely an illusion. The various categories of material actually form, so to speak, parallel and somewhat impervious worlds, despite our best efforts to establish connections. Even in the Roman installations of the Eastern Desert, which can be studied exhaustively thanks to their isolation, modest dimensions, and limited life-span, it is rare for the ostraca to provide answers to the questions raised by the non-epigraphic remains, or for these to clarify the contents of the ostraca. One of the reasons for this disappointing observation is connected to the very nature of the ostraca. If we had found the archives of a curator praesidii, or those of a centurion in charge of the quarries of Mons Claudianus, things would be different. But these archives on papyrus departed for the valley with their holders or when the sites were finally abandoned. In the dumps we find only the texts 144. Cuvigny 2009: 34 f. and 55. 145. In Egypt, this means, other than the texts, which are generally on papyrus or ostraca, large amounts of ceramics, glass, metal, bones, botanical remains and basketry, textiles, and leather. 146. Renovations, digging of new cisterns, cleaning of the well… 147. J.-P. Brun recognizes that without the ostraca he would have interpreted the stratigraphy of the dumps differently.

260

Rome in Egypt’s Eastern Desert

of a private nature (many letters exchanged between neighboring sites) or, if they are administrative, of an ephemeral interest. The value of these documents thus depends less on their content, almost always anecdotal, than on their numbers, which allow patterns to emerge and which produce series, even if they have too many gaps to allow statistical analysis: even in the Egyptian climate, where ancient deposits are much better preserved than elsewhere, even in the Egyptian desert, where, until the recent development of tourism on the Red Sea, some sites had not been disturbed since antiquity, it remains true that an infinitesimal proportion of the material has been preserved. Among all these private letters, little accounts, lists of names, and jar inscriptions, occasionally we find drafts or copies of more important documents, such as official correspondence and postal logs. We have to content ourselves with this inadequate material in order to wrest from it, even to steal from it, some scraps of history (see Chapter 41).

16 Collection of cases of irregularities in the transmission of official mail The ostracon presented here (P.Worp 51) comes from the dump outside the praesidium of Dios, now Abu Qurayya, better known to historians under the name given it by the Antonine Itinerary, Iovis. A Roman station on the road from Koptos to Berenike, Dios is located between Kompasi and Xeron. The excavations begun at Dios in 2005 brought to light the Latin dedication of this fort and provided the date of its foundation, 115/6. The document is a relative of the postal daybooks of the Desert of Berenike: daily entries, names of cavalrymen who bring and take away the mail, indications of hours. The sherd seems complete at the top, but, since it is broken at the bottom, we have only the beginning of the second entry, the date of which, nearly three months later than the preceding one, shows that we are not dealing with a postal daybook properly speaking. The complete entry, that of Epeiph 19, confirms this in the information that it contains: the scribe has not limited himself to noting the hour of arrival of the mail, but has also given the hour of its departure. This is not normally indicated in the daybooks because it is the same as arrival; but here the mail delayed for nine hours, because Herakles, the cavalryman who was to take it onward, was enjoying female company: the letters had arrived at 8:00 p.m. and sat unmoving until 2:00 a.m. Moreover, one can hardly avoid asking how Herakles managed to escape the vigilance of the curator praesidii: normally, the curator received the mail, registered it in the log, and turned it over to a dispatch rider, who in principle set out on the road right away. The ostracon was written at Dios, and it is at Dios that the letters of the prefect of Egypt (ἐπιστολαὶ ἡγεμονικαί) arrived on Epeiph 19: we indeed know that the cavalryman who brought them, Nepotianus, belongs to one of the two neighboring garrisons, because he is the author of a letter found in the same layer;1 but this letter does not allow us to determine if he was stationed at Kompasi or at Xeron and, thus, in which direction the letters of the prefect were headed. The document looks like a draft, as a crossing-out and several second thoughts indicate. But the exact circumstances in which it was drawn up and the identity of its author escape us. There are two intriguing details: (1) In a side comment (lines 8–9), the author addresses an anonymous interlocutor, 1. O.Dios inv. 40, SU 4413.

261

262

Rome in Egypt’s Eastern Desert

saying to him: “which you can verify”; (2) in the margin, opposite this comment, a note has been written, which must apparently be understood as “I have found.” The note does not seem to be from a second hand, something that makes one wonder if this document is not a copy.

Figure 52. P.Worp 51. © A. Bülow-Jacobsen

Abu Qurayya inv. 39 (magazine of the Supreme Council of Antiquities at Quft). US 4413. Dimensions: 13 × 10 cm. Sherd from the wall of an AE3 amphora. Broken at bottom. Fig. 52.

4

8

12

margin Ἐπειπ ι̅θ̅. ἐλθόντος διπλώματος τῆς ἐπίθεσις καὶ ἐπιστολῶν ἡγεμονικῶν διὰ Νεπωτιανοῦ ἱππέος {ηρα} ὥραν α̅ τῆς νυκτός, Ἡρακλῆς ἱππεὺς ⟦δ⟧ λαβὼν τὰς ἐπιστολὰς ὥραν ι̅ τῆς νυη̣ὕρ(ηκα) κτὸς ἐξῆλθε, ὃ καὶ δύνασαι ἐπιγνῶναι, μετὰ γυν̅α̅ικὸς κοιμώμενος. vacat Φαωφι ι̅ ὥραν η̅ νυκτὸς ] ἱππεὺς ε̣ . [ — — — — — — — —

2 l. ἐπιθέσεως 4 l. ἱππέως 8 η̣ὗρ(ον) ed. pr.

Collection of cases of irregularities in the transmission of official mail 263 “Epeiph 19. As a cover letter and letters of the (or for the) prefect arrived, (brought) by Nepotianus, cavalryman, at the 1st hour of the night, Herakles, cavalryman, having taken these letters, left at the 10th hour of the night, which you can also verify, because he was in bed with a woman. Left margin: I have found. “Phaophi 10, at the 8th hour of the night […], cavalryman, […]” 2.

διπλώματος τῆς ἐπίθεσις. On the meaning of this expression, see the appendix infra.

3.

ἐπιστολῶν ἡγεμονικῶν. The expression ἐπιστολὴ ἡγεμονική is not otherwise attested except in the letter P.Sarap. 84a, the editor of which thought that it did not mean “letter of the prefect” but “letter addressed to the prefect.” The author of P.Sarap. 84a (who is Heliodoros, the eldest son of the Sarapion who is the eponym of the archive) writes (ii.6–9): ἔλαβον Ἡλιοδώρου ἐπιστολὴν ἡγεμονικὴν κακῶς γεγραμμένην καὶ ἀνέδωκα ἵνα γράφηι καλῶς ὃ μεταδώσεις αὐτῶι · λαβὼν δὲ εὐθέως πέμψω, which J.  Schwartz understands this to mean “I received from Heliodoros2 the letter for the prefect; it was badly written, and I have forwarded it so that you can write it better; you will forward it to him and, as soon as I receive it, I will have it sent.” Truth to tell, this parallel is not very helpful; without knowing the situation, we will never be sure of the meaning of this passage, but there are two possible objections to the translation: the absence of a preposition before Ἡλιοδώρου is very awkward, and, above all why would γράφηι be in the middle? It would be more correct to translate “It was badly written and I have forwarded it so that it can be properly drawn up (or for him to draw up properly), which you will let him know”; but one could also take it, on a hypothesis that I find preferable, “I received the letter of the prefect for Heliodoros; it was badly phrased, and I sent it back so it can be correctly phrased, which you will let him know.” This interpretation would save the meaning of “letter of the prefect,” which seems more natural than “letter to the prefect” (cf. the neighboring unambiguous expression κέλευσις ἡγεμονική). Indeed, we know from the beginning of the document that Heliodoros son of Sarapion was for an extended time in close proximity to the prefect (as petitioner or for a professional reason?):3 it is possible that the prefectural letter which had been requested by the homonymous Heliodoros did not provide satisfaction and that Heliodoros the son of Sarapion, recognizing this, sent it back to the scribes of the prefecture for them to reconsider their draft; he therefore asks his correspondent (who was perhaps, according to J. Schwartz, his father Sarapion) to alert the homonymous Heliodoros to the delay his business was experiencing; since communication between the prefect of Egypt and ordinary private citizens did not take place by exchange of letters but by hypomnema and subscription, this epistole hegemonike must have been a letter of the prefect addressed to a subordinate, containing instructions on how to handle some request by the homonymous Heliodoros; we know that it was up to the petitioner to transmit himself the written decision of the prefect about his case to the lower official who was going to carry it out; this procedure offers a satisfactory interpretation of the expression Ἡλιοδώρου ἐπιστολὴν ἡγεμονικήν.

7.

τάς is added projecting in the margin to underline the point that it was precisely the mail brought by Nepotianus that was at stake.

2. A homonym of the author of the letter. 3. On Heliodoros, the intellectual of the family, who has taken to the high seas and works on a large (imperial?) estate at Memphis, from where he profits from his connections to aid his family, cf. J. Schwartz, CdE 34 (1959) 355. Cl. Préaux writes that he made a career in the prefect’s entourage (CdE 37 [1962] 398).

264

Rome in Egypt’s Eastern Desert

8–9.

ὃ καὶ δύνασαι ἐπιγνῶναι – η̣ὕρ(ηκα). One would readily read πυρ, which offers no intelligible meaning; moreover, the pi is not formed like other examples of pi in the text, although this note is apparently in the same hand. This is why I prefer to read an eta, the upper projection of which would have disappeared. A. Bülow-Jacobsen originally suggested to me the attractive resolution η̣ὗρ(ον), “I found,” which may respond to “which you can control.” We do not know what control is in question: a comparison of the postal daybooks of the two neighboring praesidia (in which case, the pertinent passages would be “found,” for example the hour of Nepotianus’s arrival at Dios and the hour of Herakles arrival at the following station, hours which taken together would have disclosed the negligence of Herakles)? But the reverse scenario is also imaginable: comparing the postal daybooks at Koptos, it would be apparent that too much time elapsed between the arrival of the mail at Dios and its arrival at the next station; an inquiry on the spot would then have been launched to discover the reason for this delay.

9–10.

γυ|ναικός. The stroke which appears to stand above the letters να is actually a paragraphos intended to the mark the end of the entry for Epeiph 19. The reason for Herakles’ delay was then added subsequently, and the writing is a bit tighter. Was this perhaps the result of the inquiry that led to the conclusion η̣ὕρ(ηκα)? κοιμώμενος. As the action (if one can call it that) took place before ἐξῆλθε, one would in principle expect an aorist participle; the present may have a durative force.

Appendix: The meaning of δίπλωμα τῆς ἐπιθέσεως Preisigke’s Wörterbuch gives two meanings for ἐπίθεσις: (1) das Daraufsetzen (Bauarbeiten), with just one example;4 it is the etymological sense (“place on”); (2) aggression, attack, which is the most common meaning in the papyri. Might the diploma here concern an attack of the “Barbarians,” several of which we know of (cf., e.g., several of the diplomata copied in O.Krok. 87)? But in this case we would expect περί before τῆς ἐπιθέσεως: cf. the two cavalrymen of Maximianon who left μετὰ δύπλομα περὶ Χινεδοκολπιτῶν (l. Κιναιδοκολπιτῶν).5 One of the papyrological occurrences of ἐπίθεσις offers an interesting path. It occurs in the famous Muziris papyrus SB XVIII 13167.2–4 (BL IX, 299; cf. Rathbone 2001): [δώσω τ?]ῷ σῷ καμηλείτηι α πρὸς ἐπίθ̣εσιν τῆς εἰς Κόπτον [ἀνόδο?]υ̣ καὶ ἀνοίσω διὰ τοῦ ὄρους μετὰ παραφυλακῆς καὶ ἀσ̣φ̣α̣λεί̣ας [εἰς τὰ]ς̣ ἐπὶ Κόπτου δημοσίας παραλημπτικὰς ἀποθήκας, “[I will give?] to your cameleer [(sum of money)?/what he asks?]6 for the epithesis of the journey to Koptos and I will transport (sc. the goods) across the desert under good guard and in all security to the public storehouses of the Collection at Koptos.” The first editors translated ἐπίθεσις as Benützung (“use,” sc. of the road, a toll of some sort), while recognizing that this meaning is not attested either in LSJ or in papyrological lexicons.7 Subsequently,

4. Guy Wagner thought that there was another example in O.Bahria 20, but I have shown that this ostracon was actually an entole coming from Mons Claudianus (Chapter 10). In the entolai, ἐπίθεσις seems to have a specific meaning and to refer to a salary supplement in wine for the workers who were pagani (cf. Chapter 10, pp. 176–78). 5. Chapter 26, pp. 398 f. 6. A proposal of Rathbone 2001: 41. 7. Harrauer and Sijpesteijn 1985: 137.

Collection of cases of irregularities in the transmission of official mail 265 G. Thür8 took it to mean Verladung, another unattested meaning, but one that could derive from ἐπιτίθημι in the meaning of putting on (a transport animal).9 D. Rathbone translated it as extra charges: perhaps, in his view, the toll charges for using the Berenike–Koptos road.10 The idea of looking for a solution among the meanings of ἐπιτίθημι is interesting. It is in this case tempting to connect the epithesis of the Muziris papyrus with the uses of the verb in the archive of Nikanor, another cameleer; cf., e.g., O.Petr.Mus. 152.1–4 (AD 6): Πανίσκ(ος) Ἰσιδώρο(υ) Νικάνωρι καὶ Πετεασμήφι ἀμφοτέρο̣ις χ(αίρειν). ἔχωι παρ’ ὑμῶν ἐπὶ̣ Βερενείκης ἃ̣ ὑμῖν ἐπέθηκα ἐπὶ Κόπτου οἴνο(υ) ἀμιννα̣ίο(υ) γ̣όμ(ους) β κερά(μια) τέσερα̣ς̣. About this formula, D. Rathbone writes,11 “Storage is implied by the ἐπέθηκα/ἐπέθηκε texts,” and also, “The family of Nikanor (…) ran camel trains between Koptos, the base of their operations, and the Red Sea ports in the Julio-Claudian period, and provided storage facilities at all three locations.” But the idea of putting into storage,” of “stocking” (present in ἀποθήκη or ἀπόστασις) does not connect with the semantic field of ἐπιτίθημι/ἐπίθεσις. I think that ἐπιτίθημι has the same meaning in the ostraca of Nikanor12 as when it means “entrust a letter to a messenger for him to deliver it to its addressee.” This meaning is clearly set forth in ThGrL, s.v., 1843D: nec non et epistolam ἐπιτιθέναι alicui dicimur pro Ferendam sive perferendam dare; the citation from Synesius (Ep. 86) is clear: ἐπέθηκα τῷ ἀδελφῷ πρὸς κόμητα γράμματα (fratri ad comitem perferendas literas dedi).13 Nor is this use of ἐπιτίθημι limited to letters: in the Muziris papyrus (where we have the noun ἐπίθεσις, which can be rendered “taking in charge,” although it is literally a “putting in charge”) and the archive of Nikanor, it is a matter of goods entrusted to cameleers to be delivered to their final destination or recipient; in P.Eleph. 30.2–5 (third century BC), the verb concerns a sum of money: ἀπέχω παρὰ σοῦ τὰς κ (δραχμὰς), ἃς ἐπέθηκέν σοι Ζμῆτις Ψαμματίχου ἀποδοῦναί μοι (“I received from you the 20 drachmas that Zmetis son of Psammatichos entrusted to you to give to me”). In an ostracon from Umm Balad, O.KaLa. inv. 243, it is preserved meat: πολλοὺς ἠρώτησα ἆραί συ κεράμιν κρεῶν κ’ οὐδεὶς οὐ ἦρεν (…)· ἐὰν εὕρω πῶς ἐπιθῶ, εὐθέως σ̣υ πέμψω, “I asked many people to take the amphora of meat for you, and no one took it (…); if I find a means of having it transported, I will send it to you as soon as possible.” The same usage appears in O.Krok. II 316, where Ischyras has selected a fine piglet for Parabolos but cannot find anyone to carry it to him at the next station: λοιπὸν πῶς αὐτό σοι ἐπιθῶ οὐκ οἶδα, “but I do not know how to get it transported for you”; here we observe that σοι is a dative of interest and does not represent, as in the ostraca of Nikanor or in P.Eleph. 30, the transporter to whom the object has been entrusted; all the same, the meaning of the verb remains “give for transportation,” and this has sometimes escaped translators: in O.Claud. II 365.5, the writer does not know to whom to entrust a delivery of an unknown item intended for his correspondent: οὐ γὰρ εὑρίσκ̣[ω] τίνι αὐτὰ ἐπιθῖναι διὰ τὴν ἔνδιαν τν κτηνν (mistakenly translated in the edition: “je ne trouve pas sur quoi les charger en raison de la pénurie de bêtes de somme”). The same confusion appears in the case of PSI VI 623.16–25 (third century BC): σὺ καλῶς ποήσεις δοὺς τῆι παιδίσκηι ὅ τι ἄν σοι δόξηι, ὅπως ἀναπλεύσω εἰς τὴν Θηβαίδα. εἰ δ’ ἔτι βούλη⟨ι⟩ τισὶν γράψαι τι, ἐπίθες· ἀποδο̣θ̣ή̣σεται γάρ. The anonymous author asks Zenon to give to a slave who serves them as an intermediary his travel expenses for a trip to the Thebaid, and he addition8. Thür 1987: 234, n. 8. 9. WB, s.v. 5; cf. Hdt 2.121: ἀσκοὺς πλήσαντα οἴνου ἐπιθεῖναι ἐπὶ τῶν ὄνων. 10. Rathbone 2001: 41. 11. Rathbone 2002: 189 and n. 36. 12. Where it was correctly translated by A. C. Johnson, Roman Egypt (1936), no. 230: “consign in your charge.” 13. The same construction appears in three passages of P.Panop.Beatty, but there the idea of “entrust a letter (to a messenger) for someone” has no other raison d’être than to make the style more solemn, because in each case we are dealing with letters sent by high officials: δι’ ὧν ἐπέθηκεν πρός με γραμμάτων (P.Panop.Beatty 1.364; cf. also 1.72 and 1.213).

266

Rome in Egypt’s Eastern Desert

ally proposes, “If you wish also to write any mail to these people, entrust it to me: it will be distributed to the proper recipient.” The passage does not concern adding a postscript to a letter (WB s.v. 9), nor adding mail to the travel expenses (C.Ptol. Sklav. II 150, translation). In the Dios ostracon, the context is one of consignment of mail to a postal service. The δίπλωμα τῆς ἐπιθέσεως is the cover letter that accompanies the letters of (or to) the prefect and that asks their careful handling by those responsible at each stage. The circular of which O.Krok. I 47, iii preserves a copy is an example of a δίπλωμα τῆς ἐπιθέσεως: addressed at least to the duplicarii and to the curators of the road (of Myos Hormos), it asks them to take care that the mail reach the prefect of Berenike Artorius Priscillus without delay, while recording all the information in the postal daybook concerning letters, some of which appear to come from the prefect of Egypt: τὰς δὲ ἐπιστολὰς τα̣ύτας [τοῦ κρατί]σ̣του ἡγεμόνος καὶ Ἀρτωρίου [Πρισκίλλ]ου (since the recipient is precisely Artorius Priscillus, there must be a mistake on the part of the copyist, due in my opinion to the routine character of this final recommendation). It is likely that only the most important mail was accompanied by this sort of cover letter. Herakles’ neglect is only more blatant.

