Amheida I: Ostraka from Trimithis, Volume 1 9780814771327

Citation preview

AMHEIDA I: OSTRAKA FROM TRIMITHIS VOLUME 1

This page intentionally left blank

AMHEID AI

OSTRAKA FROMTRIMITHIS Volume 1 Texts from the 2004-2007 Seasons by

Roger S. Bagnall and Giovanni R. Ruffini

with contributions by Raff aella Cribiore and Günter Vittmann

New York University Press and Institute for the Study of the Ancient World 2012

Roger S. Bagnall and Giovanni Ruffini 2012 All rights reserved Printed in the United States of America

Library of Congress Cataloging-in-Publication Data Bagnall, Roger S. Ostraka from Trimithis / by Roger S. Bagnall and Giovanni R. Ruffini ; with contributions by Raffaella Cribiore and Günter Vittmann. v. cm. — (Amheida ; 1) Includes bibliographical references and indexes. Contents: v. 1. Texts from the 2004-2007 seasons. 1. Ostraka—Egypt—Trimithis (Extinct city) 2. Inscriptions, Greek—Egypt—Trimithis (Extinct city) 3. Inscriptions, Egyptian—Egypt—Trimithis (Extinct city) 4. Written communication—Egypt—Trimithis (Extinct city) 5. Greek language—Texts. 6. Egyptian language—Texts. 7. Trimithis (Extinct city)—Social life and customs—Sources. 8. Trimithis (Extinct city)—Antiquities. 9. Amheida Site (Egypt) 10. Excavations (Archaeology)— Egypt—Amheida Site. I. Ruffini, Giovanni, 1974- II. Cribiore, Raffaella. III. Vittmann, Günther. IV. New York University. Institute for the Study of the Ancient World. V Title. PA3371.B34 2012 487'.3-dc23 2011041832 ISBN 978-0-8147-4526-7 cloth

Published under the direction of Amheida Editorial Committee Roger S. Bagnall Paola Davoli Olaf E. Kaper

This page intentionally left blank

PREFACE The ostraka presented in this volume come from excavations conducted at the site of Amheida, located in the northwestern part of the Dakhla Oasis in the western desert of Egypt. Dakhla was part of the Great Oasis of the Graeco-Roman period. Amheida is the largest surviving archaeological site in Dakhla today, although it was certainly smaller in antiquity than the historic capital of the oasis, Mothis (today's Mut), the remains of which survive only in very fragmentary form.1 In the periods from which we have Greek and Demotic texts, Amheida was called Trimithis, meaning "the storeroom in the north" in Egyptian; the name Amheida apparently represents the same root minus tri, "the storeroom", at the beginning. The excavations at Amheida have been carried out by an international team under the sponsorship of Columbia University; with the 2009 season, primary sponsorship passed to New York University, where it is based in the Institute for the Study of the Ancient World. This team has worked in 2004-2007, the years from which the ostraka published in this volume come, under the regional permit of the Dakhleh Oasis Project, directed by A. J. Mills. The Columbia project has been directed by Roger S. Bagnall, with Eugene Ball as field director in 2004 and Paola Davoli as archaeological director since 2005. Olaf E. Kaper has served as the project's associate director for Egyptology. Preliminary reports for each year appear on the project's web site (www.amheida.org). Complete lists of those who have participated in the project during these years are given in these reports. We are grateful to all for their contributions to understanding ancient Trimithis and the support they have given both in the field and between seasons to our work. The present volume has benefited above all from many discussions with Paola Davoli about the contexts and stratigraphy of the house in Area 2.1 (House Bl), and from conservation work and study of the coins by David M. Ratzan and William E. Metcalf. Infrared photography of the ostraka in 2009 by Rodney Ast helped improve readings in a number of texts. We are also indebted to all of the institutions and individuals whose financial support has made the project possible. These also are listed on the web site, but we must record here that the bulk of the funding during 2004-2007 has come from Columbia University (the Academic Quality Fund, the office of the Vice President for Arts and Sciences, the Stanwood Cockey Lodge Fund of the Department of Classics, and the Columbia University Libraries) and the Andrew W Mellon Foundation through a Distinguished Achievement Award to Roger Bagnall. We record our thanks to all of those who helped make the launching of this project possible, including (then) Provost Jonathan R. Cole, (then) Vice President for Arts and Sciences David H. Cohen, and Vice President and University Librarian James G. Neal. In Egypt, we are particularly grateful to Zahi Hawass, (then) General Secretary of the Supreme Council of Antiquities, Magdy El-Ghandour, (then) chief of foreign missions, and Mäher Bashendi Amin, the chief inspector for the Dakhla Oasis in the Pharaonic and Roman Inspectorate, for their constant support of our work. The essential logistical basis of our operations in the oasis is provided by Ashraf Barakat, assistant to the director, and Gaber Mahmoud Murad, 1 Links to publications of the work of the Dakhleh Oasis Project at Mut el-Kharab, including the recent excavations conducted by Colin Hope, may be found at http://arts.monash.edu.au /archaeology/excavations/dakhleh/mut-el-kharab/index.php.

8

OSTRAKA FROM TRIMITHIS

manager of our excavation house. Without them, there would be no excavations. Our project at Amheida was launched with as one of its explicit aims the discovery of new textual material in a fully recorded archaeological context. Anyone who has worked with papyri, ostraka, and inscriptions from excavations knows that context does not always illuminate texts nor the reverse. But equally, in many cases the relationship is deeply rewarding. Both situations occur with the ostraka presented here and will be discussed both in the introduction to this volume and in the forthcoming publication of the structures. A somewhat wider context for our work, and one of the reasons for the selection of Amheida for a field project, is the Dakhleh Oasis Project, mentioned above; the excavations at Kellis (modern Ismant el-Kharab), directed for the past twenty years by Colin A. Hope, were a particular inspiration, and it was the opportunity to work on the Kellis Agricultural Account Book at his invitation that first brought Roger Bagnall to Dakhla in 1996. We are deeply indebted to Dr. Mills and Professor Hope for their help with our work in more ways than we can recount. In the present volume, the Demotic texts have been prepared by Günter Vittmann, who also contributed onomastic suggestions.2 Raffaella Cribiore has contributed to establishing the versions of the various exercises. We are grateful to both. Professor Vittmann wishes to note that because he was unable to see the original ostraka, his transcriptions and comments are preliminary. He intends to provide a fuller publication of these texts at a later date after having the opportunity to work on the originals. The papyrologists involved in publishing the Kellis documentary and literary finds have also contributed much at every stage to our work. In the case of the present volume, above all we owe an enormous debt to Klaas Worp. He and we have shared unpublished material from both sites throughout our editorial labors, and his index of personal names from the oasis has been of great value at many points. He has also read a draft of the present volume, discussed many difficult points in the texts with us, and improved our readings. We thank also two anonymous referees for the Institute for the Study of the Ancient World. New York, November 2010

2 These indude 198, 278, 280, 305, 306 (in abnormal Hieratic), 315, 378, 384 conv., 385, 389-393, 422, 427 and 428.

CONTENTS Preface Contents Figures Note on Editorial Procedure Introduction Types of texts Dating Stratigraphy and coins in House Bl The wells and Trimithite geography Managing wells and estates Commodities and measures Trimithis as a political community House B2 (Area 1.3) Ostraka from the Temple Hill Personal names Archaeological contexts of the ostraka Texts, translations, and commentaries I. Accounts and lists (1-62) II. Ration accounts (63-70) III. Lists (71-95) IV. Well tags: Pmoun formula (96-145) V. Well tags: Hydreuma Pmoun formula (146-172) VI. Tags: Miscellaneous and uncertain (173-247) VII. Memoranda (248-277) VIII. Receipts (278-294) IX. Letters, orders, and notes (295-330) X. Writing exercises (331-337) XL Jar inscriptions (338-343) XII. Miscellaneous (344-353) XIII. Uncertain texts (354-454)

7 9 11 12 13 13 14 23 31 37 41 42 49 53 54 60 77 77 107 110 120 135 144 164 172 179 198 199 200 202