17 The postal register of Turbo, curator of the praesidium of Xeron Pelagos I. The praesidia, relay points of the public post The ostraca published in this chapter were found in the dump of the praesidium of Xeron Pelagos, the excavations of which I directed between 2010 and 2013.1 This small fort is located between Dios and Phalakron, on the road to Berenike, in a wide sandy plain, from which comes its Greek name “Dry Sea.” The place-name appears in the shortened form Xeron, also well attested in the ostraca, in the Peutinger Table and in the Ravenna Cosmography. The road to Berenike (ὁδὸς Βερενίκης) was, together with the road to Myos Hormos, one of the two great transverse roads structuring this part of the Eastern Desert, which was for the most part devoted to the Red Sea trade and which the Romans called Mons Berenicidis (var. Berenices). This territory, of which the northern and southern boundaries were a bit vague (at least for us), corresponded more or less to the ḏw Gbtyw (“Desert of Koptos”) of the pharaonic period and to the “isthmus” so clumsily conceptualized by Strabo.2 The hodos Berenikes took form only thanks to the small forts that lined it at intervals of, on average, thirty kilometers: no stone pavement, no cobbled road, not even a path cleared of stones like the via Hadriana. The loose sand of the wadis was what suited best the main means of transportation along this road: the camel. Donkeys also traveled on this road, but not wagons (in contrast to the road to Myos Hormos, which crossed firmer ground and was suitable for wagons).3 And, of course, men – and women. In a letter found at Didymoi, Nemesous, a madam who was bringing a prostitute to her place of work, describes her wrangles with the donkeydrivers: stingily, a single donkey had been hired, on which they would take turns riding, and the donkeydrivers took advantage of the weakness of these two isolated women.4 Among other functions, the praesidia along the roads of the Desert of Berenike served as relaypoints for the transportation of official mail: a small number of cavalrymen were kept ready, day and night, to carry letters and packages at once toward the next praesidium. The curator praesidii who com1. I thank Rudolf Haensch and Naïm Vanthieghem for their critical and careful reading of the manuscript of this chapter. 2. Cuvigny in Cuvigny (ed.) 2006: 3–10. 3. Bülow-Jacobsen 2013: 567. 4. O.Did. 400.

267

268

Rome in Egypt’s Eastern Desert

manded the garrison also filled the role given in modern French armies to the postmaster; he kept a register of the mail that passed through his hands. In my edition of the ostraca of Krokodilo, I attributed this arrangement to the “military mail.” This was, I think, a mistake:5 “military mail” would refer to an organization specific to the army and only for its use. But the letters of the prefect of Egypt and of the central administration in Alexandria were carried by the cavalrymen of the praesidia. I now think, therefore, that the “postal” ostraca of the Desert of Berenike bear witness to a system of public post in the Roman province of Egypt, even if this public post is strongly marked by its regional particularity: to communicate with the ports on the Red Sea, the central administration relied, for the crossing of the desert, not on the network of strategoi of the nomes,6 but on that of the praesidia. Circumstances dictated that a large part of the official mail, particularly the circulars addressed to the curators, was also of local interest, but the importance of this “military” mail is easy to overestimate, because it was a matter in essence of circulars of which the curators made copies. The mail of the civil administration did not concern them and for this reason is less “visible” in our documentation. Most of the small forts that have been excavated under my direction in the Desert of Berenike have produced inscribed pottery fragments that testify to the existence of postal registers. I have published those from Krokodilo in O.Krok. I.7 Didymoi and Maximianon are fairly poor in material of this type. In contrast, Dios and Xeron have both produced the remains of the postal register kept by curators whose names we know: Dinnis8 at Dios, Turbo at Xeron. Their diplomatic behaviors are different. Readily recognizable by their handwriting, most of the fragments of postal registers that have emerged from the trash of Xeron were drawn up by this Turbo (or under his direction). The sherds are scattered in various squares of the dump. Joins have made it possible to reconstruct our no. 1, which gives a good idea of the columnar presentation of Turbo’s books. More modest, 2 gives information otherwise unknown on the packaging of official mail. Many scraps could not be joined, but almost all of them are worth the trouble of editing, because they contain a day or a name, or confirm a reading elsewhere. Table 17.1. Complete list of the fragments of Turbo’s postal register Inventory No.

Stratigraphic Unit

Date

1

inv. 618 + inv. 1015

US 50904, 60514

Mecheir 14, 20, 22

2

inv. 257

US 80803

?

3

inv. 279

US 80807

Phamenoth

4

inv. 1030

US 60509

Phamenoth

5

inv. 1241

US 50507, 50510

Pharmouthi

6

inv. 46

US 50618

Pharmouthi

7

inv. 106

US 50618

?

8

inv. 58

US 40844

?

inv. 1013

US 60506

?

discard

US 50619

?

5. Already corrected in Chapter 18, pp. 308 f. 6. Cf. P.Ryl. II 78. 7. Chapter 11. 8. See Chapter 18, pp. 311 f.

The postal register of Turbo

269

II. Date No regnal year number appears in these documents, but the date is secured by the mention, in lines 29–30 of 1, of the prefect of Egypt L. Volusius Maecianus (13 February–15 November 161); even if his name is not only damaged but also misspelled, its reading is confirmed by the mention of the tabularii Narcissianus and Fortunatus (1, iii.35–36). This pair is attested the following year in SB XIV 11612, a copy of a letter of the prefect of Egypt who succeeded Volusius Maecianus, Annius Syriacus (end of 161 or beginning of 162 until 2 March 164): in this circular, addressed to the strategoi of the nomes, Syriacus refers to a financial report submitted to him by the ταβλάριοι Narcissianus and Fortunatus; the circular, the date of which is lost, was copied into a register of official correspondence after a letter dated to 23 November 162. The diploma reproduced in col. iii of our 1 is dated Mecheir 20, which can only be (Julian) 14 February of year 24 of Antoninus (the dies imperii of Marcus Aurelius was 7 March 161).9 As it happens, a Fortunatus is also known, this time in the preceding year, in the top echelons of the Alexandrian administration: he is the dedicant of OGIS 707,10 which honors T. Furius Victorinus, who, after having been prefect of Egypt (10 July 159–28 September 160), had just been named pretorian prefect; this promotion is thus to be situated between October 160 and January 161. In the inscription, Fortunatus identifies himself as Σεβαστοῦ ἀπελεύθερος, ἀρχιταβλάριος Αἰγύπτου καὶ ἐπίτροπος προσόδων Ἀλεξανδρείας. Sergio Daris, the first editor of SB XIV 11612, identified with him the Fortunatus who was the colleague of Narcissianus. That would require him to have been demoted under Maecianus from being head of the tabularii to a simple tabularius, or else that he had exaggerated his titles in his dedication; this last explanation, formulated by S. Daris,11 was greeted with skepticism by François Kayser12 and R. Haensch, who observed that Fortunatus was a common name.13 That is actually not the case in the papyrological and epigraphic corpus of Egypt, but this corpus probably does not adequately represent the names found in the central administration of the province, which was full of bureaucrats who belonged to the familia Caesaris.

III. What sort of postal register is involved? The activity of a curator praesidii as postmaster generated various types of documents that were related but separate. The information capable of appearing there included the following items: – Description of the mail: nature, number, senders, recipients, material characteristics; – Subject of the letters: simple mention of the object of the missive14 or a more or less full copy of the text; – Identity of the messengers who brought or took away the mail; – Date and hour of arrival, and, if appropriate, of departure of the mail. Before classifying the postal registers from the Desert of Berenike, it will be useful to distinguish the types of missives that passed through the hands of a curator: 9. An eminent jurist and valued counsellor of Antoninus and his successors, Maecianus made his whole career at Rome (Pflaum 1960, no. 141). For R. Haensch, the break that his short prefecture of Egypt represents was aimed at guaranteeing the loyalty of the province at the moment when the news of the death of Antoninus would be arriving (email of April 14, 2017). 10. = I.Alex.impér. 21. 11. Studia Papyrologica 11 (1972) 27 f. 12. I.Alex.impér., p. 83, n. 1. 13. 4102 occurrences in the EDCS database (email of April 2, 2018). 14. E.g., O.Krok. I 30.44: μετὰ διπλώματο(ς) Ἀρτωρίου Πρισκ(ίλλου) ἐπάρχου περὶ ξύλων.

270

Rome in Egypt’s Eastern Desert – Circulars addressed to officers and subordinate officers, including the curators (and sometimes only them).15 Among these we can recognize independent copies (O.Krok. I 13) or copies integrated into the books of correspondence (infra). They circulated in open form, and the curators had knowledge of them and copied them before sending them on in circulation. They are often called δίπλωμα (rather than ἐπιστολή); – Letters addressed to individual functionaries. These were not necessarily closed (i.e., sealed), but the curators would forward them without recopying them;16 – The dispatch notes accompanying a postal bag. Their technical name is δίπλωμα τῆς ἐπιθέσεως.17 In fact, they take the form of circulars addressed to those responsible for the postal relay-stations, who were essentially the curatores praesidiorum.

We have just seen that the circulars, whether dispatch lists or letters containing orders or information addressed to several recipients, tend to be called δίπλωμα. Did this noun have a weakened sense, referring simply to the multiplicity of recipients, or did it mean that the circulars were double documents with a scriptura interior protected by seals, because this is the traditional meaning of δίπλωμα, whether it was a matter of tablets or of a document on papyrus?18 I have long thought that it was not very likely that the διπλώματα mentioned in the desert ostraca were double documents, but the case of the postal dispatch notes leads me to reconsider my position: it is not inconceivable that they included a sealed copy, so as to forestall any attempt to suppress the mention of a missive that had gone astray en route, or to modify an indication of the hour in order to cover up a delay. The postal logs of the curators of the Desert of Berenike belonged to three types: (1) Transcripts of daily movements, indicating the identity of the cavalrymen who arrived and left, with the times.19 The transmission of official mail was not the only reason for these movements: there were also escorts (O.Krok. I 4 seems to register arrivals and departures with no connection to the postal service). The items transported are identified with the name of the sender, if it concerns an important person, and if it concerns a circular, with the content; ordinary letters, simply called ἐπιστολαί, are not detailed. These transcripts could be the product of the compilation of daily slips drawn up at each arrival or departure, such as O.Krok. I 120 (Trajan/Hadrian), SB XXIV 16187 (Maximianon, c. 150, published in Chapter 26), and O.Did. 52 (third century). But it is strange, if one adopts this hypothesis, that so few of these slips have been found. (2) Libri litterarum allatarum (register of copies of received circulars): Two important examples are provided by O.Krok. I 41 and 44, where only the day of receipt is indicated at the top of some of the copies; in O.Krok. 47.36, only the hour of receipt is noted before a copy. In O.Krok. 50, the presentation varies: there is no note of receipt before the first copy; date and hour (l. 14); only the date (l. 24); only the name of the messenger (l. 30).

15. E.g., O.Krok. I 13. 16. The convex side of O.Krok. I 10 forms an exception: it is the copy of a letter addressed to the prefect of Berenike by a curator who appears not to be that of Krokodilo. However, Capito, the curator of Krokodilo, copied it into his postal register. 17. Chapter 16. 18. P.Oxy. LI 3642.17–21n. 19. O.Krok. I 1 (the most complete example; one of its entries is quoted in Chapter 18, p. 309); O.Krok. 24–40; 89; 90.

The postal register of Turbo

271

(3) Registers of transmission of post bags. These include the copy of the dispatch note, followed by a note from the curator, in which he certifies the safe reception and immediate redispatch of the items enumerated in the note, indicating the identity of the messengers involved as well as the date and hour of the arrival of the mail. The departure time is not normally mentioned, because it is in principle the same as the time of arrival. (4) Finally, we should mention the possibility that the curator kept a register of his own letters sent to the hierarchy. O.Krok. I 91, which is in very poor condition, might be a liber litterarum missarum, containing copies of letters sent by the curator himself, followed by the name of the messenger who carried the letter packet (date and hour of departure, now in the lacuna, were presumably indicated). It is significant that the mail, even of a “professional” nature, exchanged between curators, was not registered in the postal logs. That did not, however, keep it from being transported by the same cavalrymen, along with the rest of the private parcels that the riders were willing to take care of. This unregistered correspondence between curators was in fact written, at least partly, on ostraca. Turbo’s registers belong to our type 3. Apart from 2.17–26, all of the sherds are written in the same handwriting, which for the sake of convenience I shall attribute to him, even if the intervention of a secretary cannot be excluded. The writing is experienced but not very careful and lacking in any pretense to elegance; this can be explained in part by the character of these documents, which were unofficial copies: the “real” postal register of Turbo was presumably drawn up on papyrus. But it is also a matter of personality: at Dios, the curator Dinnis (or his secretary) seems from all evidence to have had a passion for calligraphy.20

IV. Diplomatics The postal registers of Turbo were written in several columns on the bellies of AE3 amphoras in the shape of double truncated cones. Were these amphoras the postal register itself?21 I find this hard to credit. Although I cannot prove it, it seems to me more likely that the curators drew up a fair copy of their postal register on papyrus, which they would dispatch at regular intervals to the prefecture of Berenike for checking. I imagine that in that office the dates and hours of arrival and departure at each end of the two roads were compared and that, if the transportation of a postal bag turns out to have taken too much time, the registers of the praesidia in between would be scrutinized to identify the source of the delay. The registers on amphoras or ostraca are in my view only drafts or personal copies, written on this support in order to save on papyrus, which may have been in short supply in the praesidia, even for the curators.22

1. The dispatch note A. The Address Only three authors of these notes have been identified:

20. Chapter 18, p. 311: photo of an ostracon from the postal register of Dinnis, drawn up in Greek, but in Latin characters and in a fine scriptura actuaria. 21. So Remijsen 2007: 139. 22. Cuvigny in Cuvigny (ed.) 2006: 266.

272

Rome in Egypt’s Eastern Desert – Volussius Vindicianus (1, col. i and col. iii), several times referred to as ὁ κράτιστος ἔπαρχος. His mail originates in Koptos: he is thus very likely to be the prefect of the Desert and/or the prefect of the ala of cavalry stationed in Koptos, which was at that time the ala Vocontiorum; – Lucianus, known only by his cognomen (2.17). He is presumably to be identified with the decurion Lucianus mentioned as the addressee of mail in note 1, iii.29. He was perhaps stationed at Apollonos Hydreuma, a large station to the south of Phalakron; located on the Berenike road, Apollonos Hydreuma also commanded a by-road to the emerald mines; – Didymianus, also known only by his cognomen (6.6), who may also have been a decurion.

Other notes indicate that the mail had departed from Berenike (2.10; 5.4), but the identity of their author is lost. These notes are circulars by necessity. They are addressed collectively not only to the curators of the Berenike road but also to centurions and decurions, designated only by the appropriate symbols,  (1, i.2) and  (5.9).23 The ranks of centurion and decurion are similarly abbreviated (unlike ἐπάρχοις, δουπλικαρίοις, and κουράτορσι) in the opening of a well-preserved circular in O.Krok. 87.15. In 2 and 5, the decurions () are the first-named recipients, a fact that excludes the possibility that the note was also addressed to centurions, because these addresses observe the hierarchical order; it is all the more curious that the occurrences 1, i.2 and 5.9 appear in the notes coming from the prefect of the Desert. I wonder if there was not an error in copying in one or more cases. B. The list of items This part of the note is fairly well preserved in 1 and 2. It allowed at each stage checking to see that all of the items in the packet were still present. These items are identified by a description or by mentioning the sender and/or the recipient, but only when these are persons of a certain importance, as I already pointed out in my introduction to the edition of the Trajanic postal register O.Krok. I 1. The items enumerated are almost systematically given in the accusative in the two notes from the prefect Vindicianus (1); these accusatives depend on the verb introducing the circumstances of dispatch.24 They are in the nominative in two notes covering mail sent from Berenike (2 and 5). Does this difference reflect the drawing up of the originals, or was it introduced by Turbo? The question is also relevant, in a more interesting fashion, for the presentation of the name of the prefect Volussius Vindicianus in 1, iii.32–33. 1, iii.23–50 is a note of which the author is Vindicianus himself, and yet one of the letters listed, of which he is the sender, is described as coming from “his Excellency our prefect Vindicianus.” Did this mark of respect figure in the original note? It is hard for me to imagine such a solecism on the part of a secretary belonging to the officium of a praefectus. This addition in my view must reflect an initiative on the part of the curator, who, while copying the note, thought it proper to mark his respect for his superior; it may also be a kind of reflex action. The fact that the postal registers were destined to be checked to detect delays and failures25 may not be irrelevant in a small deviation of this sort from the original. Adaptation of copies of official documents according to the situation was a normal practice: cf. the addition of γενόμενος before the title of a functionary who had left office by the moment of the copy, or that of θεός before the name of a now-deceased emperor. 23. I leave out of account 2.17, which is an ambiguous case. 24. Cf. infra C. 25. Cf. Chapter 16.

The postal register of Turbo

273

C. Circumstances of the dispatch of the postal packet and its note They are introduced by the verb ἔπεμψα or, more often, ἀπέλυσα, followed each time by the place of dispatch (ἀπὸ Κόπτου, ἀπὸ Βερνείκης).26 The question of faithfulness to the original poses itself also with respect to what followed: in the two notes from the prefect Vindicianus, the date and hour of dispatch, as well as the identity of the messenger or messengers who carried the mail toward the first stage are indicated (1, i.9–11 and iii.46–50); in the note written at Berenike, this information is omitted. Is this because it was not on the original, or because Turbo was modifying the faithfulness of his copy to suit the importance of the author of the original? Finally, we would have expected these notes, as circular letters, to have ended with a formula valedicendi which, written on the original in the hand of the author, had an indispensable value for authentication. The ostraca of the Desert of Berenike have given us some examples. There are two in the postal register O.Krok. I 51 (109), lines 19 (in Latin) and 29; O.Did. 28 (176 or 208), a fragment of a document similar to those of the dossier of Turbo, preserves the bottom of a dispatch note with the request made to the curators to resend the mail without delay, the identity of the monomaches who brought it from Koptos,27 and the formula valedicendi, under which the curator of Didymoi added his note of redispatch. Now none of the dispatch notes copied by Turbo presents the final greetings; if Turbo added, other than his spelling errors, marks of respect attached to the name of the prefect, he cut out the end of the notes: not only the formula valedicendi, but probably also the routine exhortations about the necessity of putting the mail back en route as quickly as possible.28 Actually, he seems to have changed his mind in 1, i.8 where he added back the two or three first words of this request in the space between the lines. The copy of the notes is thus not a perfectly faithful reflection of the originals. The deviations come first because of the phenomenon of adaptation, and then because we are dealing with hasty and abridged copies.

2. The postmaster’s note This is both an acknowledgment of receipt and a declaration of dispatch. It begins with the subject of the two verbs that structure it: “Turbo, curator of the praesidium of Xeron.” The first verb, παρέλαβον/-βα, ἔλαβα, “I have received,” is immediately followed by the καθὼς πρόκειται, which occurs regularly in the subscription of declarants at the foot of contracts or of a declaration and serves to authenticate, by confirming them, the contents of the preceding text. In the present case, καθὼς πρόκειται expresses agreement between the list of items enumerated in the cover note and those that Turbo has in fact received and redispatched: in other words, Turbo certifies that no item has been lost when the post bag passed through his hands. Then, the order of listing being a matter of indifference, come the date and hour of receipt as well as the identity of the messenger(s) who brought the mail from the two neighboring forts (Dios and Phalakron), introduced by παρά or ἀπό. Only 1 has preserved the information that would allow us to calculate the speed of transmission of the mail: after Koptos on the 20th at the 9th hour of the day, it arrives on the 22nd at the 5th hour of the day, thus 44 hours to cover 192 km, giving an astonishingly 26. ἔπεμψα: 1.45; ἀπέλυσα: 1.8; 2.10; 5.4; 6.5. In O.Krok. I 51.18, the verb is ἐδόθη, calqued on the datum or data of official correspondence in Latin. 27. Unfortunately, without date or hour. 28. The same orders also appear in circulars as well as in the dispatch notes. For the latter: O.Krok. I 51.17–18, 26–27; O.Did. 28.3–6.

274

Rome in Egypt’s Eastern Desert

low average speed of 4.3 km/hour. Of course, we have to take into account the time necessary for the curator of each of the five stages between Koptos and Xeron to verify the mail and copy the cover note and any circulars, but these operations would not take much time: according to O.Krok. I 83, a half-hour would suffice, so that the curators could write, as Turbo also does, that they had sent the mail onward within the hour: καὶ εὐθέως ἀπέλυσα τῇ αὐτῇ ὥρᾳ. The identity of the messenger(s) leaving Xeron is introduced with διά. The instructions set forth fifty years earlier in the circular O.Krok. I 47.52–58 (109) were thus still in force: “these letters for his Excellency the prefect (of Egypt) and for Artorius Priscillus, after you have noted at what hours and from whom you receive them and to whom you hand them on, have them proceed swiftly to his Excellency the prefect Artorius” (sic).

V. The identity of the messengers and the hours of circulation In P.Ryl. II 78, the messengers of the official mail in the Egyptian chora, who were at the service of the strategoi of the nomes, were called ἐπιστολαφόροι. In the Desert of Berenike, a territory under army control, the dispatch riders put at the disposition of the curators for the transmission of mail by relays were partly auxiliary cavalrymen and partly μονομάχαι (or μονομάχοι). The latter are probably slaves.29 Although no ostracon from the Eastern Desert connects them with mounts, it seems to me improbable that the official post across the desert would have been entrusted to men on foot, like the ἡμεροδρόμοι of the Greek cities, even if accompanied by a donkey to carry burdens and the necessary water supply. The use of μονομάχαι to serve the official mails seems to be a post-Trajanic phenomenon: the word does not appear in Trajanic ostraca of Claudianus30 or at the metallon of Umm Balad, where the finds of ostraca date mainly from the reigns of Domitian and Trajan, except for a single occurrence in a workman’s entole, a genre that appears under Hadrian in the quarries of the Eastern Desert.31 The ostraca of Krokodilo, which date from the reign of Trajan and the start of the reign of Hadrian, contain mentions of μονομάχαι, but they never appear at that time as messengers for the official post. Their use in this service seems to have begun later. Subject to closer study, it may be said that the army seemingly turned to them more and more often in order to reinforce the regular troops. It is unclear why the mail was entrusted to one or the other category of couriers. Turbo’s register could give the impression that the monomachai were stationed in the desert praesidia rather than in the camp at Koptos: all of the messengers setting out from Koptos in fact are equites. But O.Did. 28 (176 or 208) indicates that the postal connection between Koptos and Phoinikon was provided by the monomaches Chrysoplokamos.32 The following table gives a complete list of the messengers mentioned in the cover notes and in the notes of receipt and redispatch kept by Turbo. As the post bag was put back on the road immediately, the hour of arrival at Xeron from Dios or Phalakron is the same as the hour of departure. The local modern time equivalences are given based on the hours of rising and setting of the sun on 1 March 161 at the midpoint between Koptos and Xeron, adding two hours to the times given in universal time 29. Chapter 15, p. 256. 30. With the possible exception of O.Claud. inv. 312, found in a possibly post-Antonine layer of the South Sebakh, a deposit otherwise essentially Trajanic. 31. O.KaLa. inv. 574. 32. Chrysoplokamos, to be sure, may have been attached to the praesidium of Phoinikon, from where he will have been sent to Koptos with the mail that had come from the south, and where he will have returned with the mail entrusted to him in Koptos.

The postal register of Turbo

275

zone.33 I had thought that I detected in the Trajanic postal registers from Krokodilo a reluctance to let the cavalrymen travel by night, except when they were transporting fish.34 This is not the case in our dossier: three times the courier arrived at Xeron at the 9th hour of the night (around 2 a.m.) and was immediately redispatched by Turbo. When the mail arrived at the 11th hour of the day and Krenites sets out with it immediately, it was one hour before sunset. Table 17.2. Movement of messengers in Turbo’s register No.

Place of Departure

1, i

Koptos

Isidoros s. Ammonianos (turma of Salvianus [sic])

Mecheir 14

Dios

Gigas and […], monomachai

[…]

Xeron

Narkissos and […], monomachai

[…]

9th hour of the night

Koptos

Posidonios (turma of Apollinaris) and Isidoros s. Ammonianos (turma of Sabinus)

Mecheir 20

9th hour of the day = c. 12:30

Dios

[…] (turma of [gentilicium (?)] Alexandros)

Mecheir 22

Xeron

Maron (turma of Salvianus)

Mecheir 22

Dios or Phalakron

[…] (turma of [gentilicium (?)] Alexandros) and Serenus (turma of S…)

[…]

Xeron

[…] (turma of Salvianus)

[…]

4

Dios or Phalakron

[…] monomaches

Phamenoth

5

Xeron

Glaphyrinos, monomaches

Pharmouthi

6

Dios or Phalakron

…ax and Chenos, monomachai

Pharmouthi

Xeron

Krenites, monomaches(?)

Pharmouthi

1, iii

3

Messenger

Date

Hour of Departure

Hour of Arrival at Xeron

8th hour 9th hour of the night = c. 2 a.m.

5th (?) hour of the day = c. 10:15 5th (?) hour of the day 9th hour of the night 9th hour of the night

9th hour of the night 11th hour of the day = c. 16:45 11th hour of the day

In 1, the cavalryman Isidoros son of Ammonianos, probably stationed at Koptos, is attached first to the turma of Salvianus, then to that of Sabinus. There must be a slip in one of the two cases, which can be explained by the fact that at least one cavalryman stationed locally, and thus familiar to Turbo, belonged to the turma of Salvianus. The most probable conclusion is thus that the turma to which Isidoros belonged was that of Sabinus. 33. Sunrise: 4:10 UT, sunset: 15:45 UT. 34. O.Krok. I, p. 17; Remijsen 2007: 137.

276

Rome in Egypt’s Eastern Desert

VI. The material aspect of the post Apart from the three pots containing roses sent to Berenike for the festivals of Isis, the objects transported were letters (ἐπιστολαί), circulars (διπλώματα), and bundles (ἀπόδεσμοι). Certain letters are noted as being ἀναγκαῖαι, “important, urgent, priority” (1, iii.25–26 and comm., 5.10). In the Eastern Desert, this epithet is distinctive to Turbo’s register. Some are ἐσφραγισμέναι, “sealed,” others λελυμέναι, “untied,” thus “open.” I think that λελυμέναι functions simply as the opposite of ἐσφραγισμέναι and does not imply that the open letters had originally been sealed.35 It would be natural to suppose that these last were circulars intended to be read by all of the curators through whose hands they passed. However, one of the “untied” letters is addressed to one centurion by another (1, iii.41–43). Was it of general interest, such that the sender considered it useful for the curators to be informed of it? But such a shortcut was not part of the habits of official communication: we would expect that the letter would appear as a copy, introduced by a few words in a circular addressed to the centurions, decurions, and curators.36 Just a few years before our dossier, P.Ryl. II 78 (157) may point the way to a solution: this letter, addressed by someone whose identity is lost in a lacuna37 to Herakleides, strategos of the Bousirite nome, is principally a dispatch note, despite a passage concerning mail that has gone astray and a reminder of an earlier dispatch. There is no question of ἐπιστολαὶ λελυμέναι, but the three lists of mail38 identify each letter with a short summary and their intended recipient (generally a strategos). Although brief, the summaries are too detailed to have been composed on the basis of some indication of the object of the letter on the outside. From this, then, we must conclude that these letters were not sealed. For the editors of P.Ryl. II 78, the object of having official mail circulate in this fashion was to spread the information to those who were not its addressees.39 I do not think this is right: these letters, to judge from their summaries, concerned matters as diverse as they were specific, and it would have required a sustained intellectual effort to identify local analogous situations for which they would have been relevant. And what official, already overwhelmed by paperwork, would have felt himself concerned by that of his colleagues? When the administration considered it useful to bring to the attention of other addressees a letter addressed to an official, it took care to personalize it, by having it preceded by a cover letter in which the opening would mention the new addressees. The non-closing of letters was perhaps intended to allow their identification and to follow their course in instances where they had been lost, an accident that cannot have been out of the ordinary: cf. P.Oxy. LX 4060.57–58, ἐπεὶ δὲ συμβ[α]ί̣νει παρʼ αἰτίαν τῶν διακομιζόντων παραπείπτειν τ[ι]νάς, “as it happens that some (letters) have been lost through the fault of the messengers.” None of the letters mentioned in P.Ryl. II 78 thus seems to have been sealed, and one gets the impression that it was not the usual practice to seal correspondence entrusted to the public post. R. Haensch had already reached this conclusion by another route: even if private letters, which were not handled by the official mails, normally seem to have been sealed, administrative correspondence was sealed only when the author insisted on confidentiality or when, in exceptional circumstances, the seal was necessary to authenticate the source of the letter.40 35. In the contrary case, the question would arise of knowing by what authority and at what moment during the process these letters, entrusted in sealed condition to the public post, would have been opened. 36. See, for example, O.Krok. 87.14–50 or P.Oxy. III 474 (post 184 or 216). 37. Kruse 2002: 817 and Thomas 1999: 189 consider that it is the strategos of a neighboring nome. 38. Lines 2–17 and 27–36: letters arrived with the cover note; lines 36–38: reminder of the letters of an earlier dispatch. 39. Interpretation accepted by Kruse 2002: 817. 40. Haensch 1996: 451–54, 474.

The postal register of Turbo

277

Aside from the sealed letters and the “untied” letters, documents 1 and 2 mention some ἀπόδεσμοι. The use of this term to refer to letters bundled up in packets is, in the context of the Eastern Desert, specific to Turbo’s register. In the papyri, the only instances are P.Oxy. VII 1070.39, a private letter of the third century,41 and P.Ryl. II 78.18 and 36, just four years before our dossier. The contents of such bundles are not given in detail, because they were not intended to be undone during the journey. As a result, they are identified only by their external material characteristics. In 1, iii.38–39, the apodesmos is sealed with a lead seal; the two apodesmoi mentioned in 2.3–4 and 8–9 are wrapped, one in leather, the other in linen. Katelijn Vandorpe has provided a list of six lead seals dating to the early Empire and coming from Egypt,42 the editors of which supposed that they served to seal containers of official mail, whether bags or wooden boxes. Four have been found at Rome and one at Lyon. The bottom of some of these seals has preserved the imprint of the surface of the object on which they were placed: wood, but also cloth, which makes one think of our ἀπόδεσμος σεσαβανωμένος. Various written indications have been found on these lead seals, as well as imperial portraits: Alexandrian regnal year, a mention of the fisc(us) alex(andrinus),43 or a numeral that Robert Turcan interprets as a dispatch number. Our ostracon 1 shows that the official mail internal to the province of Egypt could also be sealed with a lead seal. We know that the bundle sealed with lead contained a letter of the prefect of Egypt, thanks to information provided orally by the (special?) bearer who brought it to the prefecture in Berenike: this information (as well as the name of its author) was regarded as sufficiently important to be included in the dispatch note. In P.Ryl. II 78.18–19, an apodesmos contains also a letter of the praefectus Aegypti. This information would not have enabled the visual identification of the apodesmos, but it would tell those responsible for postal transportation that this was an object to be handled with special care. When there is an indication in P.Worp 51 that the cavalryman Herakleides had held up the transmission of the mail for several hours because he was in bed with a woman, the writer who is noting the violation does not fail to specifiy that it was letters of the prefect of Egypt: this was an aggravating circumstance. Some dispatches were thus “registered” in various ways, either because they were described as “priority” in the cover notes or because the presence of letters from the prefect of Egypt was noted, orally if necessary. These important dispatches were not, however, given any special treatment or accelerated. What then was the use of their “registration”? Was it simply to stimulate the curators’ zeal? It is not impossible that a postal bag that did not contain any registered mail might have moved less quickly. One might imagine, for example, that the cavalrymen serving the mails might not have been expected to carry out his mission at night if the sack contained only ordinary mail.

VII. Senders and recipients A new prefect of Berenike and a procurator of Berenike The authors of cover notes are the praefectus Montis Volussius Vindicianus and two decurions, Lucianus and [Did]ymianus. 41. The correspondent of P.Oxy. VII 1070, who is in Alexandria, must have sent his letter, addressed to his wife, in the ἀπόδεσμος τῶν ἐπιστολῶν, which doubtless contained letters for other members of the family, as well as, he says, copies of two petitions against his parents-in-law. 42. Vandorpe 1996: 285 (nos. 302–307). Except for the Frœhner seal, she is relying on the catalogue of Turcan 1987: 17–19 and pl. I. For the iconography, see also the drawing of the seal CIL XV 7974a in Frœhner 1890: 236 f.; on this object, cf. Alpers 1995: 283 and n. 978. 43. CIL XV 7974a.

278

Rome in Egypt’s Eastern Desert

Volussius Vindicianus is not otherwise attested. In Egypt, the gentile name Volus(s)ius is rather rare: apart from the prefect of Egypt in office at the time, L. Volusius Maecianus, we know only of Volussius Sabinianus son of a G. Volussius […] in BGU III 709.4 (138–161) and M. Volusius Rufus, associated with T. Volusius […] in P.Marmar. r° viii.37–38 (191). Volussius Sabinianus was a veteran who acquired several parcels of katoikic land in the Fayyum, in the month of Phaophi of an unknown regnal year of Antoninus Pius. He had himself represented in the transaction by another veteran, Marcus Antistius Capitolinus,44 who is known from three other documents of the reign of Marcus Aurelius. Was the prefect of the Desert Vindicianus then a relative of the prefect of Egypt Maecianus? Rudolf Haensch leads me to be reserved about this identification, and not to take lightly the question of the spelling of the gentile name, with single or double s, despite Turbo’s erratic spelling. He observes that the gentile name of the prefect Maecianus is always written Volusius in papyri and inscriptions, while the use of the double s is characteristic of North Africa.45 Egnatuleius Gallus, centurion of the legio II Traiana Fortis, figures in three cover notes, in each case as addressee: of a sealed letter sent from Koptos (1, i.7–8), of an “untied” letter sent him by a colleague (1, iii.41–42), and finally in 5.12, where the context is lacunose. Legionary centurions are rare in the ostraca of the Desert of Berenike: this one is the third to be attested, with the other two appearing in O.Did. 49 (88–96) and O.Krok. I 41.7 (109). Since O.Did. 49 is a pass delivered by the centurion (of the XXII Legion) to donkey drivers for a trip to Koptos, it is likely that this centurion was stationed at Berenike, as is also doubtless the situation of Egnatuleius Gallus. The centurion of O.Krok. I 41, who seems to be the author of a circular addressed to the curators of the road to Myos Hormos, could well be posted in that port. Egnatuleius is otherwise attested in Egypt only as the gentilicium of the epistrategos of the Thebaid L. Egnatuleius Sabinus, known from his cursus inscription ILS 1409, found at Thysdrus.46 Jérôme France prefers to set the terminus post quem for this cursus in the reign of Hadrian, and not, like Pflaum, in that of Commodus.47 There might thus be a connection between the name of the procurator and that of the centurion. Another addressee: Aelius Gemellus (1, i.6 and iii.33–34). On Mecheir 14, a letter sent to him leaves Koptos. But on Mecheir 20, a letter addressed to the decurion Lucianus which, if I understand the text correctly, is a joint composition of the prefect Vindicianus and of Aelius Gemellus, leaves Koptos (where Gemellus would thus have returned in the meantime). Gemellus is “procurator of Berenike” (ἐπίτροπος Βερενίκης, procurator Berenices). This title is new. It appears in another ostracon from Xeron, still unpublished, but later, since it belongs to a series of accounts found in an archaeological context of the third century.48 It happens that the prefect of Berenike, whose position was procuratorial, could be called not ἔπαρχος but ἐπίτροπος. In this case, ἐπίτροπος is made specific with Σεβαστοῦ or, once, Ὄρους (procurator Augusti, procurator Montis), never with Βερενίκης by itself. This person sometimes bears the title praefectus Berenicidis/ἔπαρχος Βερενίκης, but never after AD 72. In just two cases does the titulature of a prefect of Berenike combine the terms praefectus and procurator: Licinnius Licinnianus, addressee of a petition, is called ἐπίτροπος Σεβαστοῦ καὶ ἔπαρχος ἴλης Βουκουντίων (and here ἔπαρχος refers to his military command and not to his territorial prefecture, as the prefects of the Desert of Berenike were often at the same time prefects of the cavalry ala stationed at Koptos);49 44. Nachtergael 2002: 252–54. It is more difficult to imagine a scenario that would connect this Volussius Sabinianus to the prefect Maecianus. 45. Email of 2/4/2018. 46. Pflaum 1960, no. 217; Thomas 1982: 186. 47. France 2001: 162–65. 48. O.Xer. inv. 727. 49. O.Dios inv. 1460 (unpublished).

The postal register of Turbo

279

the other example is Artorius Priscillus, who in contrast did not have a military command and who seems once to be referred to as ἔπαρχος Ὄρους κ̣α̣ὶ̣ ἐπ̣[ίτ]ρ̣ο̣π̣ο̣ς̣ Ὄ̣ρους Βερενίκης.50 But in the present case, the prefect Volussius Vindicianus is accompanied by a procurator Berenices. Lacking an honorific epithet, unlike Vindicianus, Aelius Gemellus is likely to have been a freedman procurator, as his imperial nomen would tend to confirm.51 The location of the decurion Lucianus who is the addressee of the letter may perhaps give an indication: it may have been the praesidium of Apollonos Hydreuma. Located on the Berenike road, this fortlet (of which only one wing remains today) was larger than the other praesidia, which would explain the presence of a decurion in command of it, rather than a simple curator praesidii. Its importance could come from its location at the beginning of the road leading to the emerald mines of Sikayt and Nugrus.52 The procurator Berenices was perhaps also responsible for the metalla of the Desert of Berenike.