Indices I. Chronological A. Regnal years B. Indiction years C. Months and days II. Personal names III. Geographical terms IV Official and military terms V Professions, occupations, and statuses VI. Religion VII. Taxation VIII. Money

229 229 229 229 229 230 243 246 246 247 247 248

10

OSTRAKA FROM TRIMITHIS IX. Weights and measures X.Greek words XL Demotic and Abnormal Hieratic A. Regnal years B. Months and days C. Personal names D. Professions and occupations E. Words Concordance of Inventory numbers and publication numbers

248 248 254 254 255 255 255 255 256

FIGURES 1. Map of the Dakhla and Kharga Oases 18 2. Plan of Amheida, 2010 19 3. Plan of Area 2.1 (House Bl) 23 4. Plan of Areal 47 5. Plan of Area 1.3 (House B2) 48 6. Area 1.3, DSU 48, with tablet 24 in situ 74 7. Area 4.2, ostrakon 315 in situ 74 8. Area 2.1 (House Bl), Room 9, DSU 127 75 9. Area 2.1 (House Bl), Room 13, DSU 208 75 10. Area 2.1 (House Bl), Room 2, DSU 13 with 298, 299, and 360....76 11. Area 2.1 (House Bl), Room 8, DSU 85 76

NOTE ON EDITORIAL PROCEDURE Texts in this volume are presented according to the usual papyrological practices. The following signs have their usual meanings: ( ) []

{} I I αβγδε

[ ±5 ]

Resolution of an abbreviation or a symbol Lacuna in the text Letters omitted by the scribe Letters erroneously written by the scribe Letters written, then cancelled, by the scribe Letters the reading of which is uncertain or would be uncertain outside the context Letters of which part or all remain but which have not been read Approximate number of letters lost in a lacuna and not restored Letters inserted above the line by the scribe

In general, abbreviations are resolved in the text. Fractions printed are a =lA and S = Vi. The line notes indicate what form the word έτους (year) takes in the text and correct non-standard Greek except in the case of personal names. Errors of case and varieties of spelling in the latter are very numerous. NB: Egyptian proper names without terminations (i.e., undeclined or not hellenized) are not accented. Greek names and words written in a non-standard spelling are generally accented as if the correct form had been written. In translations, Greek and Egyptian names are transliterated directly; Roman names are given in Latin form.

INTRODUCTION Although considerably less visually prepossessing than the wall paintings that form the most distinctive characteristic of the house in Trimithis Area 2.1, which we have designated House Bl, or even than the pyramid tomb in Area 3, the 455 ostraka (a number that includes jar inscriptions and one bone) found throughout the site during the first four seasons (2004-2007) make a major contribution to understanding the economic, social, political, and cultural situation of its inhabitants. They also offer significant information about the geography, institutions, and onomastics of Trimithis as a whole. In presenting them here, we concentrate particularly on the texts from the two houses, the medium-sized one in Area 1.3 (B2) as well as the much larger and more productive house in Area 2.1 (Bl), analyzing them as a source for understanding the families that lived in them and for the history of the houses. We also include texts found in Area 4, the hill on which stood the Temple of Thoth, the principal divinity of Late Period, Saite, Persian, Ptolemaic, and Roman Trimithis. None of the texts from Area 4 was found in an undisturbed ancient context, because of the extensive damage done to this area by stone-robbing, pit-digging, and wind erosion. Only a couple of large ostraka dating to the New Kingdom and Late Period (also found in Area 4) are omitted from the present volume. In our descriptions of the texts here, we include basic information about the find context of each ostrakon. In the volume introduction we also provide an analysis of the stratigraphy of each room, plus a table showing brief descriptions of all contexts from which ostraka have come. When the archaeological contexts are fully published, a catalogue of the ostraka, giving a description of the contents but not the full text, will be provided along with other finds, organized by context. In the present volume, the archaeological context is discussed only where it is germane to the points at hand. This is the case principally in the discussion of the dates to be assigned to the regnal years mentioned in a number of the ostraka.

Types of Texts The single largest and most unusual category of text is the smallest in physical size, the little labels or tags; in their fullest form these bear the name of a well (sometimes of a vineyard), the name of a person, and a date by regnal year. These tags are discussed in more detail for their chronological implications in the following section. Some of the ostraka in very small format that are classified as tags below, however, are even less forthcoming with information; perhaps a personal name with no numbers and no patronymic (e.g., 189) is the extreme case. Three of these tags (161, 200, and 204) were, when discovered, still set in the top of a mud jar stopper. Although other jar stoppers were found in the house, in no other case was an ostrakon preserved in place. The first volume of ostraka from Kellis had contributed only one tag closely similar to ours, O.Kellis 270. The subsequent discovery of 14 complete mud jar stoppers like that in which 161 was embedded, in an earlier Roman building at Kellis during the 2006 season, all with small tag-like ostraka set into their tops in the same manner as 161 and 200, confirmed graphically that this was normally the original placement of these ostraka. In fact, other such ostraka and stoppers were found in earlier seasons at Kellis; some have

14

OSTRAKA FROM TRIMITHIS

not yet been published, others have texts less recognizably similar to ours.3 Probably for the most part these stoppers closed wine jars, or perhaps in some cases olive oil jars. At the same time, however, it is entirely possible that some of the other tags accompanied shipments of other commodities, whether in ceramic containers, in sacks, or in some other form of holder. As most give no information about the contents themselves, we can generally not be sure of the exact original use. These tags have mostly been found in House Bl, but they have also occurred in the excavation of the temple area (Area 4) and in House B2. They are thus not something distinctive at Trimithis to any one archaeological context. Insofar as they are readable,4 the texts published here otherwise fall into wellknown categories characteristic of ostraka: accounts, lists, memoranda or notices, and short letters. These categories do not call for any particular comment. A number of texts are damaged or illegible to such an extent that they defy ready classification; these are published together as "uncertain texts". The letters are all on business matters; no family letters have been found, and it is likely that they were (as at Kellis) written on papyrus. As is noted elsewhere, House Bl had a relatively high level of humidity at its lower levels, and very little soft organic material was preserved in its deposition levels as a result. There was therefore no papyrus found in the house. Area 4 is also relatively damp, and organic material from that area has been mainly limited to wood, rope, and carbonized remains. Area 1.3 is dryer than the other two, and more organic materials have been found there, but no papyrus was uncovered. An unusual feature of the finds is the occurrence of one clay tablet in each of House B2 (24) and House Bl (62). Only one other such tablet from Graeco-Roman Egypt has been published to date, the tablet from Kellis published as P.Bingen 116 (another has since been discovered there). The editors there cite one clay tablet from Susa with a Greek text as the only previous instance of a Greek inscription on clay tablet. They note also that clay tablets are known from the Dakhla Oasis, namely from the Old Kingdom finds at Balat. Like the example from Kellis, the two published here are both accounts. Why one would use a clay tablet when potsherds were widely available is not immediately apparent, given the greater bulk and weight of the tablet. It seems to us most likely that their advantage, as was still more the case with wax tablets, lay in the possibility of smoothing out the tablet and writing on it more than once.

Dating A primary question posed by the papyrologist and the archaeologist alike to textual material found in any structure is its potential contribution to dating the construction, use, and abandonment of the building. In the present work, we will concentrate on the case of House Bl, because the material is abundant but complicated. We will argue that the final occupational levels belong to the reigns of Constantius and the following years, probably into the first years of Valentinian I. There are no usable archaeological contexts in Area 4; material excavated there through the 2007 season has in all cases been thoroughly disturbed by the activities already mentioned. For Area 1.3 the evidence is 3 Information from Colin Hope. 4 Many are not readable or are only partially legible. To preserve the integrity of the archaeological record we have included even the illegible here. We have found that some texts have become intelligible only when parallels were found in a later season.