VIII. The documents 1 US 50904, 60514 Figs. 53–55

inv. 618 + 1015 14 × 15.5 cm

February 161 alluvial clay

The text was written in three columns on the belly of an AE3 amphora. Two of these columns (i and iii) are relatively well preserved, with each containing a copy of a diploma that accompanied the dispatch of official mail from Koptos, followed by the note in which the curator Turbo acknowledges receipt of this and declares that he has redispatched it. Of column ii only the starts of lines at the level of ll. 6–9 of col. i remain, and the ends of lines at the level of lines 55–63 of col. iii. I have not thought it useful to reproduce these isolated letters. The two dispatch cover notes were issued by the prefect of the Desert Volussius Vindicianus. That in col. i dates to Mecheir 14, that in col. iii to Mecheir 20 (February). We are in the 24th regnal year of Antoninus Pius, who was to die on 7 March. Column i (Fig. 53) The dispatch cover note gives notice of the sending of an unknown number of sealed letters, two of which are identified by their addressee: the procurator of Berenike Aelius Gemellus and the centurion Egnatuleius Gallus. The date at which this mail reached Xeron is in a lacuna. The end of lines 6–11 appears on sherd inv. 1015; this crucial connection, which I had missed in the field, appeared to me only at the moment of publication.53

4

margin ⟦⟧ [ Οὐολούσσιος Οὐινδικιανὸς (ἑκατοντάρχαις) [(δεκαδάρχαις?) κου-] ράτορσι πραισιδίου ὁδοῦ Βερν[εί-] κης χα(ίρειν). ἐ̣πιστολὰς ἐσφραγ[ισμέ-] [ν]ας πενπονένας μί[αν ]δ̣[]

50. O.Krok. I 41.47n. 51. At this period, he could be a freedman of Hadrian or of Antoninus. 52. Suggestion of Sidebotham, Hense, and Nouwens 2008: 299–301. 53. I thank Florent Jacques for the photographic assembly of this join.

280

Rome in Egypt’s Eastern Desert

8

12

16

20

[Α]ἰλίῳ Γεμέ̣λῳ ἐπιτρόπῳ Βερνεί[κ]ης, ἑτέραν Γνατουληίῳ Γάλλῳ (ἑκατοντάρχῃ) λογιῶνος ⸌φροντίσατ̣ε κταχς⸍ καὶ ἐγὼ ἀπέλυσα [ἀπ]ὸ Κόπτου τῇ ι̅δ̅ διὰ Ἰσιδώρου [Ἀ]ν̣μωνιανοῦ τύρμης Σαλβι[αν]οῦ μηνὸς Μεχειρ ὥρ(αν) η̅. Τού̣[ρ-] [βω]ν κουράτωρ πραισ̣[ιδίου Ξηροῦ] [πα]ρ̣έλαβον τὰς ἐπισ[τολὰς καὶ] [τὰ] διπλώματα κα[θὼς πρό-] κ̣ε̣ι̣ται παρὰ Γείγαντο̣[ς καὶ c. 3] δ̣ο̣ς̣ μονομαχῶν Μ̣[εχειρ  ὥρ(αν)] θ̅ νυκτερινὴν κ̣[αὶ εὐθέως] ἀπέλυσα τῇ αὐτῇ [ὥρᾳ διὰ] Ναρκίσσου καὶ Α[ c. 8 ] μονομαχῶν. vacat

2, 8  3 l. πραισιδίων 3, 4 l. Βερενίκ- 4 χα 5 l. πεμπομ- 6 l. Γεμέλλῳ 7 l. Ἐγνατουληίῳ, ω post corr. 8 l. λεγιῶνος 10 l. [Ἀ]μμων- 11  15 l. Γίγ- 16 -δος post corr. “Volussius Vindicianus to the centurions, [decurions] (?), curators of the forts along the Berenike road, greetings. Sealed letters: one for Aelius Gemellus, procurator of Berenike, another for Egnatuleius Gallus, legionary centurion. See to ⟨transmitting⟩ them rapidly. I had them sent from Koptos in the care of Isidoros son of Ammonianos, of the turma of Salvianus, on the 14th of the month of Mecheir (8 February [jul.]), at the 8th hour. “Turbo, curator of the fort of [Xeron], I have received the letters and the cover notes as specified from the hands of Gigas and of […], monomachai, on [the nth] of Mecheir at the 9th [hour] of the night and I have redispatched them at once at the same [hour in care of] Narkissos and of […], monomachai.” 2.

(δεκαδάρχαις?). The presence of the symbol  in the lacuna is uncertain.

4.

ἐ̣πιστολάς. The epsilon has been corrected from alpha: the scribe may have begun to write ἀπόδεσμον.

5.

μί[αν μέν] (cf. ll. 34–35) is not long enough to fill the end of this line, of which the traces along the top of the joined fragment show that it was longer.

8.

λογιῶνος. Other papyrological attestations of the spelling λογ- are: P.Oslo II 33 v° 7 (29), λογιῶνος; and P.Mich. IX 551.30 (103), λογηῶνος (and not λογεῶνος: the eta is somewhat effaced, but its two legs are clearly visible in the on-line photograph). φροντίσατ̣ε κταχς. The most natural reading would be φροντίσαι̣, but this would be difficult to integrate into the syntax. The tau has the advantage of taking account of the horizontal stroke which crosses the eta of -ληίῳ in the preceding line. Moreover, the imperative φροντίσατε is commonly used in the circulars, where it normally introduces an infinitive (“see

The postal register of Turbo

Figure 53. 1, col. i. © A. Bülow-Jacobsen

281

282

Rome in Egypt’s Eastern Desert

Figure 54. 1, col. iii. © A. Bülow-Jacobsen

The postal register of Turbo

283

Figure 55. 1, col. iii, bottom. © A. Bülow-Jacobsen

to it that …”). In the Eastern Desert, however, it is found only in O.Did. 29.9. What follows is uncertain. It is straightforward to read ταχ, and thus naturally ταχέως (or τάχους, proposed by N. Vanthieghem), but then what to do with the kappa? φροντίσατ̣ε ⟨ἐ⟩κ ταχέ̣ω̣ς?̣ But the expression ἐκ ταχέως is without a parallel (and ἐκ τάχους has a meaning different from “rapidly”); or, supposing the presence of an erased alpha between kappa and tau, φροντίσατ̣ε κα̣τὰ ⟨τά⟩χο̣υ̣ς (l. κατὰ τάχος). The general idea is apparently “see to the immediate transmission (of the two letters mentioned earlier)” (cf. O.Krok. I 47.56: ἐν τάχ⟨ε⟩ι διαπέμψεσθε). These readings suppose the omission of an infinitive (which Turbo perhaps thought he did not have space to write, with the prolongation of the final sigma indicating that the rest had to be restored implicitly). Imagining a genitive noun as object of φροντίσατ̣ε would be another possibility, but I do not see what one might read. καὶ ἐγώ, which is rather unexpected, must in my opinion be in contrast to the subject of φροντίσατ̣ε. Turbo, who had decided to omit, for the sake of shortening, the routine instructions, perhaps changed his mind in order to give some sense to καὶ ἐγώ. 9–11.

The author of the cover note (or Turbo while copying it) mistakenly wrote the day of the month before the name of the messenger, thus dislocating the expression of the date.

10.

[Ἀ]ν̣μωνιανοῦ. What remains of the nu could not belong to a mu. The same error occurs in this patronymic in line 48.

10–11. Σαλβι|[αν]οῦ. It is more likely that Isidoros belonged to the squadron of Sabinus (supra p. 275).

284

Rome in Egypt’s Eastern Desert Column iii (Fig. 54 and Fig. 55)

Starting with line 55, the encroachment of col. ii forced the scribe to push the left edge of the column towards the right. The deterioration of the text makes it impossible to establish with complete certainty the list of objects enumerated in the dispatch cover note of Volussius Vindicianus; here is how I understand it: – urgent letters destined for Berenike (are these some unidentified letters, or is this a heading announcing some items mentioned further on?); – three pots containing roses for the goddess Isis; – two letters addressed to the decurion Lucianus: one, sealed, from the prefect of Egypt Volusius Maecianus, another from the prefect Vindicianus and the procurator Aelius Gemellus; – two open letters, the addressee of which is not mentioned: one sent by the directors of an accounting office in Alexandria, the tabularii Narcissianus and Fortunatus; the other sent by Vindicianus; – a bundle with a lead seal containing, according to the messenger who handed them over to the prefecture of the Desert, a letter from the prefect of Egypt (it is not clear if the intended recipient of the prefect’s letter is anonymous or if it was the centurion Egnatuleius Gallus); – an open letter of the centurion Flavius Flavianus addressed to the centurion Egnatuleius Gallus; – an open letter (sender and addressee unnamed, unless the latter is again Egnatuleius Gallus). 21 24

28

32

36

40

44

margin ]υ λελυμ̣έν̣η[ 1–2 [Οὐολο]ύσσιος Οὐινδικιανὸς (ἑκατοντάρχαις?) [(δεκαδάρχαις)?] κ̣[ο]υράτορσι πραισειδίου ὁδο̣ῦ̣ Βερνείκης χ̣α̣(ίρειν). ἐπιστολὰς ἀνα̣[γ-] καίας πενπομένας ἰς Βερνεί[κην] καὶ βείκους τρεῖς̣ ῥό̣δων Eἴσ̣[ιδος] θ̣ε̣ᾶς μεγίστης καὶ εἰς δὲ Ἀπο̣[ Λουκιανῷ (δεκαδάρχῃ) δύο ἀπλ̣σίου Μ̣[αι-] κ̣ιανοῦ {(ἑκατοντάρχου)} τοῦ λαμπροτάτου ἡ̣[γε-] μ̣έν̣ος̣ ἐσφραγισμένην̣, ἑτέ̣[ραν] [δ]ὲ Οὐινδικιανοῦ τοῦ κρατί[στου] [ἐπ]άρχου ἡμῶν καὶ Αἰλίου Γεν[ [ἐπ]ιτρόπου · δύ̣ο λελυμέναι, μία̣ [μ]ὲν ἀπὸ Ναρκισσιανοῦ καὶ Φορτο̣[υ-] [ν]ά̣του ταβουλαρίου, ἑτέρα δὲ Οὐιν[δι]κ̣ιανοῦ τοῦ ἐπάρχου ἡμῶν· καὶ [ἕτ]ερο̣ν ἀπόδεσμον μολυβῇ ἐσ[φρ]αγισμένον ἣν δέδωκεν []ικίας λέγων ἡγεμονικὴ̣ν [ἐπ]ιστολὴν εἶνα̣ι̣ ἐ̣ν̣τὸς· Ἐκνατ̣ου[λη]ίῳ Γάλλ[ῳ] (ἑκατοντάρχῃ) [παρ]ὰ Φλαυ̣ίου [Φλ]α̣υιανοῦ κολλήγ̣α̣ς̣ [α]ὐτοῦ λε[λυ]μένη μία καὶ ὑπὲ̣ρ̣ το̣υ-

The postal register of Turbo

48

52

56

60

64

285

[ο]υ λελυμένη α̅ ἔπε[μ]ψ̣α [ἀ]π̣ὸ Κόπτου διὰ Ποσειδωνίου [ἱ]ππέος τύρμης Ἀπολιναρίου καὶ Ἰσιδώρου Ἀνμωνια̣νοῦ τύρμης Σαβείνου Μεχειρ κ̅ ὥρ(αν) θ̅ ἡμέρας. [2–4] Τούρβων κουράτωρ [πρ]αισιδίου Ξηροῦ παρέλα[βο]ν τὰς ἐπιστολὰς καὶ τ[ὸ] [δ]ίπλωμα κ̣α̣ὶ β̣ε̣[ί]κ̣[ους Εἴσ]ε̣ιδι̣ ὑπ̣άρχ̣ο̣ν̣[τας τρε]ῖς καιρωο̣ς̣ τούρμη(ς) φ Ἀλεξάνδρου τῇ κ̅β̅ ὥ̣ρ̣(αν) ε̣̅ τῆς ἡμέρα(ς) καὶ εὐθέως ἀπέλυσα τῇ αὐτῇ ὥρᾳ διὰ Μάρωνος ἱπ̣πέος τούρμης Σ̣[αλ]β̣ι̣ανοῦ [κ]αὶ ωρα̣ []κ̣ου [ ]υ̣η̣ϲ ]ϲαν ]ων ] – – – – – – – –

24 l. πραισιδίων 25, 27 l. Βερενίκ- 25 χα 26 l. πεμπ-, εἰς 27 l. βίκ-, Ἴσιδος 29  30 λαμπρο-  30–31 l. ἡγεμόνος 31 ἐσφραγισμένην: ην ex ας corr. 33 l. Γεμ- 36 l. ταβουλαρίων 37 ημνω 39 l. ὅν 41 εκνατου l. Ἐγν- 42  Φλαυ̣ίου: υ post corr. 47 l. ἱππέως 48 l. Ἀμμωνι- 56 τουρμη 58 ημερα 61 l. ἱππέως 62 -βιανου 64 ]υηϲ “Volussius Vindicianus to the centurions (?), [to the decurions (?)] (and) to the curators of the forts of the Berenike road. Sent to Berenike: urgent letters and three pots of roses for Isis, the very great goddess; to Apollonos (?), for Lucianus, decurion, two letters: one sealed, from Volusius Maecianus, vir clarissimus, governor, the other from his Excellency our prefect Vindicianus and from the procurator Aelius Gemellus. Two open letters: one from Narcissianus and Fortunatus, accounting agents, the other from Vindicianus, our prefect. Another bundle, sealed with lead, handed over by Nikias (?), who indicated that it contained a letter of the prefect. An open letter for Egnatuleius, centurion, from Flavius Flavianus, his colleague, and an open letter … I have sent (the enumerated items) from Koptos, in care of Poseidonios, cavalryman of the turma of Apollinaris, and Isidoros son of Ammonios, of the turma of Sabinus, on Mecheir 20 (14 February [jul.]) at the 9th hour of the day. “Turbo, curator of the fort of Xeron, I have received the letters, the cover note, and the pots for Isis, three in number, from the hands of …, of the turma of … Alexandros on the 22nd at the 5th hour of the day, and I have sent them on immediately within the hour in the care of Maron, cavalryman of the turma of Salvianus and of …”

286

Rome in Egypt’s Eastern Desert

21.

This line, written above line level, is in the same hand as what follows, but in a larger size; the pen is finer. Is it a title, λελυμένη being the open letter (the δίπλωμα τῆς ἐπιθέσεως) that follows? Or is it the beginning of an abandoned text?

23.

(ἑκατοντάρχαις?) [(δεκαδάρχαις)?]. Of the first symbol only a part of the chi remains. Was the other letter, now lost, a rho as in line 2 of column i, or an iota as in 5.9? And if the first-named were centurions, did the symbol for decurions appear in the lacuna?

24.

κ̣[ο]υράτορσι. The second bit of ink at the start of the line probably does not belong to the omicron; it is rather the tip of the lower branch of kappa.

25–26. ἀνα̣[γ]|καίας. I have not found another example of ἀναγκαῖος modifying ἐπιστολή, except in Philostratus, Ap. 4.46: τὰς δὲ οὐχ ὑπὲρ μεγάλων ἐπιστολὰς ἐάσαντες τὰς ἀναγκαίας παραθησόμεθα κἀξ ὧν ὑπάρχει κατιδεῖν τι μέγα, “leaving to one side the insignificant letters, we will mention those which are important and which allow us to see something important” (with respect to philosophical correspondence). But in our cover notes, it has a technical meaning: it refers to priority letters (see supra p. 276). This modifier may have made it possible to mark as urgent those letters of which the authors or the recipients were not individuals important enough for their names to have appeared in the note and for mail to or from them to have been given priority automatically. 27.

βείκους. The term βῖκος can refer to vessels of very diverse shapes and sizes.54 It must have referred here to containers with a large opening that allowed cut flowers that still had blossoms to be kept fresh (no doubt they were moistened at each stage).

27–28. ῥό̣δων Eἴσ̣[ιδος]| θ̣ε̣ᾶς μεγίστης. Presumably, these “roses of Isis” (if the genitive is not an error for the dative, which is used by Turbo in lines 54–55) were meant for the temple of Isis Great Goddess, built under Tiberius and which was the main temple in Berenike (Ast 2020). The flowers were perhaps intended for the festival saluting the return of spring, which is known to have been celebrated in Mecheir. At Soknopaiou Nesos, the rhodophoria went on from Mecheir 12 to 24 (SPP XXII 183.76 [(post) 138], P.Louvre 4.56n.).55 29.

ἀπλ̣σίου. The writer became muddled, and the traces, some of which seem to be hanging, cannot be interpreted with confidence; this is curious, because the same nomen (despite the question of the doubled s, discussed on p. 278) is written without difficulty in the case of the prefect of the Desert. However that may be, the intention was ἀπὸ Οὐολουσίου.

30.

λαμπροτάτου. Rho is written over pi, forming a monogram found also in πρόκειται in 3.7. This epithet (vir clarissimus) was in use for the prefect of Egypt between 145 and 267.56

54. Bonati 2016: 27–31. 55. Cf. Youtie 1951: 193. Youtie tacitly identified this festival with that of the “germination of the divine plants” celebrated at Edfu between the 22nd and the 30th of Mecheir. A letter found at Berenike, O.Berenike II 195 (50–75), with commentary in O.Berenike III, p. 14, evokes a large purchase of roses for 100 drachmas, not including the cost of transportation. 56. Arjava 1991: 19 f., nn. 11 and 13.

The postal register of Turbo

287

Figure 56. 2. © A. Bülow-Jacobsen

33

Γεν[. It would be very unlikely that this referred to a procurator other than Aelius Gemellus, even if the tiny bit remaining of the letter that follows nu does not naturally make one think of epsilon.

38.

[ἕτ]ερο̣ν supposes in principle that another ἀπόδεσμον has already been mentioned, which is not the case. μολυβῇ. σφραγῖδι is understood or forgotten: cf. O.Krok. I 17 (σφρ̣[αγ]ε̣ῖδι βολυμ̣[ῇ). The day slip O.Did. 23 also mentions an indeterminate postal item sealed with lead.

40.

[Ν]ε̣ικίας?

288

Rome in Egypt’s Eastern Desert

44–45. ὑπὲ̣ρ̣ το̣υ|[ο]υ. ὑπὲ̣ρ̣ το̣ῦ |[ο]υ is tempting, but what then is to be made of the following descending vertical stroke? Should one consider ὑπὲρ Τα̣υρ̣ί̣|[νο]υ? το̣ῦ φ̣ί̣|[λο]υ? 48–49. Ἀνμωνιανοῦ. Cf. Ἀνμων̣[ίου in P.Hamb. IV 270.3 and the quarry name Νειλάνμων in O.Claud. IV 734–738. 51.

[2–4] Τούρβων. It is hard to say if the width of the lacuna should be estimated based on the preceding or the following line. There is visibly something before Τούρβων. A gentilicium?

54–56. I am not entirely sure how the beginnings and ends of these three lines match up. 57.

Φλ̣α̣ο̣υ̣ί̣ο̣υ̣ is proposed with all reserve.

55.

The two letters visible before δι belong to column ii.

56.

ἱππέος is the word expected before τούρμης and is not an impossible reading. But there then remains little space for the cavalryman’s name.

2 (Fig. 56) US 80803, 80804

inv. 257 14 × 15.5 cm

161 alluvial clay

In contrast to the presentation of 1, there are two diplomata, separated by a line, in the same column. Moreover, these may have been copied by different scribes: if the first hand is indeed that of the rest of the dossier, the second, as far as this tiny sampling allows us to judge, seems to me to present strokes significantly different for certain letters (ω, β) and particularly for υ, systematically written in the form y. → 4

8

12

16 m.2

– – – – – – – – – ][ [τῷ λαμπ]ρ̣[ο]τ̣άτῳ ἡ̣γ̣ε̣[μόνι·] [ἀπόδε]σμος εἷς δεδερμα[τωμ]ένος τῷ κρατίστῳ ἐ[π]ά̣ρχῳ ἡμῶν Οὐολυσσίῳ Οὐινδικιανῷ· ὁμοίως δύο ἐπιστολαὶ ἄλλαι ἐσφραγισμέναι τῷ κρατίστῳ ἐπάρχῳ, ἀπό̣δεσμος εἷς σεσαβανωμ̣[ένος·] ἀπ̣[έ]λυσα ἀπὸ Βερνείκ̣[ης.] Τούρβων κουράτ[ωρ 5–6 ] τὸ δίπλωμα καὶ τ[ 7–8 ] καθὼς πρόκειται [ c. 8 ] [ἐ]ν̣ε̣σ̣τ̣ῶτος μηνὸς Φ̣[ [?]ω τῆς ἡμέρα[ς [][]εικου̣ϲουκ[][ –––––––––––––––––––[ [2–3]2–3 ο̣ς Λουκειαν̣ὸς (δεκαδάρχ-) [

The postal register of Turbo

20

24

10 l. Βερενίκ[ης

289

[πραισι]δ̣ί̣ων̣ ὡδοῦ Βε[ρε]νε[ίκης [ἀπόδεσ]μ̣ο̣ν ἐπιστ̣[ο]λ̣ῶ̣[ν ] τοῦ κρατίσ[του ]ειαϲτρω[ Ἀπόλλ]ω̣νος ῾Υδρ̣ε̣υ̣[μα-δίπλω]μα τοῦ αὐ[τοῦ ἐπισ]τ̣ωλὴν το[ῦ ]κ[]τι[ ][ – – – – – – – 17 l. Λουκιανός 

18 l. ὁδοῦ Βε[ρε]ν[ίκης

24 l. ἐπισ]τολήν

“ … for the governor, v.c.; a bundle wrapped in leather for his Excellency our prefect Volussius Vindicianus. Likewise, two other sealed letters for his Excellency the prefect; a bundle wrapped in linen. I have dispatched (the enumerated items) from Berenike. Turbo, curator, [I have received] the cover note and […] as indicated above, on the [nth] of the current month of Ph[amenoth) at the – hour and I have redispatched (?)] them the same day [hour, identity of messengers.] “[Name] Lucianus, decurion (or to the decurions), [to the curators] of the praesidia of the Berenike road, [greetings.] A bundle of letters …” 3–4.

δεδερμα[τωμ]ένος. S.v. δερματόω, LSJ indicates, for the passive, the meaning “be turned into a hide.” Better translations are offered by the Th.Gr.L. (pelliculo), and the DGE s.v. δερματόομαι, formar piel, a translation suitable for rendering the passage of Galen cited, in which the subject is σάρξ. This verb is not attested in the active voice, where it would have the general sense of “cover with skin or leather.”

4.

κράτιστος (vir egregius) becomes the epithet for the rank of equestrian procurator under Marcus Aurelius, but it is found before then, in informal usage, for the prefect of Egypt and the equestrian and freedman procurators.

9.

σεσαβανωμ̣[ένος. The verb σαβανόω is derived from τὸ σάβανον, a borrowing from the Egyptian sbn (linen strips used to wrap mummies). Τὸ σάβανον in the papyrus refers to a piece of linen for various uses (napkin, scarf …). In P.Kellis I 72.34, the diminutive σαβάνιον refers to a linen fabric used to protect threads or bits of wool dyed purple, during transport. Some of the manuscripts from Qumran were wrapped in rectangles of linen with leather straps.57 As to the verb, it was known before now only in very late texts: in Digenis Akritas, an epic poem of the twelfth century, in the demotic language, and in the translation into Greek, made in the fourteenth century in the Peloponnesos, of the Romance of Troy, a poem in Romance by Benoît de Sainte Maure (a Norman, unless from the region of Tours, poet of the twelfth century). In each case it means “put (a cadaver) into a coffin.” It is not certain that the bundle wrapped in linen was intended for Volussius Vindicianus. It is also possible that, as in the postal register O.Krok. I 1 (c. 107), no one took the trouble to specify the recipients of lower status. It is therefore not to be excluded that the choice of leather or linen to wrap the mail may have been an indicator of the importance of the recipients. 57. Bélis 2003.

290

Rome in Egypt’s Eastern Desert

Figure 57. 3. © A. Bülow-Jacobsen

10.

This is the end of the cover note. This version of the postal register is obviously abridged, because ἀπέλυσα (var. ἔπεμψα, cf. 1, ii.45) then normally introduces the day and hour of departure as well as the identity of the messenger(s).

11.

This is the only place to restore ἔλαβον, used in O.Xer. inv. 574 (rather than παρέλαβον, which is also found at Xeron in this context, but which would be too long). The name of the praesidium is thus not indicated as it is elsewhere.

13.

τ̣[ῇ 1–2 τοῦ]?

15.

We expect here the end of the verb ἀπέλυσα, but this reading does not fit the traces.

16.

The identification of the messenger is expected here.

17.

(δεκαδάρχης) or (δεκαδάρχαις). Cf. 6.6.

19.

ἀπόδεσμος ἐπιστολῶν: P.Ryl. II 78 (157).

21.

χ]ρ̣είας τρωχ̣[, l. τροχ[οῦ? Mentions of τροχοί, water wheels, are rare and always doubtful in the ostraca of the desert (O.Did. 61.4; O.KaLa. inv. 619).

The postal register of Turbo 3 (Fig. 57) US 80807

inv. 279 14 × 14 cm

291 25 February–26 March 161 wall of a Dressel 2/4 amphora

The ostracon is so faded that the remains of ink are barely visible to the naked eye. It contains the remains of two columns. Of that at the right only the beginnings of thirteen lines remain, with two to four letters, often doubtfully read. Of the left column, which is edited here, the lower right corner remains, corresponding to the acknowledgment of receipt by Turbo.

4

8

12

5 l. πραισιδίου

– – – – – – – – – – – traces of 3 lines ] Τούρ[βων κουράτ]ω̣ρ π̣ρ̣α̣ι̣σ̣ε̣ι̣δ̣ί̣ο̣υ̣ [Ξηροῦ] π̣α̣ρ̣έ̣λ̣α̣β̣α τὸ̣ δίπλωμα [καὶ τὰς ἐπισ]τολὰς ὡς πρ(ό)κειται τῇ [1–2] [τοῦ μηνὸς] Φαμενωθ ὥρᾳ θ̅ τῆς νυκτ̣ὸ̣ς̣ [ἀπὸ --] ἱ̣π̣πέος τύρμη̣ς̣ ] Ἀλεξάνδρου καὶ Σερήνου ἱππ]έ̣ο̣ς̣ τύρμης Σει̣τ̣α καὶ εὐ̣[θέως ἀπέ]λ̣υ̣σ̣α τ[ῇ αὐ]τῇ ὥρᾳ διὰ [Μάρωνος?] ἱππέος τύρμης Σαλβιανοῦ. vacat 6 διπλωμα

8 της 9, 11, 13 l. ἱππέως

6.

τὰ̣ διπλώμα(τα) cannot be excluded.

7.

πρ(ό)κειται. The same monogram pi-rho appears in λαμπροτάτου in 1.30, but here the monogram represents the entire syllable προ. In P.Oxy. XL 2915n., John Rea collects examples of this monogram representing the verbal prefix προσ-.

11.

Σει̣τ̣α. One might envisage ρ or υ in place of ι and υ in place of τ.

13.

Μάρωνος is restored on the basis of 1, ii.61.

292

Rome in Egypt’s Eastern Desert

Figure 58. 4. © A. Bülow-Jacobsen 4 (Fig. 58) US 60509

inv. 1030 5.6 × 6.8 cm

25 February–26 March 161 alluvial clay

A curator’s note of redispatch.

4

8

– – – – – – ][ ]υ [ Τούρ]βων κουράτω̣[ρ πρό]κ̣ειται ἀπὸ Σ̣τ̣ρ̣[ μον]ομάχου ὥρ(αν) [ ἀ]πέλυσα διὰ [ ] Φαμενω[θ ] vacat [ ][ – – – – –

5 5 (Fig. 59) US 50507, 50510

inv. 1241 9 × 11 cm

27 March–25 April 161 alluvial clay

The fragment preserves the end of a cover note of mail originating in Berenike (lines 1–5), the curator’s note (lines 5–8) and the beginning of a second cover note originating in the office of the prefect of the Desert (lines 9–15).

The postal register of Turbo

Figure 59. 5. © A. Bülow-Jacobsen

4

8

12

4 l. Βερεν-

– – – – – – – – – – – ]σαι̣⸌⸍και[ ]μενος κ[ ἐσ]φ̣ρ̣[αγισ]μένη τῷ κρ̣[ατίστῳ ] ἀπέλ̣[υσ]α ἀπὸ Βερν[είκης Φ]α̣ρμουθ̣[ι.] Τούρβων κ[ουράτωρ παρέλ]αβα τὰς ἐπιστολ(ὰς) καθὼ[ς ]ων ὥρ(αν) ϛ̅ νυκτὸς καὶ ε̣[ὐθέως ] Γλαφυρείνου̣ μ̣ονομ(άχ-) vac. [ [Οὐολούσσι]ο̣ς Οὐινδικιανὸς (δεκαδάρχαις) καὶ [ ] χα(ίρειν). ἐπιστολὰς ἀνα[γκαίας ἐ]⟦πάρχῳ ἡμῶν⟧[ Ἐγνατου]λ̣ηίῳ Γάλ[λῳ Τραι]ανῆς Ἰσ̣[χυρᾶς ]α̣πο[ ][ – – – – – – – – – – – – – 6 επιστολ 7 

8 μονομ

293

294

Rome in Egypt’s Eastern Desert

Figure 60. 6. © A. Bülow-Jacobsen

7.

]ων. Perhaps παρά τινος καί τινος μονομαχ]ῶν. In this case, the number of the day would come directly after καθὼ[ς πρόκειται.

6 (Fig. 60) US 50618

inv. 46 14.5 × 7.5 cm

3 April 161 alluvial clay

Lines 1–5 belong to a notice of receipt by Turbo, while lines 6–7 are the beginning of a new diploma.

4

3 μονομ

4 ημερινην

– – – – – – – – – – ] ἔλαβα τ̣[ ] καθὼς πρόκειται vac. ([) ]α̣κος καὶ Χένου μ̣ο̣νομ(αχῶν) ([) Φαρ]μουθι η̅ ὥρ(αν) ι̅α̅ ἡμερινήν vac. ([) ὥρᾳ] τ̣ῇ̣ αὐτῇ ἀπέλυσα διὰ Κρηνείτ(ου) ([) ][]υμιαν(ὸς) (δεκαδάρχ-) κουράτορ(σι?) ([) ] Β̣ε̣ρ̣νεί(κης) χα(ίρειν). α([) – – – – – – – – – – – – – 5 κρηνειτ 6 ]υμιαν  7 l. Βερενί- χα

The postal register of Turbo

295

1.

ἔλαβα. The traces do not support παρέλαβα.

3.

]α̣κος καὶ Χένου. These names are new in the prosopography of the monomachai of the desert. Χένος is found neither as a personal name nor as a dictionary entry (except in the lexicographers and grammarians who present it as derived from χέω and an orthographic variant of κενός (παρὰ τὸ χέω χενὸς καὶ κενός). Would this then be a variant of Χέννος, the nickname of the Alexandrian grammarian Ptolemaios Chennos, who lived in the time of Trajan and Hadrian?58 This nickname could have been formed from a noun *χέννος, of which τὸ χέννιον, which is in contrast well attested, would be the diminutive form. From *χέννος could also be drawn the personal names Χεννᾶς (O.Berenike I 75, cf. O.Berenike I, p. 27) and the feminine Χ̣εν̣οῦς (SB XX 14116.3). According to LSJ, χέννιον is a borrowing from the Egyptian chennu (more precisely, ḫnn.t, written also ḫnnw, derived from ḫn, “to alight”) and meaning “quail.” This identification comes from Athenaeus, particularly 9.48.40 Kaibel = 9.393c Casaubon: τῶν δὲ καλουμένων ΧΕΝΝΙΩΝ (μικρὸν δ’ ἐστὶν ὀρτύγιον) μνημονεύει Κλεομένης ἐν τῇ πρὸς Ἀλέξανδρον ἐπιστολῇ γράφων οὕτως· “φαληρίδας ταριχηρὰς μυρίας, τυλάδας πεντακισχιλίας, χέννια ταριχηρὰ μύρια,” “Cleomenes mentions the ‘chennia’ – the term refers to a small quail – in his letter to Alexander: “ten thousand salted coots, five thousand thrushes, ten thousand salted quails.” According to the evidence of Athenaeus and the lexicographers, but also Herodotus (2.77, cited infra), the Egyptians consumed quail salted. Athenaeus in the same place quotes some verses of Hipparchus, a dramatist of New Comedy, in which the poet makes fun of the Egyptians who spend their lives plucking χέννια. The Egyptian origin of the word thus seems to be inescapable; but it does encounter two problems: first, ḫnn.t is rare after the New Kingdom and is not at all attested in demotic;59 second, ḫnn.t, ḫnnw is a generic term for “bird”;60 in Egyptian, quail is called p‘rt, p‘r in demotic. More precisely, the ḫnn.t are birds that are hunted and eaten, as opposed to birds of prey, as Maurice Alliot has clearly explained: “The Egyptians divided birds into two categories: pȝyt nbt ḫnnt nbt meant “all the birds.” Those of the first category are not only those who fly, but those who fly high, and who spend most of their existence flying (raptors, for example); those of the second category are not only those who “alight,” but those who fly low, walking or swimming (water birds or barnyard birds, for example, wild or domestic “fowl”).”61 How would the generic ḫnn.t have come to mean “quail” when borrowed by Greek? The reply is perhaps in Herodotus (2.77): ὀρνίθων δὲ τούς τε ὄρτυγας καὶ τὰς νήσσας καὶ τὰ σμικρὰ τῶν ὀρνιθίων ὠμὰ σιτέονται προταριχεύσαντες, “as to birds, (the Egyptians) eat quails, ducks, and small birds raw and salted beforehand.” *χέννος could refer to any small bird that was prepared in this type of salting (ḫnn.t would have been specialized in this meaning in the Late Period?), quail being the most emblematic, hence the slippage toward the meaning of “quail” to which the literary sources testify. But what do the papyri indicate? Attestations of χέννια are not very numerous, but extend from the third century BC down to the Byzantine period. Two papyri from the archive of Zenon show that a distinction was made between a pot of chennia and one of quails.62 If in the papyri the ὄρτυγες (quails) are sometimes reckoned in units or 58. Suda, π 3037. 59. Marie-Pierre Chaufray and François Herbin, per litt. 60. Fournet 1989: 74. 61. Alliot 1946: 72, n. 3. I thank Fr. Herbin for this reference. 62. P.Iand. Zen. 53.12; 55; 81 and PSI VII 862.10–11: ὀρτύγων βανωτὸν α, χεννίων βανωτὸν α.

296

Rome in Egypt’s Eastern Desert

Figure 61. 7. © A. Bülow-Jacobsen

Figure 62. 8. © A. Bülow-Jacobsen

The postal register of Turbo

297

even mentioned as living animals to be fed, the chennia form an indistinct mass, almost always reckoned in pots. The constant diminutive form χέννιον might be explained by the fact that these birds are referred to not so much as birds but as a food product.63 If the link between the personal name Χένος and the Egyptian χέννιον is correct, our monomaches “Birdie” would be particularly thin, unless the opposite, “beefy,” is meant jokingly. 5.

Κρηνείτ(ου). The name of this monomaches is also attested in O.Xer. inv. 847, 1144, and 1199, which do not belong to the dossier of Turbo. It is homonymous with two ethnic adjectives: according to Artemidoros in Herodian, De prosod. 3.1 (ed. A.  Lentz 1867, p. 188, l.19), the Κρηνῖται were the inhabitants of the city of Κρηνίδες in Macedonia, renamed Φίλιπποι by Philip after he had assisted it against the Thracians. It can also be the hellenized form of an Armenian ethnic that became a family name in the Byzantine period; the form Κρινίτης is then found, which is also present in O.Xer. inv. 1199.64

6.

][]υμιαν(ός). The only personal name found in Egypt at this period that fits this ending is Διδυμιανός, which is also borne by soldiers, for example the decurion Aponius Didymianus, who addresses the short message O.Florida 5 (second century) to the curator Iulianus. At Xeron itself, an amphora lid with a dipinto bearing the name of the decurion Ulpius Didymianus was found (inv. 1212). The first stroke of the line is too high to belong to this name and should belong to a descending letter from the previous word. (δεκαδαρχ-). It is not clear if this refers to the rank of Didymianus or to the addressees of this circular; cf. the similar uncertainty in 2.17.

7.

α([). Perhaps the beginning of ἀπόδεσμον (a suggestion of N. Vanthieghem).

7 (Fig. 61) US 50618

inv. 106 10.5 × 7.5 cm

161 alluvial clay

Lines 4–6 belong to Turbo’s note of receipt, while line 7 is the start of a new diploma. Turbo’s hand becomes erratic, which complicates reading.

4

8

– – – – – ] ]α̣ ] τ̣ῇ κ̣̅θ̣̅ ]α πραισιδίο̣υ̣  καθὼ]ς πρόκειται τῇ α̅ ] ἀπέλ̣υ̣σ̣α τῇ αὐτῇ ] vacat ] κουρά̣τ̣ο̣ρ̣σ̣ι̣ [ ][ – – – –

63. Cf. S. Amigues, comm. ad Theophr. Hist. pl. 7,1, CUF, 79, citing notably the opposition between ἧπαρ, which belongs to the anatomical vocabulary, and ἡπάτιον, a culinary term. 64. Porph. Const., de administrando imperio, 165, comm. ad 43/137.

298

Rome in Egypt’s Eastern Desert

2.

At the end of the line, we appear to have a monogram (φρ?), unless there has been a deletion.

4.

I do not know how to explain the sinusoidal stroke after πραισιδίου. It rises too high to represent Ξ(ηροῦ).

8 (Fig. 62) US 40844

inv. 58 6.6 × 8 cm

161 alluvial clay

Fragment of a cover note. Several words appear that we do not find in the other fragments.

4

8

2 σαγ

4 επιστολ

– – – – – – ]α̣[ ]σαγ̣( ) τῷ κρατίσ̣[τῳ ] ἡμῶν Οὐολο̣[υσσίῳ ]⸌ν̣⸍ γ̅ ἐπιστολ(αὶ) τ̣[ ] Ἀλεξανδ( ) Ε̣ὐπ[ ]βντες ἀ]πολύσαντ( ) ἀπὸ π[ ] ὑποση( ) [ ] τ̣ο̣υρ[ ]κ[ – – – – – 5 αλεξανδ

7 ]πολυσαντ

8 υποση

2.

]σαγ̣( ). The hanging letter which resembles a gamma might be a sigma (cf. the sigmas of -καίας in 1, iii.26 and of -μενος 1, iii.31). In this scribe, suspension does not always imply abbreviation, cf. τῆς in 3.8.

6.

]βντες. Interline. ]λ̣α̣βό̣ντες, which comes to mind naturally, does not fit the strokes very well.

7.

πρ̣[αισιδίου?

8.

Mention of a subscription (ὑποσημείωσις)?

9.

Τ̣ο̣ύρ[βων or τ̣ο̣ύρ[μης.