INTRODUCTION

15

too slight to support much precision, but we will note in a few cases evidence that points to a range from the second quarter of the third century to the first part of the fourth. A considerable number of the ostraka from Area 2.1 contain numerals followed by the sign in the shape of an L that is widely used in papyri for (έτους), 'in the year'. The numbers run up to 32, a level reached in Roman history down to the apparent abandonment of Amheida only in the reigns of Augustus, Commodus, and some of the fourth-century emperors. Neither archaeological context nor palaeography will permit a date under either of the first two, and the fairly abundant coins leave no doubt that the overall context is in fact of the fourth century. Although context is lacking, we should also mention that a year 33 appears in 127, from Area 4.1. But a problem arises immediately, in that there is never any second or third (or fourth) regnal year mentioned, as is typically the case in fourth-century documents using regnal dating (see CSBE2 43-45 and 224-51). These facts initially caused us some hesitation in interpreting the L sign as an indication of year numbers, wondering if it might instead have some other significance. The sign is, for example, often found meaning one-half (ήμισυ). But we eventually excluded all such alternative meanings. There is, moreover, enough evidence to show that the sign in fact has its normal meaning as "year" in these texts. Although this may seem obvious to papyrological readers, we will set this evidence forth so that there may be no doubt. First, there are many texts in which the L sign precedes the numeral, in the manner familiar when we have the start of a dating clause. Of these, 114 is particularly probative, because there the numeral is followed by the sinusoidal curve and diagonal stroke as so commonly in fourth-century regnal dates: LieS/. Second, in several cases (e.g., 104 and 148), instead of L the numeral is simply followed by S. The only other word that can be rendered both by L and by S is ήμισυ, half. It is not probable that our tags record only amounts ending in Vi. Third, in one case (149), the numeral and sign are followed by a month and day. These facts seem to us cumulatively to exclude the possibility of any other interpretation. As a basis for the following discussion, we set out the complete list of these dates from texts found in Area 2.1 in tabular form. "Hyd. Pmoun?" indicates whether the text has the characteristic formula beginning υδ(ρευμα) Πμο(υν) found in many of the tags. "Yr lst/last" indicates whether the year appears at the start of the ostrakon's text or at its end. Year

Ostrakon

Room

Strat. Unit

1 1 2 2 3 3? 4 4 4

244 235 146 147

12 12 15 4 1 9 8 9 15

245 245 157 214 53 212 F74 129 157

100 213 102 103 32

Pmoun or Hyd. Pm.? X X Hyd. Pm. Hyd. Pm. Pm. Pm.? Pm. Pm. X

Yr first/ last first first last last last last last first first

Year symbol aS/ aS/

Lß ßL

yL yS 5L 58 5S/

16

OSTRAKA FROM TRIMITHIS Year

5 5 6 7 7 7 7 7 7 7 9 9 10,11 11 11 13 13 13 13 14 14? 14? 14 14 14 15 15 15 15 15 15 16 16 16 16 16 16 17 20 20 20 20 20

Ostrakon

Room

Strat. Unit

104

4 14 12 10 8 4 10 9 4 3 3 8 4 9 10 4 4 9 9 9 9 14 9 9 11 8 4 3 9 9 12 8 4 1 1 9 4 4 9 9 9 4 9

29 211 245 191 102 29 142 132 214 22 93 102 35D+G 132 142 35A 35A 167 200 151 127 194 218 200 235 98 35C 36, F38 212 212 245 102 35D+G 168 168 218 214 214 104 104 107 23 128

270 243 241 196 175 202 106 142 357 148 108 149 110 111 178 179 150 223

151 152 154 96 112 233 115 113 114 116 215 242 117 183 118 119 121 122 123 155 156 125 124 157

Pmoun or Hyd. Pm.? Pm. X X X X X X Pm. X ?

Hyd. Pm. Pm. Hyd. Pm. Pm. Pm. X X Hyd. Pm. X Hyd. Pm. Hyd. Pm. Hyd. Pm. Pm. Pm. X Pm. Pm. Pm. Pm. Pm. X Pm. X Pm. Pm. Pm. Pm. Pm. Hyd. Pm. Hyd. Pm. Pm. Pm. Hyd. Pm.

Yr first/ last last last first first first first last first first first last last both first first last last last last last last last first last first last first first first first first first first last last first first last last last last first? last

Year symbol

ES ES

rs// ζ8

LÇ LÇ ζ8/ ζ8/

LÇ ζ8

es

BL iL, iaL laS iaS/

lyL lyL ,yL lyL i5L i5L i5L? Li5 i5L Si8/ ιε8 Lie

LieS/ Lie IEL

IES/

Lir Lir irt irL

Lir Lir ,CL KL KL KL KL KL

INTRODUCTION Year

Ostrakon

Room

Strat. Unit

20 21 21 22 22 23 23 25 25 25 27 28 30

158 160 159 161 228 162 194 163 164 231 165 166 126?

10 11 4 4 10 4 3 1 1 9 8 1 4

30? 32

214 167

13 9

17

191 79 9 12 220 9 95 4 4 232 63 4 35

Pmoun or Hyd. Pm.? Hyd. Pm. Hyd. Pm. Hyd. Pm. Hyd. Pm. X Hyd. Pm. X Hyd. Pm. Hyd. Pm. X Hyd. Pm. Hyd. Pm. Pm.

Yr first/ last last last last last last last first last last first last last last

207 151

X Hyd. Pm.

first last

Year symbol

KL KaL xaL KßL Lxß xyL KyS/ xeL xeL xeL κζΕ Kj|L

λ[ (uncertain if year) AS LAß

Table 1: Regnal Year Dates from Area 2.1 In order to have the material from the other areas available for comparative purposes, we provide a table also with these: Year 1 1 2 2 2 4 5 8 10 10 10 14? 30 33

Ostrakon 200 203 97 98 99 101 105 107 109 195 199 153 230 127

Area 1.3 4.1 4.1 4.1 1.4 1.1 4.1 4.1 4.1 4.1 4.1 4.2 1.4 4.1

Strat. Unit 19 64 61 77 3 4 61 43 99 11 43 13 1 18

Pmoun or Hyd. Pm.? X X Pm. Pm. Pm. ? Pm. Pm. Pm. Pm. X X Hyd. Pm. ?? X Pm.

Yr first/ last first first first first last last first first first first first first first first

Table 2: Regnal Dates from Areas 1 and 4

Year symbol aS aS/ β// β//

Lß 5L ES/

r|L .S/ iS/

iS// iSS/ AS AyS

18

OSTRAKA FROM TRIMITHIS

Figure 1. Map of the Dakhla and Kharga Oases. Based on ASTER GDEM data, a product of METÍ and NASA.

INTRODUCTION

Figure 2. Plan of Amheida, 2010.

19

20

OSTRAKA FROM TRIMITHIS

We will first look at these data from the point of view of formal criteria and patterns, then passing on to a stratigraphie analysis that takes the coins into consideration.5 The tabular presentation above gives an orderly impression, which might lead one to think that we are dealing with a consecutive span of years from years 1 to 32 (or 33 including the text from Area 4). But that impression is deceptive, as we shall see. It was noted in CSBE2 44 that the use of single regnal dates, omitting the numerals for emperors other than the senior one, is first attested in the Kellis ostraka, with O.Kellis \ and 2 giving us years 10 and 11 for what elsewhere would be called 10-9-2 and 11-10-3 under the First Tetrarchy Such references are generally retrospective, rather than referring to the years currently in progress, in their first occurrences in papyri from the Nile valley, as they probably are in these Kellis texts. Even such dating is, however, rare until after 308, when experimentation with shorter datings and the use of Galerius' years as a kind of era, persisting after his death, comes to be visible. For the Trimithis tags with years from 10 up, then, we have in theory a choice among Diocletian, Galerius, Constantine I, Constantine II, and Constantius II, although not all of these were senior emperors in all of the years in question. But, to complicate matters further, (a) Diocletian's regnal years do not go past 21, and posthumous reckoning by Diocletian did not begin until much later; and (b) the posthumous use of Galerius' years is not attested after year 26 (317/8). These facts raise doubts about the advisability of dating the Hydreuma Pmoun series, which runs up to year 32, under either Diocletian or Galerius (cf. CSBE2 8-9), even if we did not have the evidence of the coins and of stratigraphy. Constantius II, on the contrary, was senior emperor only from his year 17. He cannot, therefore, be the emperor of the dates to years up to 16. Nor can Constantine II, who was senior emperor only from years 22 to 24. There are therefore insuperable obstacles in the way of adopting a chronology placing all of the tags in a single sequence from 1 to 32 (33). The possibilities for each range of years is as follows: Year(s) 1 2-9 10-13 14 15-16 17-21 22-23 24 25-31 32-

Possible senior emperor(s) Diocletian (or earlier) Diocletian; Valentinian I Diocletian; Constantine I; Valentinian I Diocletian; Constantine I Diocletian; Galerius; Constantine I Diocletian; Galerius; Constantine I; Constantius II Galerius; Constantine I; Constantine II; Constantius II Constantine I; Constantine II; Constantius II Constantine I; Constantius II Constantius II

It must be said that we lack good evidence for the a priori likelihood of such regnal dates under Constantine and Constantius. Regnal dates under Constantine are until now known only from the Oxyrhynchite and Arsinoite nomes (CSBE2 250 gives a list). None are known at present from years 27 and 28, but that is probably just a quirk of the 5 Only coins and ostraka offer sufficiently precise evidence for dating the stratigraphy at Amheida. Pottery cannot be dated exactly enough to help in this respect.

INTRODUCTION

21

documentation, because dates to years 29 and 30 (with long trains of additional numerals, to be sure) are known; year 30 is missing, but 31 is again present. Given our ignorance of oasite practice at this time, we certainly must reckon with the possibility that the calendarical conservatism of the Dakhla Oasis extended to retention of regnal years at a time when few other nomes kept them. But we have had until now no certain evidence to this effect, either. The chronological distribution of the published Kellis papyri, unfortunately, prevents them from telling us anything very precise about usage of regnal years versus indictions in Dakhla in the early fourth century; the first exactly dated papyrus using an indiction is P.Kellis 1 G. 30, with a seventh new indiction mentioned in 363. The last regnal date in the same volume is of year 15-3-1 of Galerius and colleagues (306/7), P.Kellis 1 G. 40.6 That is quite a gap in the evidence. Kellis papyri dated to the intervening period have dates by consulates; we do not have the kind of short text that would have been likely to use either regnal year or indiction to identify a tax year. The reason for the general dearth of regnal years under Constantine and following emperors in the Nile valley is certainly the adoption of consular dating for formal dating clauses and indictional dating for short reference to years. Indictions never appear in Kellis papyri from before mid-century They do appear in small numbers in the Kellis ostraka, in no case securely datable to a particular year. It is likely that O. Kellis 24 dates before the middle of the century, but for none of the others is there any indication.7 Indictions are also uncommon in the Trimithis ostraka. But we do have dates to indictions 1, 2, 5, perhaps 7, 13, and 15 (see Index IB). Whether these in fact fall in that order or should be put in the order 13 to 5 (or 7), a shorter span, is impossible to say. It seems generally likely, from the evidence of Kellis, that these do not date before midcentury but each stratigraphie unit must be analyzed independently. It is striking that Constantius II's regnal year 33 (356/7) coincided with the last year of an indiction cycle. A tidy mind might suppose that the oasis began using indiction years with the start of the cycle beginning in 357/8. But O.Kellis 24 suggests that this is not true in any absolute sense. Overlap of the systems is very likely. It is thus by no means to be excluded that the indictions 13-15 attested in the Trimithis ostraka belong to the years 354-357. Alternatively, they could belong to 369-372. It remains up to this point in the argument unclear to which emperor most of the dates are to be assigned. The latest tag from Area 2.1 and the latest from Area 4.1, from regnal years 32 and 33 respectively, can only belong to Constantius. But there is no hypothesis in which the single sequence can belong to one emperor, as we have indicated. We must of course notice the fact that there is no evidence for the lifetime use of Constantius' regnal years anywhere else in Egypt until his year 32, when Julian was made Caesar and a count of 32-1 was adopted. This count is attested in Oxyrhynchite documents beyond the death of both emperors, becoming for reasons still unknown to us the basis of the Oxyrhynchite era (CSBE2 55-62) and of an era used in the Small Oasis that uses Constantius' years but not Julian's.8 Constantius' posthumous regnal years are 6 Given that the instances of use of Galerius' posthumous count as regnal years (rather than as indiction numbers) are all Arsinoite, identifying any numbers higher than 19 (the year at Galerius' death) as belonging to him seems most unlikely. 7 For the date of O.Kellis 24, see line 3n. 8 As Klaas Worp points out to us, the combination of year 87 with indiction 9 in O.Dor. 2 = SB 20.14919 is possible only if the years of Constantius are the base; no other plausible emperor's year 87 would coincide with indiction 9. (There are problems of reading in this ostrakon, but the chronological data

22

OSTRAKA FROM TRIMITHIS

not attested so far in use in any other part of Egypt. Taking all that into consideration, however, there is no reason to doubt that the latest dates are by Constantius. The chart of regnal years above shows that in principle one might assign almost any regnal year to one of two or more senior emperors; only the lowest numbers and the highest are more constrained. However, multiple factors, even apart from stratigraphie evidence, make random assignments impossible. One of these is the coherent series of tags beginning with the phrase Hydreuma Pmoun. All certain instances come from Area 2.1, all give the date in last place, all but one uses the L symbol for year, and only one out of 20 places the L symbol before the numeral rather than after it. This uniformity entitles us to look for as economical a hypothesis as possible to explain the distribution in time of this group. It is particularly striking when one looks at the lack of consistency in the group of texts mentioning only a Pmoun, where the placement of the date at the start or end appears to be random (6 in each place), or in the group of texts not mentioning a well at all, in which the placement of the date at the start outnumbers placement at the end by 13-9. (In tags from other areas, first place outnumbers last overall by 12 to 2.) One other factor needs to be taken into consideration, which is that Egyptian documentary papyri mentioning indiction years sometimes use the same sinusuoidal curve or single or double stroke used to designate regnal years, without the word ίνδικτίων. (This is not true of the L sign.) It will be evident that a significant number of years up to 15 in the table use one of these markers, whereas almost all of those over 15 use the L sign. Year numbers 1-5 6-10 11-15 16-20 21-25 26-30 31+

L sign 6 6 14 12 9 2 1

S or / sign following 12 9 5 0 1 1 (+1?) 1

The very small numbers at the upper range may not mean much, unless to suggest that familiarity with the sinusoidal curve with indiction numbers was causing it also to be used more commonly with regnal years. But the concentration in the lower range suggests that we must consider seriously the possibility that those numbers are indiction numbers when followed by S. This point is taken into consideration in the stratigraphie discussion below.

are not in doubt.)

INTRODUCTION

23

Figure 3. Plan of Area 2.1, House Bl.

Stratigraphy and coins in House Bl At this point we turn to the evidence of stratigraphy and the coins, which we believe will allow us to assign the Hydreuma Pmoun sequence and much else to the reign of Constantius. The descriptions of the stratigraphie units will be set out in more detail, and with ceramic evidence as well, in the publication of the house of Area 2.1, which is designated House Bl, including the many stratigraphie units where evidence is lacking (i.e., there were no readable coins or ostraka with dates) or where the stratigraphy itself is unclear. Here we set out only those contexts where reasonably clear information emerges and helps to date levels and their contents.