18 Men and gods in a network* In memory of Jean Bingen During the principate, forts controlling quarries or roads started spreading across the Eastern desert zone stretching between the latitudes of the Porphyrites and Berenike. Some of these praesidia have remained practically in the same state as they were when last occupied by a garrison. A great deal of writing took place there, and part of it was left on the spot. However, the garrisons’ archives were taken to the military camps in the Nile valley. What is left are ephemeral documents generated by service as well as the quirks of everyday life. Such texts may only be found in refuse dumps, since they were usually written on ostraca then quickly sent to the scrap heap. Refuse dumps are found in front of the gate of every fort, provided they have not been taken away by floods or razed by bulldozers. Other, later dumps have built up inside the buildings in condemned rooms or cisterns, when smaller or less disciplined garrisons settled there, towards the end of the second century or the beginning of the third.1 As to the kind of texts that may be found on these ostraca,2 there is, first, a great deal of tituli, whether on amphoras or other containers; then, lists of names, whether beneficiaries of rations, accounts, or duty rosters for the guards; tiny notes also, so terse that their meaning remains unclear. There are a lot of private letters aimed at accompanying packages so that those would not be “lost” on the way, and a few school assignments which reveal efforts to learn how to write Latin and Greek. The ostraca which historians are most interested in – as it happens, the rarest ones – are rough versions and copies of official documents: postal journals, requests, and administrative and military letters. It was on Mons Claudianus that I first became aware of the papyrological potential of the Eastern desert. During our last campaign there, Adam Bülow-Jacobsen and I took advantage of the fact that Nicolas Grimal was there on a visit – he was at the time the director of the IFAO – to take him to AlZarqaʾ, which is the best-preserved fort among those that lie on the road from Quft to Qusayr. On the previous year, we had noted that the refuse dump which lay before the gate had recently been visited by illicit excavators. N. Grimal was easily persuaded of the relevance of a rescue excavation. He made a promise that the IFAO would offer logistical support and introduced a request for a grant from the Ministry of Foreign Affairs excavation Committee, whose support has never failed us. I was fortunate * Translated by Guillemette Monge Romero. 1. Chapter 15, p. 245. 2. See Chapter 41, Fig. 140, p. 599 (chart of ostraca types found at Xeron).

299

300

Rome in Egypt’s Eastern Desert

enough to have two leading archaeologists joining the venture, Michel Reddé, a specialist in military architecture, together with Jean-Pierre Brun, who was not deterred by the prospect of having to disentangle the confused stratigraphy of a refuse dump. They quickly realized that it would be necessary to excavate other forts in order to understand better the history of such a network and the way it functioned. This is how the project came about, with the unfailing support of N. Grimal and the IFAO directors who came after him. In 1998 we moved on to the twin road of Koptos to Berenike, not to leave it again, but for a short mission in Umm Balad, a granodiorite quarry tucked into the southern slope of Porphyrites (2002–2003). We shall leave that aside here and strictly deal with that part of the desert stretching between the latitudes of Qusayr and Berenike, which took the name of “Desert of Berenike” (Ὄρος Βερενίκης, Mons Berenices/Berenicidis) at the beginning of the Roman period. Mons Berenicidis was under the authority of a governor, whose title was prefect. Of equestrian status, he could be a unit commander or just a civil administrator, and held sometimes a military post as well as being an imperial procurator: the first person known to have cumulated both tasks is Servius Sulpicius Serenus under Hadrian (Chapter 3). From that time onwards, it seems that the prefect of Berenike always commanded the cavalry wing which was based in Koptos, the ala Vocontiorum until c. 180, then, after that date, the ala Herculiana.3 The curatores praesidiorum, who were the non-commissioned officers in command of the forts, would answer directly to the prefect of Berenike. The “Eastern desert praesidia” program, which began in January 1994, ended in January 2013 with the fourth campaign in Xeron Pelagos. Other archaeologists joined us in between, and readily embraced the game of papyrological search:4 Emmanuel Botte, Bérangère Redon, and Thomas Faucher. The corpus of ostraca is now closed, and we cannot rely on further discoveries to address unanswered questions, except minimal ones coming from better readings or remaining corrections. It is now time to take stock. The trade of the goods which were exchanged between the Mediterranean world and the countries to the north of the Indian Ocean was the raison d’être of those forts, ports, and roads. The praesidia, which were fortified quadrangles built around a well, ensured a military presence in a territory where aggressive Beduins roamed, and also supplied the caravans with water as well as acting as mail relay stations. Although about 5,000 ostraca have been inventoried, they have not taught us anything at all about the Red Sea trade. The world of international trade and the world of the garrisons hardly ever communicate, or if they do, one cannot tell from the ostraca. The main elements of knowledge that we have relate to the history and the running of a network of the Roman army’s advanced posts in a border zone, and with garrison life. They will be illustrated by a few unpublished documents which will enable us to give an insight into the linguistic atmosphere in those forts, where the written vernacular was Greek, but the languages spoken were diverse. What was mostly heard there must have been a pidgin of Greek and Egyptian.

1. The forts and their gods: religious geography and toponymy Our knowledge of the toponymy of the Eastern desert has increased. We now know that the two roads which crossed the Berenike Desert during the principate derived their names from the Red Sea port to which they led from Koptos; the road from Quft to Qusayr was called “Myos Hormos road” (ὁδὸς Μυὸς Ὅρμου, ὁδὸς Μυσορμιτική), while the road which ran from Koptos to Berenike was called “Berenike 3. For a prosopography of Berenike’s prefects, see Chapter 3, pp. 105–9. 4. It goes without saying that we did not limit ourselves to “papyrological search” only – archaeologists would have balked at it! However, Jean-Pierre Brun has always ensured that both archaeologists’ and papyrologists’ interests were reconciled.

Men and gods in a network

301

road” (ὁδὸς Βερενίκης and also Βερνικησία ὁδός, once).5 Their Latin names are not mentioned anywhere. Both roads were only progressively equipped with forts. In piecing together a series of identical, though unequally preserved, dedications, we realized that a building program had been ordained by Egypt’s prefect Iulius Ursus in 76/7.6 However no dedication sufficiently complete has been found on the Myos Hormos road, though the latter is likely to have been equipped at the same time: the architectural typology in Krokodilo and Maximianon is not foreign to the building style of the Flavian period, even though there are not any ostraca prior to Trajan in the praesidia of this road. According to Jean-Pierre Brun, this is the result of Flavian dumps being swept away by floods. It remains true nevertheless that merchants at the time did not travel on the Myos Hormos road so much as on the Berenike road; it seems indeed that Myos Hormos was rather used for ship building and ship repair.7 In the study of the names of the forts, the tutelary deities invoked in the proskynemata in the private letters written there should not be overlooked. Those deities, first of all, were sometimes eponyms for the praesidia which they protected, while proskynemata also reveal the origin of the letters found on a site.8 Yet, private letters – at least those that were written on ostraca, unlike official mail – rarely went beyond the two directly neighboring forts. Indeed, they could only be entrusted either to an official mail rider, someone whom you knew, but who did not go any farther than the next fort, or else to an unknown donkey-driver or camel-driver who would travel longer distances. Another possibility was to write a second letter to someone you were acquainted with in the next fort, asking them to forward the first letter by entrusting it to a reliable person.9 The names of the praesidia on the Berenike road, which, although it was opened after the Myos Hormos road, became better travelled, were already known, even though with deformations, through the Antonine Itinerary, the Peutinger Table, and the Ravenna Cosmography.10 Pliny the Elder, who describes this road before it became equipped with praesidia, knows of but two toponyms, Hydreuma Apollinis and Novum Hydreuma. The ostraca have corroborated the evidence in the Antonine Itinerary and the Ravenna Cosmography as opposed to that found in the Peutinger Table, which confuses Didymoi (for which the ostraca have restored the correct plural masculine form) with Aphrodites. In the case of Xeron, the ostraca enable us to lay aside the form Aristonis of the Antonine Itinerary, which could be a corruption of Aridum, itself a calque for the Greek toponym11 (as in the case of Dios). The correct ending for Φαλακρόν has been restored: the underlying neuter appellative must be ὄρος (which should then be taken here as meaning “mountain” and not “desert” as in Mons Berenices) or else ἄκρον, ἀκρωτήριον, of which we find occurrences in the geographical literature. It is likely that this “Bald Mountain” is the one that can be seen behind the praesidium in Phalakron.12 Given that the ostraca from a praesidium rarely offer glimpses of sites other than a couple of its closest neighbors, the knowledge we have of the toponymy of the praesidia south of Phalakron, none of which has been investigated, is more fragile. One ostracon found in Xeron in 2013, unfortunately quite damaged, probably bears the name of the fort which comes under the form Cabalsi/Cabau/Cabaum in 5. O.Dios inv. 457 (219). On road names, see Chapter 1, pp. 308 f. 6. Bagnall, Bülow-Jacobsen, and Cuvigny 2001. 7. Bülow-Jacobsen 2013: 567. 8. See Table 18.1 on p. 304. 9. O.Did., p. 19. 10. Table 18.2, pp. 321 f. 11. A hypothesis of A. Bülow-Jacobsen, to which I subscribe. 12. See Fig. 28, p. 51.

302

Rome in Egypt’s Eastern Desert

the Itineraries.13 What we have is a narrative poem in the first-person singular and first-person plural of a journey from Koptos to Berenike. Each halt is described by a sentence characterizing, as in a vignette, the water resources which the narrator has found; vignettes are separated by oblique strokes. One hexameter thus relates the halt in Phalakron: ἤλθαμεν εἰς Φαλακρὸν κοὐχ εὗρον ὕδωρ ἀγοράσαι, “We arrived in Phalakron and I could not find any water to buy.” The following verse, which is ill-preserved, is probably about the following halt, Apollonos (the epithet Πυθίου may be read), then, after an oblique stroke, we read (Fig. 63): ἤλθαμεν εἰς Καυ

Figure 63. O.Xer. inv. 995, fr. C, end of line 14. © H. Cuvigny

We cannot be certain of the total amount of unread characters because the word is written at the end of a line and because the scribe has pressed the letters against the right-hand edge of the sherd. Καυα̣λ⸌̣ σ̣(ι)⸍ could be, but does not have to be, one of the possible readings derived from the Itineraries. The following stage of this imaginary journey is Kainon Hydreuma,14 just before Berenike: / Καινὸν Ὕδ(ρευμα) ἰδὼν []ον πλατὺ Παλλάδος ὕδωρ / The “water of Pallas” mentioned here reminds us of a similar poem written in praise of the Xeron well in which Pallas is both described as a queen reigning on the admirable soil of Xeron and as a water purveyor.15 It is worth noting that Athena was the tutelary deity of Xeron: could she have been that of Kainon Hydreuma as well? At any rate, she was the genius loci at Persou, on the Myos Hormos road. Athena seems to have enjoyed great popularity in Koptos, where she appears as the second-best represented deity in the corpus of terracotta figurines found by Adolphe Reinach in the Kom al-Ahmar.16 Reinach attributed this to the presence of the military. One is tempted to think also of an interpretatio of the goddess Allath, who was venerated among the Arabs, but also among Nabataeans and Palmyrenes. Strabo already in his time – he was a contemporary of the first Roman constructions in the Eastern desert – described Koptos as a “town [which was] shared by the Arabs and the Egyptians.”17 It has to be said, however, that those peoples did not leave many traces in the praesidia, particularly at the time when the network enjoyed its floruit, before the Antonine plague. Apart from Phoinikon, a halt that was common to both roads, the names of the praesidia on the Myos Hormos road had remained unknown until our campaigns. We can now name three of them: Maximianon, Krokodilo (named after the shape of the rock towering over it, when seen from a certain angle: see figs. 26–27 in Chapter 1) and Persou, which was the name of the now disappeared praesidium 13. O.Xer. inv. 995. 14. For the identification of Kainon Hydreuma, see Chapter 1, pp. 52 f. 15. A. Bülow-Jacobsen, “Den syngende korporal,” Festskrift til Chr. Gorm Tortzen = Aigis. Supplementum I (Copenhagen 2011): . 16. Cuvigny (ed.) 2006: 597. 17. 17.1.44.

Men and gods in a network

303

in Biʾr Umm Fawakhir.18 Simiou must be quite close to Maximianon, since two epistolary proskynemata were found to the Tyche of Simiou. The most likely site for identification is Biʾr al-Sayyala, the architectural complexity of which may be revealing of an ancient origin; this would agree with the idea that Σιμίου would be the genitive for Σιμ(μ)ίας, the name of a famous explorer sent to the African coasts by Ptolemy III.19 Epistolary proskynemata appeared earlier in the Desert of Berenike than in the northern zone of quarries of Claudianus and Umm Balad.20 The way the gods are distributed in the proskynemata in letters found in forts is worth examination. While doing so, the following facts should be taken into account: (1) the three forts under consideration on the Myos Hormos road were direct neighbors (the fort in Biʾr al-Hammamat did not come into existence until Krokodilo stopped functioning and was abandoned at the beginning of Hadrian’s reign); (2) while Dios and Xeron are neighbors, two forts which have remained uninvestigated, Aphrodites and Kompasi, stand between Didymoi and Dios; (3) a strip of road connected Krokodilo with Didymoi. I have indicated the proskynemata “before the gods of the place” and “before all the gods” even though they do not help us in identifying where the letters come from, only to highlight the fact – for which I cannot account – that proskynemata before several undetermined gods are more common on the Myos Hormos road. I consider them as alternatives to the invocation of one god, which depended on personal mood or choice, rather than as a local practice linked to a specific fort. Interestingly enough, one of those proskynemata reads παρὰ τῇ κυρίᾳ Ἀ[θηνᾷ] καὶ τοῖς ἐνθάδε θε[οῖς].21 Indeed the discovery of the chapel in Dios, which is well preserved, has shown that in that fort, in which letter writers would traditionally invoke Zeus, several σύνναοι θεοί shared the sanctuary (Chapter 31). In the three forts situated on the Berenike road, the link between the god who is invoked and the origin of the letters has been established, except for a slight uncertainty as regards Pan. He was most likely the patron of Phoinikon, or maybe of Krokodilo. Strangely enough, Phoinikon is underrepresented in the proskynemata in letters found in the two forts directly neighboring it, Krokodilo and Didymoi, even though it was situated at an intersection between two roads and must have been an important halt.22 Still on the Berenike road, we find that the proskynemata before the god of the fort where the letter was found (which means that it was never sent) come in second or third position. Dios and Maximianon are on balanced terms with both their neighbors, whereas Xeron and Krokodilo maintain one-sided relationships: Krokodilo towards Persou, and Xeron towards Dios, through which it receives vegetables from Kompasi. There is no concentration of proskynemata revealing regular exchanges of any sort between Xeron and its southern neighbor, so much so that without O.Krok. 61, an official letter dated 102/3 or 121/2 which mentions Phalakron, one could have believed the fort to have been built later, just like Qusur al-Banat. Both Persou and Kompasi are – or were – gold mines, and enjoyed abundant water resources as well as being market gardening centers which supplied at least both immediate neighbors with fresh vegetables; one may thus account for the prevalence of proskynemata before Athena and Techosis. The latter is, like Philotera, a remnant of a time previous to the Romans’ taking control of the desert. While Τέχωσις is a well attested personal name in the whole of Egypt, it is not otherwise known as a theonym. That the god Apollo appears in a proskynema found in Krokodilo does not imply that the letter comes from Apollonos Hydreuma, as in the ostraca from Dios and Xeron; by a process of elimination, our conclusion is that he could be Krokodilo’s genius loci. 18. Cuvigny (ed.) 2006: 99. 19. Cuvigny (ed.) 2006: 133. 20. O.Did., p. 5. 21. O.Krok. II 234. 22. At the time of the Napoleonic Expedition to Egypt, it was still a busy road junction, where three large wells could be seen (cf. Cuvigny [ed.] 2006: 16).

304

Rome in Egypt’s Eastern Desert

Table 18.1. Proskynemata in ostracological corpora from praesidia in the Berenike Desert Berenike Road Didymoi Aphrodite: 11 Pan: 4 Dioscuri: 1 ἐνθάδε θεοί: 3

Dios Techosis: 60 Athena: 42 Zeus: 16 Apollo: 2 Aphrodite: 2 Dioscuri: 1 ἐνθάδε θεοί: 9

Myos Hormos Road Xeron

Krokodilo

Persou

Maximianon

Zeus: 23 Athena: 9 Apollo: 2 Techosis: 2 Pan: 1 ἐνθάδε θεοί: 5

Athena: 32 Apollo: 1 Philotera: 1 Pan: 1 Dioscuri: 1 ἐνθάδε θεοί: 20 πάντες οἱ θεοί: 7

Sarapis: 1 Athena: 1 ἐνθάδε θεοί: 3

Athena: 62 Sarapis: 30 Philotera: 17 Tyche of Simiou: 2 ἐνθάδε θεοί: 22 πάντες οἱ θεοί: 1

The identity of Maximianon’s tutelary deity remains uncertain: either Sarapis, or Philotera. The proskynema before Sarapis in an ostracon from the limited corpus in Persou23 may be put forward when arguing in favor of Sarapis. Nevertheless, the large number of proskynemata before Sarapis, which represent so many letters that were not sent, would have me tip the scales in Philotera’s favor, as the latter also appears in a proskynema from Krokodilo. The author of the letter is a certain Prokla, who from other sources is also known to have been a prostitute, which led her to come and go in the forts. We know actually that she was sent to Simiou, but also to Maximianon.24 In that case, Sarapis would be worshipped in Al-Hamraʾ, a station that seems to have been built at the same time as Biʾr al-Hammamat, towards mid-second century, after Krokodilo was abandoned at the beginning of Hadrian’s reign because it was too much at risk of flooding from the wadi.25 That the Romans should have chosen, when founding Maximianon on a new site, to place it under the protection of the deified sister of a Ptolemy, remains puzzling. This is why Adam Bülow-Jacobsen thinks that Philotera could be Simiou’s protector, given that the latter was founded earlier, possibly in the Ptolemaic period judging from the remains in Biʾr al-Sayyala.26 Philotera was then identified with Simiou’s Tyche;27 indeed both proskynemata before Simiou’s Tyche which were found in Maximianon come from the latest layers (late second–early third century). Appealing though it may be, such a hypothesis leaves us without a god for Al-Hamraʾ, a fort which must have been built earlier than Qusur al-Banat, judging from its architecture and rare ceramic finds, and which must have been in use for at least half a century at the same time as Maximianon. It would be quite surprising also, that after the building of this fort, Maximianon should have maintained such sustained epistolary exchange with Biʾr al-Sayyala/Simiou. The cult of Philotera may have been embedded enough in the region for the founders of Maximianon to choose her as the fort’s patron. One port near Myos Hormos still bore the name Philotera under Trajan, though probably a distinction must be made with the Philoterion, or Philotera’s temple, to which two women wanted to go and had to get a permit delivered from a duplicarius.28 One of the letters sent from Persou to Maximianon not only bears a proskynema before Athena, but also includes, in a strongly-worded passage, an oath by Philotera.29

23. O.Fawakhir 24 (= SB VI 9017/24). Cf. Cuvigny (ed.) 2006: 53. 24. Prokla in Simiou: O.Krok. II 182; in Maximianon: O.Krok. II 221. 25. Cuvigny (ed.) 2006: 199 f. 26. Cuvigny (ed.) 2006: 133. 27. In the ostraca from Mons Claudianus, the letters which come from Raïma contain proskynemata in front of Isis or in front of Raïma’s Tyche, which are two ways of designating the same goddess (O.Claud. II, p. 66). 28. Van Rengen 2001: 233–36. Chapter 1, pp. 72 f. 29. O.Max. inv. 172.