24

OSTRAKA FROM TRIMITHIS

Room 1: This is the central reception room with mythological paintings. This room's occupation layers are not easy to read, because the domed roof, along with upper parts of the walls, collapsed into a layer (DSU 4) containing dome bricks and debris left on the roof as well as occupation debris on the floor at the time of collapse. DSU 4 contains several ostraka, including 163 and 164, both dated to year 25 and using the Hydreuma Pmoun formula. 166 is similar and from year 28. This material is probably from the roof and represents occupation debris. The same unit also yielded a letter of Gelasios to Herakleios (295). Gelasios is securely linked elsewhere to the last period of occupation. Another unit containing collapse material, closely tied to DSU 4, is DSU 53. This, however, is made up of rubble from the upper parts of walls, triggered relatively quickly after abandonment by removal of the lintels of the doors to Rooms 14 and 11. Ostrakon 100 belongs to this unit. If it comes from occupational debris, one would suppose that the year 3 mentioned in it belonged to Valentinian. But the ostrakon is badly effaced, and it seems more likely that it was a chinking sherd in the walls that collapsed; it would in this case belong to the reign of Diocletian (286/7), and this seems to us more likely. Beneath the collapses is the room's last and only surviving floor, F43. DSU 164, underneath that floor, produced a coin datable to 355-361. DSU 164 is in turn at a higher elevation than DSU 168, from which come 118 and 119. These both have dates to year 16 and probably belong to the ashy fill used in making a floor. As there is only one floor level surviving in Room 1, it seems likely that this is a replacement floor, laid when the house was already in use and after removal of the remains of any earlier floors. Year 16 (118 and 119), as a vintage year, probably does not in any event belong to Constantius II, because he became senior emperor only in the last part of that year after the death of Constantine II in 340. If it belongs to Constantine I, it would date to 321/2. Conclusion: 321/2 is thus probably a terminus post quern for the deposit of the underlying ash layers from construction. The surviving floor, however, is probably a new one laid after 355, to judge from the coin. The occupational material above that floor is therefore from the reign of Constantius or later; years 25 and 28, from 163, 164 and 166 above, seem certain to be his. Room 2: Room 1 was entered from Room 2, which probably had a flat, lightweight roof; no vault collapse was found. DSU 13 was found just above the floor level. It contained a bit of mud-brick debris and some clean sand, together with some ostraka. Above it was a 2meter thick layer of windblown sand (DSU 2) sealing the context. DSU 13 is thus to be seen as occupational debris of the period just before abandonment. Most of its ostraka were found in a group below a niche in wall F3. These included the memorandum mentioning a warden of foreigners (252), and three letters of Serenos, two of them to Philippos (297, 298, 299). Philippos also appears in a text signed by Nikokles (358) from this context. This cluster of individuals is consistently associated with the final period of occupation of House Bl. None of the ostraka from Room 2, however, contains a regnal year. The other stratigraphie unit to contain significant numbers of ostraka was DSU 34, the remains of the second floor level from the top, i.e., a floor before that of the last period of occupation but belonging to the habitation history of Bl. Unfortunately, these were mostly poorly preserved and uninformative. Only 177 is perhaps of interest, as one

INTRODUCTION

25

of the group of tags with alpha followed by the name Psennesis. Similar tags also occurred in lower layers in the same sequence, DSU 38 (180) and 41 (188). These ostraka suggest that the sequence of layers here was of fairly compact duration. Room 2 did not yield any coins to assist in assigning an exact date to any of the levels. Conclusion: The data from Room 2 do little to allow secure dating of the finds therein. Room 4: Room 4,9 to the north of Room 2, was a vaulted room. The uppermost layer of material in this room came from DSU 9 and 10, which are to be taken together, as the similarity of 249 (DSU 10) and 250 (DSU 9) suggests. This context contained mud-brick collapse from the roof and vaults and thus could in principle contain both occupation material and fill or chinking sherds. There are two dated ostraka in DSU 9,159 of year 21 and 162 of year 23. Both ostraka use the Hydreuma Pmoun formula. DSU 23 is the uppermost part of a series of floor levels in this room, that is, probably the last occupation level. (The sequence of the floors is, from top to bottom, 23, 35A, 35C.) It contains debris including datable material. Of 16 probably datable coins, 12 belong to the reign of Constantius (337-361). One belongs to the last years of Constantine (306-337), and one is out of range entirely (Neronian). Two come from the reign of Valentinian I (364-378). Seventeen ostraka were found in this unit. The only tag with a year date is from year 20 (124). There are, however, ostraka mentioning Serenos (300: "to the house of my lord Serenos"), Domnion (253, 279), Gelasios (302), and Nikokles (9, 279). Serenos and Nikokles also appear in the last occupational level of Room 2 and Gelasios in the final level of Room 1. Below DSU 23 was DSU 35, a subfloor with several likely subphases. The one dated coin, from 351-361, belongs to DSU 35A, the highest level, with an upper elevation of 136.85 meters above sea level (abbreviated below as m.a.s.l.). The ostraka are dated as follows: - year 13 (178 and 179). Their elevations are 136.754 and 136.729 m.a.s.l. - year 15 (113). Its elevation is 136.619 m.a.s.l. - years 10 and 11 (149), 16 (183). Their elevations are 136.428 and 136.433 m.a.s.l. - Demotic ostrakon of late Ptolemaic or Augustan period (278) at 136.378 m.a.s.l. All of these fall below the bottom of the room's last floor, DSU 23, at 136.80 and also below the stated tops of 35A and 35C. This entire deposit should be considered to belong to a pre-floor phase and of mixed character (like DSU 214). The year sequence 10-16 is compatible with Constantine I but not with Constantius II, who was not senior emperor during those years of his reign. The one coin mentioned above is higher than any of the ostraka in elevation, and it could perhaps belong to occupation debris or have been (like the coin in Room 1) deposited in the making of the last floor. DSU 29 is a layer of fill behind a low wall (F19). It contained six coins, three with certain dates 335-348, and another perhaps dated 337-347. There are ostraka with dates to years 5 (104) and 7 (175). As this context represents fill, it is perhaps not all of a consistent date. It should be noted that 104 uses the S indicator and could refer to an indiction. 9 Room 3 belongs to a different house, to the south of Bl; it has therefore been omitted from the discussion of the stratigraphy of Bl.

26

OSTRAKA FROM TRIMITHIS

DSU 214 is a foundation fill, thus consisting entirely of material from before the construction of House Bl. It yielded 10 ostraka, including from years 2 (147), 7 (142), 16 (122), and 17 (123). That this material is not a great deal older than much of what is in DSU 35 is suggested by the presence in both of tags giving only a number followed by a name: Seiles, in DSU 35 (182, 186, 187) and 214 (217, 219); Horos, in DSU 35 (181, 185) and 214 (218). This reinforces the view that most or all of what is in DSU 35 dates before the construction of the house. Conclusion: Despite some difficulties in understanding the sequence in DSU 35, it looks as if material from the reign of Constantine I or before all belongs to preconstruction debris, while all material from Constantius II or later (i.e., AD 340) belongs to occupational layers. The undated material in these contexts should be assigned to the same periods. Room 5: This is the staircase to the roof. Few ostraka were found here, and none of them with dates. In DSU 213, a Serenos is mentioned (271). This is collapse material at the foot of the staircase and of uncertain relative date. Also from this context came 318, in which an amount in talents is mentioned, probably from the second quarter of the fourth century. It is possible that this comes from a chinking sherd in collapsed wall, but it cannot be excluded that it came from the roof. Few ostraka, none with useful contexts or information, were found in Rooms 6 and 7 on the east side of Bl. Room 8: This is a room just to the east of Room 4, on the northeast corner of the house, which contains evidence of food preparation. The uppermost unit of significance is DSU 63, which contains vault collapse. As usual, this could in principle reflect either occupational debris or chinking sherds. One ostrakon of this unit, 165, contains a date to year 27 that is likely to be occupation debris. There are no coins. DSU 85 (a wall collapse) contains the account 19, which has amounts of money that must be later than 351 and probably not later than the start of the 360s. It was found on the threshold between Room 8 and Room 7 and is evidently occupation material. There are again no coins. DSU 74 (a matrix of sand and ash just above the floor F64) contains a number of coins, including two datable to 350-355 and "before 361". That is compatible with the thousands of talents mentioned in 16, one of the 6 ostraka from this unit, which must postdate 351. The other texts from this unit have no dating information, but it is worth noting that 63 is a ration account of a type encountered in Rooms 13 and 14 and discussed there. DSU 98 (below floor F64) and 102 have no legible coins. DSU 98 has one date to year 15 (115), and DSU 102 has dates to year 7 (196), 9 (108), and 16 (117). The spectrum is overall like that in DSU 35 in Room 4, and like it this context appears to be fill. Conclusion: In Room 8 also the pre-Constantian material belongs to fill below floor level, while material from the late 340s into the 360s dates to the occupation levels. Room 11: With Room 11 we begin the series of rooms on the west side of House Bl. It is the room painted with a green decorative pattern. It produced just 8 ostraka. DSU 79 (sand filling

INTRODUCTION

27

above collapse) includes ostrakon 160, dated to year 21 and using the Hydreuma Pmoun formula. This should, by comparison with other rooms, come from the roof, the collapse from which is below this sand deposit. As a superficial sand deposit it is not very secure. DSU 235 (collapse, mainly from the eastern wall F16 and vault) is more secure and was more productive but the materials are unfortunately not easy to date. This DSU has a text dated to year 14, referring to the vineyard of Horigenes (233), one mention of a delivery to the oikodespoina (mistress of the household) (273), and one text with Mouses son of Psais, who also appears in a text from DSU 194 in Room 14 (234). The year 14 text (of necessity from the reign of Constantine or earlier) would be surprisingly early for something from an occupation context, but of course this unit is, given its character as collapse, not necessarily homogeneous. An early text could perhaps come from a wall chinking sherd. The coins from the floor level that were datable belong to 347-363. One coin from DSU 260, just above floor level, dates tentatively to 364-388, too wide a span to be useful. Conclusion: the year 21 ostrakon 160, potentially from roof collapse, and the coins from the floor level are both consistent with occupation during the reign of Constantius. But none of the contexts producing ostraka were as well defined as one would like. Room 12: Room 12 is the west entryway to the house from the alley. It was undecorated. DSU 245 (the collapse of the north wall and north part of the roof vault) is one of the richest contexts in the house for ostraka. It is, however, not immediately easy to make sense of, as the year numbers seen to be widely spread: year 1 (235 and 244, both referring to a vineyard of Sarapion), year 6 (243), and year 15 (242). The name Ephrem (unlikely before mid-century we believe) shows up in 94. This is more likely a feature of the house's occupation period. But the early year numbers mixed in with it are surprising. The problem is analogous to that in DSU 235 in Room 11; the context could again be mixed, given its origin as wall and roof collapse. Another explanation is possible, however: the dates from years 1, 6, and 15 do not use the L sign for year, but are all numerals followed by the S sign (and sometimes a stroke). These could easily refer to indictions, in which case the years would probably range from 356/7 (indiction 15) to 362/3 (indiction 6). That this is the correct solution is indicated by 245, where the date is explicitly to the 15th indiction (ieS// ίνδ(ικτίονος)). These years then fall after the last known regnal date from the house, year 32 of Constantius, which equates to 355/6. Conclusion: The datable occupation debris from the room is all likely to come from the period approximately 356-362. Room 13: This room is the northernmost of the series of rooms on the west side of House Bl. It was decorated with a series of panels in purple. It yielded in all 32 ostraka. The context of the surface layer DSU 185 is insecure, but it is noteworthy that 261 mentions Psais son of Kapiton, known from Kellis texts (including the KAB) from the 360s (cf. 276, from Room 12, in which we restore his patronymic). That text is in turn closely similar to 272, found in DSU 199, which is vault collapse from Room 13 that fell outside its wall. Together these suggest a date in the 360s for the last material from the collapse. (DSU 199 contains other ostraka, but none have any chronological indications.) Another vault

28

OSTRAKA FROM TRIMITHIS

collapse unit is DSU 207, coming from above the doorway. Ostrakon 214, from this context, has a date to year 30 and a reference to a vineyard of Sarapion. It should be noted, however, that the lambda of the date is not read with absolute certainty. If this is year 30 of Constantius, it would date to 353/4. DSU 208 represents the fill of the northwest bin in this room. One of the ostraka found in this fill, 288, is signed by Nikokles, who is known from the final phases of Rooms 2 and 4. DSU 209 is the fill of the northeast bin. It contained a ration account (64). DSU 216 was a layer of sand on floor F150, the latest occupational level in the room. The dated coins here are placed after 364. The clearest sign of a late date here in the ostraka is the 2ndindiction in 41. This indiction 2 is unlikely to be earlier than 358/9, and this indeed seems the most likely year. There is also one signature of Serenos (290), thus also of the final phase. In 42, an account of tiphagia, talents in the thousands are mentioned: this thus dates after 351. One coin from F150, the floor, is assigned to 347-8 and a second with less certainty to 355-363. That is consistent with the picture elsewhere. Conclusion: The picture in this room is clearly one with material from the 350s and 360s in occupation deposits and perhaps the stage of abandonment as well. There do not appear to be any contrary indications. Room 14: Room 14 is the southernmost of the west rooms, the one painted with a red decorative motif. DSU 194 is a layer of collapse. A year number appears in 154, but the reading is very doubtful; it seems to be a 14 followed by the year sign and then a 1. We do not understand this. Other ostraka from this DSU are not very helpful. This is a context in which material could belong either to masonry or to deposits on the roof, and is thus not very secure. DSU 211, which is, with DSU 233, floor debris, produced a year 5 in ostrakon 270. This could possibly be year 5 of Valentinian, but more probably it is an indiction number, as the marker is the S curve. Nothing in the contents of the ostrakon is distinctively linkable to the final occupation phase. The date would thus be either 367/8 or 361/2. DSU 221, which certainly belongs to the room's occupation level, has mostly names, but in 66 we find Domnion and Gelasios in a ration account. The coin dated from this context is from 364-378. It seems likely that this context is in fact part of DSU 233. DSU 233 is dominated by Domnion. He appears in 323 (with Psenpnouthes), 324 (with Antinoos and Psenpnouthes), 293 (with Paesios), 292 (with Paesios), and 322 (with a camelarius and a centurion). 65, 67, 68, and 69 are all ration accounts. Antinoos also appears in 88. The numerous coins have mostly not been dated with much certainty. Three are dated before 340, 7 from 347-361, and 3 probably from after 364 but without a secure terminus ante quern. Conclusion: Apart from some uncertainties in dealing with collapse, the picture in this room is again consistent with the occupation levels belonging entirely to the reigns of Constantius and his immediate successors. Northern Annexes of House Bl The courtyards numbered 9 and 10, and Rooms 15 and 16, belong to a structure that was originally not part of House Bl. The bulk of this area was occupied by a multi-room

INTRODUCTION

29

school. After the school ceased to function, the rooms were added to the house, and a passage cut from Room 17 through the north wall of the house into a short passageway numbered Room 16, which in turn leads into Room 15. Another area was turned into the work areas called by us Rooms 9 and 10. Because Room 17 is for our purposes part of this late phase northern complex, it is also included below. Room 15: This is the schoolroom with rhetorical verses on the wall, then turned into a storage magazine.10 The most productive unit is DSU 157, a layer of occupational debris intermixed with bricks. This filling was accumulated when the baked brick piers supporting the raised floor of the magazine period were already in place. Low year numbers appear in 146 (year 2) and 32 (year 4; with an S curve this could be an indiction). If both regnal years, these could indicate 363/4 (2 Jovian) or 364/5 (2 Valentinian) and 366/7 (4 Valentinian); a fourth indiction might be 360/1. The contents are otherwise unhelpful for dating. The coins from this DSU date to 347-8 and two tentatively to after 364. That they belong to the last phase of the house is consistent with the situation elsewhere. From DSU 152 = 145, a layer of flat roof collapse on occupation debris, come 281 and 283, both signed (it appears) by Serenos. Conclusion: Room 15 has not supplied any evidence from ostraka or coins useful for dating its original construction and use as a schoolroom. Its final period of reuse as a magazine appears to correspond to that of the rest of the house, i.e., roughly the late 350s and 360s. Room 16+17: Joins of fragments of ostraka (282, 312) show that DSU 153 (wind-blown sand), 155 (sand and dust above floor), 158 (wind-blown sand over floor but below collapse), and 161 (sand and dust above floor) are all of a single chronological horizon. They are decisively of the latest phase of the house. 205 speaks of hundreds of talents and mentions Tou son of Ploutammon, who also appears in 249, in a similar context in Room 4; the sum of 500 talents appears in 282, where we find Nikokles. Nikokles also appears in 312 and 313. We have already seen him linked to other key figures in other rooms. Serenos is the author of 310 and recipient in 311. A priest (presbytères) Makarios appears in 27. Conclusion: Overall, the chronological horizon seems likely to be the end of the house's occupation and in any case after 351, given the size of amounts in talents. Courtyard 9: The central area of this courtyard shows a sequence from the top of windblown sand (DSU 68 + 114), with no useful dating material; a floor (F107, with deposits DSU 126 and 131), also with no ostraka or coins; and a large body of dumped material apparently laid down under the floor at the time of construction (DSU 128, 107, 104, 127). This material includes ostraka with dates to year 14 (152) and especially year 20 (125,155,156,157). If these belong to the same sequence, they must be Constantinian and thus 319/320 and 325/6. The same layers also included 347, with a Christian symbol, the chi-rho. The pit filling DSU 132 is probably slightly later than this dump layer; it contained ostraka 10 See Raffaella Cribiore, Paola Davoli, and David M. Ratean, "A Teacher's dipinto from Trimithis (Dakhleh Oasis)," Journal of Roman Archaeology 21 (2008) 170-92.