Men and gods in a network

305

2. Oracular consultation in the chapels Were the gods who were invoked in proskynemata the same gods as those worshipped in the praesidia’s chapels? Chapels in Roman military camps were dedicated to the cult of the emperor, that of the dii militares, and above all the cult of the deified standards that were kept there. What was the situation in the praesidia? The phrase τὰ πρινκίπια τῶν κυρίων, which refers to the chapel of Aphrodites in O.Did. 31, suggests at least that the imperial cult was practiced there. On the Myos Hormos road, only two chapels have been identified, in Maximianon and Qusur al-Banat. They are but skeletons, of which only the walls, the side benches and a base against the back wall, remain. The equipment has completely disappeared, which is not the case however in the sand-filled chapels in the forts on the Berenike road. In Didymoi, the furniture which survived an old pillaging is in an Isiac, even Egyptian-looking style (libation tables). It includes several representations of Sarapis, as well as two inscriptions in which a soldier relates an injunction which he thought the god had sent to him in a dream. On the other hand, there is no trace of the Dioscuri, the gods of the proskynemata, except for a small steatite plaque representing an armor-clad figure leading his horse in a bridle.30 In Dios, where letter writers perform their proskynemata before Zeus (without an epiclesis), we have found on a rostrum the remains of two cult statues which we were able to identify, thanks to the dedication, as Zeus Helios Megas Sarapis (flanked by Cerberus) and the Tyche of the praesidium, with its preserved head showing the so-called Libyan locks which are characteristic of Roman Isis (Fig. 110 in Chapter 31). Had Zeus Helios Megas Sarapis, which was created under Trajan, come to monopolize the forts’ chapels? The answer could come from Xeron, where letter writers entrusted Athena with those they were writing to; if we were able to find the chapel, there would be no possible confusion, as in Dios, between Zeus and Sarapis. The aedes of Xeron was discovered by Michel Reddé in January 2012. Of a less elaborate architecture than the chapels in Dios and Didymoi, it was poorly equipped, with only two objects which enabled us to identify a deity: a small terracotta figurine of Athena and an arm made of bronze from another statuette of the same goddess, which had fallen in the neighboring cistern. Its use was confirmed by an ostracon which was a draft or a copy of a report addressed by Xeron’s curator to the prefect of Berenike concerning a night-time incident which had occurred in the Athenadion.31 The only possible attestation of a cult of signa comes from Dios, in this case the standard of a tiny vexillation of about fifteen men which constituted the garrison – if, that is, such a standard actually existed. This ostracon, O.Dios inv. 1005, is a substitute for an inscription, yet it is not devoid of ambiguity (ἐπ’ ἀγαθῷ τ̣ῷ οὐ vac. η̣ξίλλου).32 If my reading and interpretation are correct, this unassuming dedication reads: “good luck to our flag.” It comes from the filling of a tiny staircase fitted inside the rostrum supporting the statues, with eleven more ostraca which are the membra disiecta of a collection of oracles, the content and presentation of which share common points with the oracles using astragali in the southwest of Asia Minor and also, to a lesser extent, with the Sortes Astrampsychi. They give advice based on common sense and make predictions which are just as vague as they are soothing, using metaphors which can be applied to any type of circumstances: “Do not fear the waves of the sea. Do not fear the storm, and you shall not suffer from it. Trust when you pray to the gods, and they will guide you to your destination…” As is the case with the oracles using astragali, the oracles found in Dios mostly seem to be aimed at travellers. The dedication made in Dios by an Alexandrian naukleros indeed shows that people who were just passing could also find solace in the chapels of the praesidia. The oracular 30. Cuvigny (ed.) 2011: 89, fig. 121. 31. O.Xer. inv. 731. 32. Chapter 31, esp. pp. 500 f. An in-depth study of the oracles in Dios may be found in this chapter.

306

Rome in Egypt’s Eastern Desert

answers are preceded by a heading made up of two or three elements: a numeral, the time of the day which is most suitable for the consultation of the oracle (or this paradoxical warning: “Do not consult!”) and sometimes, the name of a god in the genitive, the latter being the god who is supposedly granting the oracle to the consultant and who is not among the gods worshipped in the chapel, for instance: γ̅ μὴ χρῶ Λητοῦς, “3. Do not consult! (oracle) of Leto” (Fig. 121 in Chapter 31). Three tiny ostraca coming from the Athenadion in Xeron may shed a light on the way consultations worked, while showing that the use of oracles was not confined to Dios.

Figure 64. O.Xer. inv. 809. © A. Bülow-Jacobsen

O.Xer. inv. 809 (Fig. 64) US 38e02

8 × 5 cm →

2 l. Ἑρμοῦ “1. (Oracle) of Hermes.”

α Ἑρμοῦς

second–third century Nile-silt clay

Men and gods in a network

307

Figure 65. O.Xer. inv. 810. © A. Bülow-Jacobsen

O.Xer. inv. 810 (Fig. 65) US 38e02

5 × 5 cm ↓

ιδ Ἄρεως

“14. (Oracle) of Ares.”

Figure 66. O.Xer. inv. 829. © A. Bülow-Jacobsen

second–third century Nile-silt clay

308

Rome in Egypt’s Eastern Desert

O.Xer. inv. 829 (Fig. 66) US 38e02

5.2 × 4.5 cm

second–third century Nile-silt clay

The top part of the support has been destroyed by salt, but there seem to remain too many traces of ink for it to be a simple numeral. There may have been another theonym above line 2. → 4

traces of 1–2 lines δέ̣μ̣ωνος καταχθωνίου

“… (Oracle) of the subterranean god.”  My idea is that these notes are tokens which enabled one to draw oracles; hypothetically thus, number 14 would refer to oracle number 14 in the collection, and it was explained to the consultant that the oracle had been sent to him by Ares, whose name could also appear in the heading of the oracle as it was presented in the collection.

3. The official mail and the meaning of δίπλωμα The written proof that the praesidia in Qusur al-Banat and Biʾr al-Hammamat did not exist under Trajan lies in the postal journals which were found in Krokodilo: the two halts mentioned upwards and downwards from Krokodilo are Persou and Phoinikon. We know little about the way official mail worked in the Nile valley during the principate. The most informative document we have on that topic is P.Ryl. II 78 (157), which seems to indicate that the offices of nome strategoi served as stations, beside secondary stations called στατιῶνες, the mail being transported by ἐπιστολαφόροι. In the Eastern Desert, praesidia were used as stations for the official mail, the working of which we now know quite well thanks to the journals held by the curator praesidii. Praesidia were placed every 25–30 kilometers on average. Given the nature of the country, sometimes sandy, sometimes rocky, a horse rider could cover the distance in two or three hours without forcing his mount. Each garrison included a small number of cavalrymen who would take turns and go on an errand whenever there was mail to transport or someone to escort towards the next station, in either direction. In some cases, the mail was transported not by the military but by the so-called μονομάχαι.33 We have distinguished in Chapter 17 three types of registers produced by the curatores praesidiorum in their functions as postmasters: (1) the postal journals as such, with daily entries; (2) the books with the mail that was received, on which the circulars that had passed through the curator’s hands were copied in chronological order (libri litterarum allatarum); (3) a mixed genre, in which the official mail packing slips were copied out with, at the bottom, a statement by the curator himself, in which he specified the day and time at which the mail had reached him, as well as the identity of the messengers who had brought it and taken it away. Let us keep in mind that those registers can only be drafts or personal copies of official documents written on papyrus. It must be emphasized that the curator did not include the letters, even those of administrative content, that had been addressed to him by other curators of the network; only the mail that 33. Chapter 15, p. 256 and Chapter 17, pp. 273–75.

Men and gods in a network

309

emanated from administrations run by leading figures of procuratorial rank (in this case, the prefecture of Egypt and the prefecture of Berenike) was submitted to this type of double checking. The same is true for P.Ryl. II 78, mentioned earlier. The most instructive example for type (1) is O.Krok. I 1 (c. 108, Fig. 67). This journal covers a fortynine-day period, which allows us to know with certainty the number of cavalrymen from the Krokodilo garrison who were assigned to mail transportation (three). In Persou, there are four of them, possibly five. Here is an extract from this journal (lines 11–13): κζ̅ γ κλῆ(ρος)· ἐπιστολαὶ Κοσ[κωνίου ἐπάρ]χ̣(ου) καὶ δίπλω(μα) δι’ οὗ γρά[φει κουρά-] τορσι Αὐρήλιον (ἑκατοντάρχην) ἀπ̣ο̣[καταστῆσα]ι ⸌ἀ̣ν̣[c. 10 Ὅ]ρμου⸍ ἀπὸ πραισιδίου ἰς πρ[αισ]ί[διον][1–3] ἑτέραι ἐπιστολαὶ ἠνέκθ(ησαν) ἀ[πὸ] Φοινικῶνο(ς) διὰ Καρπουρνίου ὅραν ϛ̅ ἡμέρα{ι}[ς]· ἰς Πέρσου Αἴστις. “On 27th. No. 3. Letters from Prefect Cosconius and a diploma in which the latter writes to the curators, telling them to escort centurion Aurelius (…) from fort to fort, (as well as) other letters, were brought from Phoinikon by Calpurnius at the 6th hour of the day. In Persou, Aestiv(i)us.” This entry records the coming of official mail on the Mecheir 27. It consists of several letters from Berenike’s prefect, a diploma also from the latter, the content of which is summarized, and other letters, for which the curator did not find it necessary to mention senders; in O.Krok. 1, only important senders are indicated. Calpurnius, a cavalryman from the garrison in Phoinikon, brought the mail towards 11 o’clock in the morning. Aestiv(i)us, a cavalryman based in Krokodilo, almost immediately took it to Persou; the curator hardly had time to read the diploma and make a copy of it. The mention of a numbered kleros is specific to this journal and is evidence of a system which aimed at an equitable sharing of turns between the three cavalrymen, to whom a fixed number had been assigned, Aestiv(i)us thus being number 3. The word κλῆρος in the sense of “turn of duty” had not been attested as yet, but it is indeed a meaning of the Latin sors, for which κλῆρος is a semantic calque here. The only complete postal journal found in Didymoi covers a period from Mesore 8 to Thoth 2 (O.Did. 22). It is very succinct and gets into few details, just a few lines on an ostracon. The writer does not make any difference between letters and diplomata; epistolai are all that are mentioned, without further precision as to quantity or senders. The identity of the messengers is similarly left out, only the direction of circulation (from Phoinikon to Aphrodites, or vice versa) and the arrival time are specified. As it was important nevertheless that the identity of the messengers should be known so that they could be sanctioned in case they lost something or were late, this document must be but a summary of a journal. Type (2) is attested only in Krokodilo (copies of official circulars are found on other sites, but they are written on separate ostraca, as far as the fragmentary state of the greater part of the sherds enables one to decide whether one is dealing with a fragment of a register or not). O.Krok. 87, the “amphora of the Barbarians,” offers a spectacular example of a collection of copied circulars addressed to commanders in charge in the Eastern Desert, including curators, keeping them informed of the clashes with the Beduins (who are called Barbaroi) taking place here and there. This correspondence dates to the second year of Hadrian’s reign. Each circular is introduced by the mention ἀντίγραφον διπλώματος, “copy of diploma.”

310

Rome in Egypt’s Eastern Desert

Figure 67. O.Krok. I 1. © A. Bülow-Jacobsen

Men and gods in a network

311

Type (3) is mostly represented by the archives of curators Dinnis in Dios and Turbo in Xeron. Such a type of register may have been integrated late into the bureaucratic habits of the Roman army; it is at any rate unknown in Krokodilo, a fort which was abandoned towards 120.34 Turbo’s archive is published in Chapter 17. Dinnis, a Thracian, wrote his journals, or had them written,35 sometimes in Latin, sometimes in Greek, yet always in a beautiful Latin scriptura actuaria, which indicates that the message writer is a bilingual whose “strong” language is Latin.36 The following ostracon has kept the delivery receipt for a mail arrival. It is in Greek, but there are others in Latin. I do not know of the reasons which led Dinnis, or his secretary, to prefer one language or the other.

Figure 68. O.Dios inv. 807. © A. Bülow-Jacobsen

O.Dios inv. 807 (Fig. 68) US 4805

14.2 × 6.5 cm

second century Nile-silt clay

Greek aspirations (rough breathings, aspirated consonants) are not transcribed by the scribe. There are some unnecessary doublings (epistullas, dipplomate, Accillas, Luppu, orran) and influence of Latin (epistullas in place of epistolas, dipplomate in place of diplomati).

4

↓

_ _ _ _ _ _ _ _ _ _ [Dinnis curator Dios] parelaba [ta]s procimenas epistullas esprag[is-] menas sun dipplomate XXVIII Paopi oran X• emeras dia Accillas ippes turmes Luppu cei aute orran apelusa dia Arpocras ippes turmes Bassianes Paopi XXVIII oran X emera

Leg. 2 proceimenas epistolas 3 diplomati 4 horan hemeras Achillatos hippeos cai auten horan 6 Harpochratos hippeos 7 hemeras

5 Loupou

34. The stratigraphy of the dumps in Dios and Xeron has not been fully studied yet. It is bound to be lacking in precision, for dated texts are scant, and will thus mainly rely on ceramic typology. In Didymoi, where dated or datable texts are more numerous, the oldest type 3 document was thrown away towards 140–150. 35. We have one letter in Greek from Dinnis: the handwriting, which is very clumsy, appears to be his own. I do not think that the scribe who calligraphed the mail logbooks would have had such a childish hand in Greek, even if he had been totally ignorant of that language. 36. Cf. Adams 2003: 30 (about the opposite case of Latin texts written in Greek letters).

312

Rome in Egypt’s Eastern Desert

“[Dinnis curator in Dios,] I have received the aforementioned sealed letters together with the diploma which were brought by Achillas, a cavalryman in Lupus’s squadron, on the 28th of Phaophi at the 10th hour of the day, and I have sent them forward at the same hour with Harpokras, a cavalryman from the squadron formerly of Bassus, on the 28th of Phaophi at the 10th hour of the day.” Turbo’s register allows us to understand that what Dinnis calls diploma in his acknowledgements of receipt, is in fact the packing list, or dispatch note, which very often came from the prefect of Berenike himself, who detailed the list of letters and commended them to the diligence of the curators. In the Eastern Desert ostraca, the word diploma may refer to missives dealing with any type of official business, but also to such packing slips used for official mail. The common point between them, which accounts for both of them bearing the same diploma label, is in my view the fact that we are dealing with circulars in which the prescribed formula mentions several addressees. In one case only is diploma used instead of ἐπιστολή in a letter written to a unique addressee, a report on the attack of the fort of Patkoua sent by the cavalryman Antonius Celer to the centurion Cassius Victor (O.Krok. I 87.18 and 27). This report however is meant to be widely distributed; unless the scribe made a mistake, this is probably why it is referred to as a diploma. The packing slip-diploma itself bore a technical name, δίπλωμα ἐπιθέσεως, which appeared in a fragment of a register found in Dios, in which the setbacks in the distribution of the mail could possibly have been logged (Chapter 16). In this expression, the word ἐπίθεσις refers to the act of entrusting a messenger with some mail. Ancient authors mention this mode of transmission, as do also quite likely two papyri from Dura Europos, yet this is not certain. Messengers were placed (disponere) at intervals, and took turns to convey a piece of news or some mail.37 This is how Domitian was able to follow the various stages of Agricola’s death throes: momenta ipsa deficientis per dispositos cursores nuntiata (Tac. Agr. 43). The annotation dispositus may be found next to about fifteen names of soldiers in the registers P. Dura 100 and 101. Editors have speculated that these were riders placed at intervals to take turns, yet this may not be necessarily so, since disponere may apply to the positioning of sentinels as well as the laying out of an ambush. The word disponere finally appeared, though as a compound, in an ostracon from Dios.

37. Kolb 2000: 25 f., 289, 305.

Men and gods in a network

313

Figure 69. O.Dios inv. 145. © A. Bülow-Jacobsen

O.Dios inv. 145 (Fig. 69) US 2402

second century Nile-silt clay

13.5 × 11.5 cm

This letter was sent from Kompasi, as indicated by the proskynema, the vegetables, and the prefixes in the verbs expressing movement. ↓ 4

8

12 Leg. 1 Ψεντουάνσει 7–8 ἱππέα 12 ἐγγύς

Αἴλις ⟨Σ⟩α̣ραπίων Ψεντουάσι τῷ ἀδελφῷ πλῖστα χαίριν. τὸ προσκύνημά σου ποιῶ παρὰ τῇ κυρίᾳ Τεχώσι. κόμισαι ἀπὸ τοῦ ἱππῆες τοῦ προδησποσίτου τοῦ ἀναβάντος σήμερον δέσμην κράμβης· ἀπέσταλκά σοι ἄλλην καὶ ᾔτηκα τὸν ἱππῆα τ̣ὸ̣ κέρμα. ἐ̣ὰ̣ν̣ ε̣ὕ̣ρ̣ω ἀναβαίν̣ο̣ν̣τα̣, πέμψω σοι π̣ά̣λ̣ι̣. ἀσπάζομαι ρ̣ι̣ν̣ καὶ Ἑρμῖνο̣ν̣ καὶ Πετόσιριν. ἐὰν εὕρῃς εὐκαιρίαν, κάτελθε ἐνγὺς ἡ̣μ̣ῶ̣ν̣. ἐρρῶσθαί σε εὔχ(ομαι). 2 πλεῖστα χαίρειν

3–4 Τεχώσει

ἱππέως

5 προδισποσίτου

314

Rome in Egypt’s Eastern Desert

“Aelius Sarapion to Psentouansis, his brother, best greetings. I make prayers on your behalf before our lady Techosis. Receive a bunch of cabbages from the cavalry messenger who will have gone up today. I have sent you another one and I have asked the rider for the money. I will send you some more of them when I find someone going up. My greetings to [--]ris, Herminos, and Petosiris. When you get a chance, come down and see us. I wish you good health.” 1.

Ψεντουάσι. The name Ψεντούωνσις, less often written Ψεντούανσις, is almost solely attested by mummy labels from the Panopolite.

6.

“A bunch of cabbages,” that is to say, small broccoli stems, cf. Chapter 38.

6–8.

The context implies that Sarapion previously sent this other bunch. ᾔτηκα, “I have asked the rider for the money.” My personal feeling is that Sarapion asked the prodispositus rider to claim from Psentouansis the money for the two bunches of cabbages that he successively sent and to bring back that money to him. The claim will more likely be heard if introduced twice, once in the written form and once orally.

9 and 11. ἐάν. For the translation “when,” cf. Chapter 40. Prodisponere is a hapax. However, the Thesaurus Linguae Latinae has an entry “? praedispono.” It refers to a passage of Livy (40.56.11), whose meaning remains unclear and which has been thus edited in the CUF: Medicus Calligenes, qui curationi praeerat, non expectata morte regis, a primis desperationis notis nuntios per dispositos ⟨equos⟩, ita ut convenerat, misit ad Perseum.38 The end of book 40 by Livy, in which this passage may be found, relies on one manuscript only, the Moguntinus, which is lost today. Carbachius used it for his 1519 edition. Yet he had a hard time transcribing the paleography of the Moguntinus, and ended up introducing many misreadings. After autopsy of the Moguntinus, Gelenius corrected many of them in his 1535 edition.39 What Carbachius had seen in the Moguntinus was nuntios perdispositos, ita ut convenerat, misit – which he faithfully copied insofar as he was able. Gelenius’s reading was nuntios praedispositos, ita ut convenerat, misit. In 1555 however, in the edition which he published in Venice, Sigonius suggested restoring per dispositos equos. Such a restoration relied in particular on a Livian parallel, 37.7.11: Gracchus … per dispositos equos… Pellam pervenit. Since praedispositus is not otherwise attested, modern scholars have followed suit.40 Thanks to the Dios ostracon, we now know that a compound did exist, and that it was precisely used in that type of contexts.