30

OSTRAKA FROM TRIMITHIS

dating to years 7 and 11, most likely Diocletianic (290/1 and 294/5), although year 11 could also be 316/7 (106 and 110). But it should be noted that in both cases the year indication is S rather than L, and it is thus not impossible that we are dealing with indiction years, if the chronology of the pit is insecure. Under this layer was a sand layer (108 = 129) with one ostrakon bearing a date to year 4 (103). (This uses the S-curve and could refer to an indiction.) Still lower are layers of ash, probably deriving from the waste from the public baths that stood in this area before the construction of the school and House Bl. Year 4 would seem likely to belong to 287/8, but as dumped material it cannot be placed accurately with respect to the period of the bath's use. In the southeast corner of the courtyard, we find another sequence, at a somewhat higher level. The top dump levels, DSU 117 = 100, contain no coins or datable ostraka. The same is true of the mud floor 133 = 131, but there is one ostrakon reading "Neilos son of Neilammon" (201). Underneath is another dump layer DSU 138, in which appears another ostrakon referring to the same man (394). It is noteworthy that Neilos appears yet again in 201a (inv. 11561), coming from DSU 130, recovered in 2008 from pottery found in 2006. In the west part of the courtyard we find a sequence of windblown sand (DSU 114), dumped material (DSU 118 = 146 = 116, DSU 148 = 195), then the mud floor F107. Nearby dump layers DSU 151 and 167 were attributed to the same level as DSU 118 etc. This dumped material above the floor was taken to belong to the period after the closure of the courtyard and as dump is likely to be of little chronological value. It included ostraka with dates in years 13 (150), 14 (151), and 32 (167). This is not particularly helpful, given its source as debris, and the fact that these years must come from the reigns of different emperors. Under the floor were several layers of pre-construction dumped material, including DSU 196 and 197; 200 and 203; and 206, 212, 218, 234, and 237. Debris from the various sub-floor levels of dumping is somewhat more consistent in character, with dates to years 10 (144), 13 (223), 14 (96,112), 15 (116, 215), and 16 (121). In 213 there is a date ending in 3, which could well be 13. These dates seem most likely to be Constantinian as elsewhere just below floors, and to range from 315/6 to 321/2. Finally and perhaps most interestingly, DSU 232 was fill in the foundation trench for the wall F22 + F19 + F52 (the north wall of the house). It contains a date to year 25 (231). That would presumably be Constantinian again, and refer to 330/1. That is the latest secure indication of a terminus post quern for the construction of the house discovered so far. (Construction of the house after Constantius' year 25, or 348/9, does not appear consistent with the other evidence.) Courtyard 10: Almost all of the datable evidence from the excavation of Courtyard 10 comes from material dumped under the floor F106, corresponding to the pre-construction material in Courtyard 9. This material, DSU 142, DSU 191, has a familiar range, with dates to years 7 (202, 241), 11 (111), and 20 (158). It presumably dates from the period down to 325/6, although 241 and 111 use the S year indicator and could refer to indictions as well. Horos son of Mer sis in 241 (dating to year 7) may be the same man appearing in P.Kellis (see below, p. 34) in the 320s and 330s, in which case his appearance in 241, a text found under F106, would be an earlier attestation, from 290 (Diocletian's seventh year).

INTRODUCTION

31

Another DSU that goes back to the fill underlying construction of the complex is DSU 240, filling one of the channels from the bath. The clay tablet 62, relating to a pagan cult, belongs to this fill. It is also interesting that Neilos son of Neilammon appears in 394, from DSU 119, which is part of the debris above the floor level. Comparing the information from Courtyard 9, one may conclude that his activity straddles the period of the laying of this floor and the activity taking part in the area after construction. From DSU 220, debris contemporary with floor F106 or postdating it, we also have one tag referring to Nikokles (226) and an uncertainly read reference to year 22 (228). At least the first of these presumably comes from a final occupation phase, but the character of this unit is not clear enough to lead us to assign all of its contents to that phase. General conclusion: There seems to be nothing clearly deriving from the reign of Constantius in any level below the floor except for the remade floor in Room 1. There are a few earlier coins in occupation levels, but that is consistent with a normal pattern of older coins remaining in circulation and provides nothing more than a terminus post quern. All of the ostraka in occupation levels probably belong to the reign of Constantius, starting with his year 20 (343/4), except for perhaps a few belonging to the years immediately after Constantius' death; almost all such low numbers, however, are capable of being interpreted as indictional dates, meshing well with several certain or probable indictional dates from the late 350s and early 360s. The coins are consistent with this dating. At all events, it looks as if there is a horizon between 330 (year 25 in the foundation trench for F22+F19+F52, the north wall of the house) and 340 between pre-construction material and occupation material. Nothing in the final phase of occupation can be securely placed after about 365/6, year 3 of Valentinian, but it is not excluded that occupation continued for a few years longer. The one uncertainty in this reconstruction is the fact that dates of years 22-24 could belong to Constantine I, Constantine II, and Constantius II, while those of years 26-31 could belong to Constantine I or Constantius II. This uncertainty makes it hard to pin down the date of construction between 330 and 340. The collocation of dates to years 21 and 23 in Room 4 (DSU 9 and 10), however, excludes Constantine II (who was not senior emperor in his year 21), and if this context is in fact occupation debris, as it appears to be, it also cannot belong to Constantine I (because construction came in or after his year 25). The dates to years 25 and 28 in Room 1 could in principle belong to either Constantine I or Constantius II, but their location above a floor laid no earlier than 355 makes a date in the early 330s far less attractive than dates in 348/9 and 351/2. There is thus no evidence plausibly to be attributed to the decade of the 330s in occupation layers, making it likely that construction fell nearer the end than the start of the 330-340 window.

The Wells and Trimithite Geography Thanks to the well tags in particular, the ostraka from Trimithis provide considerable material on regional topography. At least 30 places are attested in the Trimithis ostraka for the first time, the majority of these in texts from Area 2.1. These place-names almost invariably begin with Πμουν (Pmoun), "the well" or "the water." Many ostraka, primarily the well tags described above, use the redundant phrase hydreuma pmoun.