4. Rotating women We have explained how it came about that the letters from the prefect of Egypt were kept for nine hours in Dios because they had inauspiciously arrived towards eight o’clock in the evening, while the prodispositus rider was engaged with a woman. Praesidia were not without female company, which goes

38. “The physician Calligenes, who was in charge of his treatment, without waiting for the death of the king, at the first signs that the case was hopeless sent messengers to Perseus, as he had arranged to do, on relays of horses, and until his arrival concealed the death of the king from all who were outside the palace.” (trad. E. T. Sage and A. C. Schlesinger 1938.) 39. Though possibly not all of them, cf. Livy, Histoire romaine (CUF), volume XXX, p. CXXVII. 40. Thus J. Briscoe, A Commentary on Livy. Books 38–40 (Oxford 2008)  560: “There can be little doubt that Sigonius’ supplement equos is correct (…).”

Men and gods in a network

315

against the idea one usually has of discipline in Roman camps.41 Several letters written by procurers, their brokers, their military clients, and even angry prostitutes give us quite a clear idea of how prostitution was organized in the fort network.42 The garrison soldiers, once they had agreed on the choice of a prostitute, clubbed together to hire her on a monthly basis. This hiring contract had a technical name, the (μίσθωμα) κυκλευτικόν, or “rotation contract.” I thought at first that this term hinted at the ambulatory nature of a prostitute’s job in the Desert of Berenike, since these girls rotated between the praesidia, just like the gyrovago monks in the Apophthegmata Patrum, whose peregrinations from one hermitage to another are also described using the verb κυκλεύειν. Yet one letter, O.Dios inv. 1420, clearly shows that the prostitutes did indeed rotate, not from one fort to the other, but from one client to another. Its author, Faustus, writes to [--]dora, with rough spelling: εἰὰν θέλῃς ἐλθεῖν ὧδαι, ἐλθὲ εἵνα κυκλεύσῃς “if you want to come here, come to rotate.” It is worth noting that the addressee has a say in the matter, which is also the case in another letter, also found in Dios and coming from Kompasi, just like the text mentioned above. O.Dios inv. 1246 (Fig. 70) US 3746

12.5 × 17 cm

second century Nile-silt clay

Right from the first lines of the letter we understand that there exists a rivalry between two prostitutes and their supporters. The prostitute who “rotated” in the fort where Arabion, the author of the letter, finds himself, has left, thus making way for Sarapias, of whom one part of the garrison in Kompasi was already particularly fond. It is hard to see Arabion as a procurer, since he will not keep for himself the amount of the kykleutikon. It seems rather that he could be a faithful client who has maintained a good relationship with Sarapias and looks forward to seeing her again. The kykleutikon was worth 30 staters, an amount which could not be negotiated, since the prostitute’s fee was usually made known to all. In the hiring offer O.Did. 430, the price is even used to identify the paidiske on whom the garrison has set their sights: “that little 60 drachmas’ worth winner of yours.” Sixty drachmas (15 staters) is indeed the most common price for a kykleutikon. Do the 30 staters imply that Arabion expects Sarapias to hire her services for two months? Yet it seems likely that he would have specified that, given that a contract usually lasted a month. Another possibility is that this woman, who worked for herself and is likely to have been a star in the Berenike Desert,43 could have charged high fees. The list of food products which Arabion intended to send to her – though he was unable to do so in the end – makes it less likely however for her to have been a wealthy call girl. →

4

Ἀραβίων vacat Σαραπιάδι τῇ ἀδελφῇ χαίριν. γινώσκιν σε θέλω ὅτι 4–5 κατέβη· εὐχαριστῶ δὲ τοῖς θεοῖς ὅ̣τι καὶ ̣δε αὐτῆς οὔσης οἱ φίλο̣ι̣ μου ἤθελάν σε ἐλθεῖν καὶ πάλι ἄ̣ρ̣-

41. Cuvigny (ed.) 2006: 361 f. 42. Chapter 24. 43. Several ostraca from various forts mention prostitutes with that name, who could be one and the same person, obviously of the querulous type.

316

Rome in Egypt’s Eastern Desert 8

12

16

20

τι. εἰὰν θέλῃς ἐλθεῖν, γράψον μ[οι] καὶ πέμψω σοι τριάκοντα στατῆ[ρας.] ὄ̣νος ἔνι ἐνθάδε, ἐμοῦ διδόντος τὸ φόλετρον καὶ ἐρχομένης καὶ ὑπαγαγούσης· τὸ περὶ τῆς κου[ιν-] τάνας ἐμοὶ μελήσι. vac. κόμισαι παρὰ Φήλικος δέσμας β κράμβη̣ς̣· ἐδίδου δὲ αὐτῷ ἐν τῷ κιθων 1–4 μάτειν φοίνικος καὶ εἱμιμάτιν πτομάτων καὶ ε⟨ἱ⟩μιμάτιν ἐρεβι⟨ν⟩θίων καὶ οὐκ ἠθέλησε βαστάξαι ὅτι καὶ ἄ̣ρ̣τος εἶχεν ἀρ̣ῖν.

Leg. 3 χαίρειν 4 γινώσκειν 6 ὧδε 10 ὄ̣νος: ς ex ν corr. 7 ἤθελον 13 μελήσει 16 μάτιον 16–17 ἡμιμάτιον 17 πτωμάτων 19–20 ἄρτους

8 ἐάν 11 φόρετρον 20 ἀρεῖν

“Arabion to Sarapias, his sister, greetings. I want you to know that Isis(?) has come down. I give thanks to the gods that even while she was here my friends wanted you to come down, and that it is still the case now. If you agree to come here, write to me and I shall send you thirty staters. There is a donkey here and I will pay for your coming and going. As for the quintana, I’ll see to that. Receive from Felix two bunches of cabbages. I wanted to give him in the tunic a mation of dates, half a mation of fallen olives and half a mation of chick peas, but he would not take them with him for he also had bread to take.” 1.

Another letter from Arabion to a “sister” called Aphroditous, herself the author of letters sent to a certain Boubas based in Xeron, was found in Dios (Chapter 22). The handwriting is different, which does not necessarily mean that this is a different Arabion: many inhabitants in the forts were illiterate and had someone else write their letters.

4.

The word which remains unread at the end of the line must have been a female anthroponym. Εἶσεις (l. Ἶσις) could match the traces.

5.

κατέβη. In the Eastern Desert, the preverb κατα- indicates a journey in the direction of the Nile valley.

11-12. φόλετρον, κουιντάνας. The expense of transportation as well as the quintana are also borne by the clients in O.Did. 430, while in O.Did. 390, the procurer is responsible for them. As regards the quintana (a lease tax on trade transactions in military zones), see Chapter 23. ὑπαγαγούσης. Nothing can account for the aorist but the clumsiness of the writer. On the importance for a prostitute not to linger in a garrison after the expiry of the contract, see Chapter 25. 15.

ἐδίδου. Conative imperfect (Mandilaras, Verb, § 293). ἐν τῷ κιθων 1–4. The end of the line is illegible: κιθῶνι or κιθωνίῳ. In the first case, the word was followed by ἡμι- (spelled in one way or the other). The tunic could have been coming

Men and gods in a network

Figure 70. O.Dios inv. 1246. © A. Bülow-Jacobsen

317

318

Rome in Egypt’s Eastern Desert back from laundering (residents in Dios and Xeron sent their clothes to be laundered to Kompasi). Arabion may have wanted to use it in order to wrap the groceries.

17.

πτομάτων. When the term πτῶμα applies to food, it means “fruit which has fallen from the tree.” This is an uncommon meaning in papyri. In the Nile valley, I have only noted P.Fay. 102 (105) and BGU XV 2483.4 (205), in which the collective singular πτῶμα applies to olives which have spontaneously fallen as opposed to those which have been picked or knocked down; in SB XXII 15637.4 and 19, πτῶμα, again a collective singular, is clarified by ἀκάνθης; this papyrus is a receipt for the sale of acacia seeds. The plural πτώματα is attested with the same meaning as here in a small number of letters from the Berenike Desert (Krokodilo, Maximianon, Dios, and Didymoi) which belong for the most part to the dossier of Philokles, a procurer and supplier. Three letters from him in particular record each time the sending of a baukalion (bulging pot) μέστον πτωμάτων (O.Did. 376, 377, and 385). Given that Philokles was probably based in Phoinikon, the editor believes we are dealing with dates (O.Did. 377, introd.), yet the mention of φοῖνιξ here goes against this hypothesis – unless the ptomata are only fallen dates, with the same consequences in terms of quality as with olives. Be that as it may, the term is mostly used for olives, and corresponds to the oleae caducae which Cato keeps for the slaves’ diet (Agr. 58). As indeed they were often maggoty (which is why they fell prematurely)44 and spoiled from lying on the ground, they could not be sold. In O.Did. 377, Philokles specifies the amount of such wretched fruits (forty) which his agent Kapparis will have to give Theanous and Isarous, who may have been prostitutes (could Isarous be the Isis in our letter?): this miserable counting suggests a great deal about the pitiful life women led in the praesidia.

5. The end of praesidia and the “Protoblemmyes” During the last third of the second century, or the first half of the third, at some point which has not been identified for certain, possibly because the process was a progressive one, residents in the forts on the Berenike road stopped taking their rubbish outdoors, and instead left it to pile up in the areas in which they lived, when they did not dump it into disused premises. Floor levels rose, and the interior design became more and more slum-like, with the regular quarters being razed or subdivided, giving way to a clutter of tiny spaces – though the chapels remained remarkably clean. The Roman army still lived there, but we can no longer rely on duty rosters in order to get an idea of the size of the garrisons, which may well have grown much smaller. In such a general slackening of standards, some curators no longer concealed their Egyptian origins under false Greek or Latin names, and now used vernacular names such as Psenthamephis or Psenosiris. Recruitment changed also, with Eastern archers replacing Thracian and Dacian soldiers. One ostracon, which was found inside the fort of Xeron, gives a detailed account of their origins:45 they are Palmyrenes, Hemesenians, and Parthians. The first two ethnic groups have already been attested in the Eastern Desert textual corpus, but it is the first time that we encounter Parthians. Since we cannot give a precise date for the ostracon, we are unable to say whether it comes prior to, or later than, the fall of the Parthian empire to the Sassanians, towards 226. It was not from that time on, as I wrote in the original version of this chapter, but shortly before being abandoned, that the praesidia seem to rely on the nomads from the surrounding desert, with whom 44. I thank J.-P. Brun for this piece of information. 45. O.Xer. inv. 794. Jean-Pierre Brun attributes the slumification of the forts to these Eastern soldiers (Brun 2018: §25 and Chapter 15, p. 245).

Men and gods in a network

319

they were at war at the beginning of the second century. Traces of correspondence in Greek between the praesidia on the Berenike road and a man called Baratit, who introduces himself as the hypotyrannos of the Barbaroi, were found in Dios and in Didymoi. This is the oldest testimony to the title hypotyrannos, which is otherwise known only in a small cluster of sources from the fifth and sixth centuries for two peoples from Nubia, the Blemmyes and the Nubades. Finally, it was in Xeron that Michel Reddé unearthed, in the last occupation layer of the fort, about a hundred orders for delivery of wheat, written on ostraca, for the benefit of men who almost always bear Barbarian names. Among these names the anthroponym Baratit is actually found again.46 These vouchers do not bear any prescript, so the sender or the addressee are unknown. They are all written in the same expert hand, a slanting cursive from the third century, and are dated to the 11th year of an unknown emperor whom Jean-Pierre Brun, after some hesitation, identified as Gallienus.47 Every single one of these orders has been crossed out, obviously so that the same voucher would not be used several times. They all end with the entry ωευρευεξ, written by the same hand as the rest of the text. This does not look like an anthroponym and must be the transliteration of a word or a phrase meaning “I have signed” or “I agree” in the Barbarian vernacular. This entry reveals that the issuing authority is a Barbarian, not a Roman one. Its purpose does not lie in conveying information to the people in charge of distributing the wheat, but rather in proving the validity of the vouchers. In order to get to that result, the scribe would not have passed the calamus to the Barbarian chieftain, who must have been illiterate and would have clumsily traced a subscription of some sort, but he transcribed himself an utterance in the language of the Barbaroi.

46. A great deal of these ostraca had been re-used as tesserae to carpet the bottom of small silos. They are edited in O.Blem. 47. Chapter 27, pp. 429 f.

320

Rome in Egypt’s Eastern Desert

Figure 71. O.Blem. 57. © A. Bülow-Jacobsen

O.Xer. inv. 344 = O.Blem. 57 (Fig. 71) US 1602 ↗

10.5 × 7.5 cm

17 April 264 Nile-silt clay

(ἔτους) ια// Φαρμουθι κβ̅. μέτρ(ησον) Μασαδ ὑπ(ὲρ) ὀνόμ(ατος) Βαδιτ πυρ(οῦ) ἀρτ(άβης) ἥμισου, γί(νεται) (πυροῦ ἀρτάβης) (ἥμισυ). ωευρευεξ.

3 l. ἥμισυ “Year 11, Pharmouthi 22. Measure out to Masad, on behalf of Badit, half an artaba of wheat, that is ½ art. of wheat. Oeureuex.” Even such a simple voucher as that yields two unknown anthroponyms at a glance. It is striking to see the robust onomastic diversity of these people, which allows the scribe not to clutter his text with filiations. This stands in welcome contrast to the monotonous Greco-Egyptian or Latin anthroponymy of the residents in the forts, where everyone had a tendency to be called Ammonios, Isidoros, Antonius, Apollinaris, Serenus, or Sarapion! I have entrusted Helmut Satzinger with the study of this onomastic treasure, which he undertook to meticulously analyze.48 The issue was whether our Barbaroi were the ancestors of the Blemmyes. Even though H. Satzinger was unable to make a connection between the names of the Barbarians in Xeron and those of the Blemmyes from the fifth and sixth centuries, which are known to us through a handful of Greek documents,49 he nevertheless noted that both anthroponymic corpora obey the same structural and phonemic rules. He remains cautious in his conclusion that Baratit and his peers probably did not speak the same lan48. Satzinger 2014. 49. Mainly fifth century inscriptions in Talmis (Kalabsha, in Lower Nubia), the capital city of the ephemeral Blemmyes kingdom; the letter of Phonen, king of the Blemmyes, to Abourni, king of the Nubades, found in Qasr Ibrim (c. 450); the documents on gazelle skin found in Gebelen (sixth cent.).

Men and gods in a network

321

Table 18.2. Appendix. The names of the praesidia on the roads to Myos Hormos and to Berenike1 Modern Name

Name According to the Ostraca

According to the Ancient Itineraries

Tutelary Deity

Inventoried Ostraca

Koptos-Phoinikon stretch common to both roads Al-Matula

unknown

not mentioned

unknown

Al-Laqita

Φοινικών

Poeniconon (It. Ant.) Phenice (Tab. Peut.) Phinice (Cosm. Rav.)

Pan (probable)

*Qusur al-Banat

unknown

not mentioned

unknown

36

*Al-Muwayh

Κροκοδιλώ

not mentioned

Pan? Apollo?

772 (O.Krok.)

ὁδὸς Μυὸς Ὅρμου

*Biʾr al-Hammamat unknown

not mentioned

unknown

1

Biʾr al-Fawakhir

Πέρσου

not mentioned

Athena

59 2

*Al-Zarqaʾ

Μαξιμιανόν

not mentioned

Philotera or Sarapis

1540 (O.Max.)

*Al-Hamraʾ

unknown

not mentioned

Philotera or Sarapis

3

*Biʾr al-Sayyala

Σιμίου?

not mentioned

Philotera? Tyche of Simiou?

10

*Dawwi

unknown

not mentioned

unknown

ὁδὸς Βερενίκης *Khasm al-Minayh

Δίδυμοι

Minayh al-Hir

Ἀφροδίτης (Ὄρους) Afrodito (It. Ant.) Dydymos (Tab. Peut.) Afroditis (Cosm. Rav.)

Aphrodite

Biʾr Daghbag

Κομπασι

Compasi, Comparsi (It. Ant.) Conpasin (Tab. Peut.) Comuasim, Connasim (Cosm. Rav.)

Techosis

*Abu Qurayya

Διός, Iouis

Iouis (It. Ant.) Dios (Tab. Peut., Cosm. Rav.)

Zeus

*Biʾr Bayza

unknown

not mentioned

unknown

*Al-Faysaliyya/Jirf

Ξηρὸν (Πέλαγος)

Aristonis (It. Ant.) Xeron (Tab. Peut., Cosm. Rav.)

Athena

*Al-Duwayj

Φάλακρον

Falacro (It. Ant.) Philakon (Tab. Peut.) Phalacox, Phalacoxs (Cosm. Rav.)

unknown (Athe- 8 na?)

Wadi Jamal

Ἀπόλλωνος (Ὕδρευμα)

Hydreuma Apollinis (Pliny) Apollonos (It. Ant., Tab. Peut., Cosm. Rav.)

Apollo

Abu Ghusun?

Καυα̣1-2

Cabalsi (It. Ant.) Cabau (Tab. Peut.) Gabaum (Cosm. Rav.)

unknown

Dydime, Didime (It. Ant.) Affrodites (Tab. Peut.) Didimus (Cosm. Rav.)

Dioscuri

949 (O.Did.)

1567 (O.Dios)

1303 (O.Xer.)

1. The asterisk marks forts where we have excavated. 2. Found in 1940 or 1941 during an attempt at re-exploiting the tailings of the ancient gold-mines; published by O. Guéraud (BIFAO 41 [1942] = SB VI 9017).

322 Abu Qurayya

Rome in Egypt’s Eastern Desert Καινὸν Ὕδρευμα

Novum Hydreuma (Pliny) Cenon Hidreuma (It. Ant.) Cenonnydroma (Tab. Peut.) Cenon Idrima (Cosm. Rav.)

unknown (Athena?)

guage as the Blemmyes. This language, of which we can only get an idea through anthroponyms, is thought to be an old form of Beja, which is still spoken today, along with Arabic, by the Ababda Beduins in the Egyptian Eastern desert.50 Still, H. Satzinger acknowledges indisputable similarities between protobyzantine Blemmyes’ old Bedja and the idiom of the Barbaroi who came to stock up in the praesidia during the third century. Since then, material found in the Ptolemaic fort at Biʾr Samut has produced undoubtable arguments in favor of an identification. On them, I refer the reader to Chapter 27, p. 435.

50. Claude Rilly, with whom I had the opportunity to discuss this anthroponymy, thus defines the Beja (email dated 14 April 2011): “the language of the Blemmyes is not a ‘Nubian dialect,’ but the ancestor of the Beja, which, like Egyptian or Arabic, belongs to the Afro-Asiatic superfamily (or phylum), Chamito-Semitic in the French tradition. To be more precise, it is the only known language of the Northen Cushitic branch and it is, as such, a cousin of the languages in the Horn of Africa: Afar, Somali, and so on.”