32

OSTRAKA FROM TRIMITHIS

Only once does a well toponym appear solely as a hydreuma (also "the well"), with the pmoun absent.11 (See below, s.v. "Chorion of Psenamounis.") Pmoun toponyms are common in texts from Kellis as well.12 Presumably the word hydreuma means much the same in Trimithis as it does elsewhere in Egypt. According to the standard reference work on water management in Graeco-Roman Egypt, a hydreuma "n'est pas un appareil hydraulique; il peut être donné comme limite d'une terre; c'est une citerne destine à garder l'eau qui s'y accumule de façon naturelle... par des puits artésiens."13 The appearance in one Trimithis ostrakon, 53, of a well-cleaner or ύδροκαθαρτής confirms that actual wells are at stake. But the toponyms presumably refer not solely to the wells themselves but to cultivated areas around the wells, which may have had minor settlements named after them. As a vital part of the agricultural economy of the Oasis, it is no surprise that these wells comprise a large portion of the region's topography14 The names of the wells generally suggest an Egyptian linguistic context. We find wells named after Egyptian deities - Pmoun Horos and Pmoun Osire - and wells whose names are merely theophoric personal names - the water of the chorion of Psenamounis, Pmoun Taesis. We also find wells with a variety of Egyptian names more generally: Harau, Phib, Palou, Pkes, etc. Pmoun Beri, which appears in the Kellis papyri as well, is an historical descriptor of sorts, stemming from the Coptic βρρε, for "new, young."15 Pmoun Tsoa is named with a Coptic word, "to give to drink." Pmoun Psoi may be a topographical description, from Coptic φωι, "high, above." Choria, in contrast, the vineyard-orchard properties presumably named after their owners, can readily have names of a Greek or Roman origin, as in the chorion of Ammonios son of Klaudios. The majority of the well toponyms appear only once in our corpus. Pmoun Harau appears five times in Trimithis ostraka, in texts from House Bl of Area 2.1, the temple hill, and Area 1.3. The last of these (21) is an account including payment to a donkeydriver, perhaps for transportation of goods to and from the well. Pmoun Tpake appears three times in ostraka from House Bl. Tpake alone appears once in O.Kellis 59 and nine times in the Kellis Agricultural Account Book. The first appearance of Tpake in the KAB subordinates Tpake to Bait( ), which was in Mesobe. The account book's editor took Mesobe to be east of Kellis. Pmoun Tpake could have been there. Alternatively, since Tpake means "the town" in Coptic, this name could have been used more than once in the area of any town, including Trimithis. If only one Tpake is at stake, its appearance in texts from both Trimithis and Kellis may suggest a location between the two settlements. The description of Tpake in P.Kellis \ G. 35 as "of the city of the Mothites" (της Μωθειτών Ι [πόλεως]), that is, modern Mut, may help place the well near there, but equally may refer simply to the Oasis in general. 11 To this one case add the apparently unnamed hydreuma appearing in a Trimithis ostrakon from the 2008 season, inv. 13205, not published herein, which records payment of one animal ύπ(έρ) ύδρ(εύματο^) ίερατικ( ). A search of the Duke Databank does not return any other attestations of this phrase. 12 See, e.g., the references collected in Index VI(b) to P.Kellis 1. 13 D. Bonneau, Le Régime administratif de l'eau du Nil dans l'Egypte grecque, romaine et byzantine (Leiden 1993) 61. 14 G. Wagner, Les oasis d'Egypte à l'époque grecque, romaine et byzantine, d'après les documents grecs (Cairo 1987) 279-83. 15 Wagner, Les oasis, 163. This toponym has also appeared in an ostrakon from Ain el-Gedida, an Arabic name that reproduces the Coptic one. It is far from certain that there was only one place with this name in the Dakhla Oasis, however.

INTRODUCTION

33

Pmoun Psoi is by far the most frequently attested well in Trimithite ostraka to date, appearing in twelve ostraka from Area 2.1. Was it the closest of the wells to Trimithis itself, or the largest serving the city? The fact that Pmoun Psoi does not appear in texts from outside Area 2.1, at least with that spelling, may suggest that Pmoun Psoi had a special relationship to the large estate operated from the house associated with the Trimithite councilman Serenos. Most references to Pmoun Psoi are well tags, too laconic to help answer the question. Pmoun Psoi appears on the reverse side of an account naming an Aphrodisieion, perhaps that of the Small Oasis, which suggests association with an institution, perhaps a large estate, with some financial reach. Pmoun Psoi appears in texts ranging from year 3 to year 16, on tags cataloguing over half a dozen payers. Tithoes son of Thaesios and Ammonios son of Tithoes both pay from Pmoun Psoi; this could be a father and son, but the commonness of the name makes it impossible to know. Pmoun Pso is the next most frequently attested pmoun, appearing eight times in Trimithite ostraka, six times in texts from Area 2.1.16 Indeed, it seems likely that Pmoun Pso and Pmoun Psoi are the same. One attestation of Pmoun Pso is an unusual doublesided tag (155) from a year 10, the reverse of which mentions a Pmoun Όσιρε. This is the same well attested in a tag from the temple hill, a tag from Area 2.1, and an account from Area 1 detailing a delivery of 22 loads. That same account (22) includes another payment, of 15 loads, involving a Pmoun Έμβ(ωου). These overlaps may suggest the physical proximity of Pmoun Pso, Osire and Emb(oou) to one another. If Pso and Psoi are the same, Osire and Emb(oou) in turn are likely near Trimithis itself, or at least near the central properties of Serenos and family. Pmoun Tkele appears in 17, an account from Area 2.1, recording 6 donkey-loads of wood, 4 moia of chaff, et al. Pmoun Tkele also appears in the KAB, in which one of the references to Trimithis records the payment of the rent of Pmoun Tkele to Horigenes in Trimithis. Since the reference to Pmoun Tkele from Area 2.1 is not in the form of a well tag, we do not know if tenants were paying to the house of Serenos from Pmoun Tkele the same way they were from other wells. The overlap with the Kellis text may suggest some interaction between the Kellis estate and that of Serenos, but without more chronological information, we cannot assess the possibility. More likely, the two texts refer to multiple properties or places at the same location. One well tag from Area 2.1, 166, names a Pmoun Pamo( ), probably to be resolved as Pamo(nthes). This may be the same site as Pmoun Pam( ) in O.Kellis 96 and Pmoun Pamo in P.Kellis 1 G. 21.12, where reading an abbreviation may be possible. In a third possible connection, this may in turn be the same site as the Pmoun Pem( ) reported in 17 an ostrakon from Deir el-Hagar, due west of Trimithis. Trimithis is much closer to Deir el-Hagar than to Kellis. If the same site is at stake in all three cases, the location of Pmoun Pamo( ) was likely closer to Trimithis, or at least closer to central Dakhla than anywhere to the east of Kellis. A similar point may be made about Pmoun Phib, which appears in an account of days, 39; another Deir el-Hagar ostrakon mentions hydreuma pmoun Pheb, an otherwise unattested toponym which may be the same as Pmoun Phib. The site may be between Trimithis and Deir el-Hagar, or near to them both. The Trimithis ostraka attest to at least half a dozen choria or vineyards, several of which are associated with specific wells. One ostrakon from Area 2.1 (10) is a rent 16 Hydrological Pso names appear elsewhere in Egypt; cf. the médiane Pso in P.Oxy. 34.2724. 17 Anthony J. Mills and Klaas A. Worp, "Four Greek Ostraka from Deir el-Hagar," ZPE 146 (2004) 155-58.

34

OSTRAKA FROM TRIMITHIS

account "for the winter period" of the hydreuma of the Chorion of Psenamounis. The Chorion of Petosiris appears in texts from the temple hill and Area 2.1, in the latter case on a well-tag from Pmoun Hor( ). This is perhaps the same as Pmoun Horou, which appears in an account or name-list also from Area 2.1 (27). A Chorion of Sarapion son of Paleb appears in two well tags, one naming Pmoun Berri (114). The northern part of a 18 Pmoun Beri appears in P.Kellis l G 5.12, and may refer to the same locale. The Chorion of Pal( ) son of Pete( ) appears in an ostrakon found on the temple hill, a well tag naming Pmoun Thaut. Finally, two well tags from Area 2.1, 112 and 151, mention a Pmoun name - apparently different in each case - followed by reference to a χω(ρίου) I αύτουργ( ), "self-cultivated vineyard." Does this imply a contrast to the other choria, which might then have been worked by a large estate work-force rather than their tenant cultivators? If this is the case, then the personal names attached to these other choria might then be the names of their respective tenants. Finally, we note in 241 the appearance of an unread vineyard name described as "of Horos son of Mer sis." A camel-driver named Horos son of Mersis appears six times in the Kellis papyri, in a variety of contexts, e.g. purchasing half of a foal (P.Kellis 1.34), delivering a camel load of goods (P.Kellis 1.51) and as a neighbor listed in the borders of a land-grant (P.Kellis 1.38). The appearance of a vineyard related to the same personal name may indicate another topographical link between Kellis and Trimithis, although several decades separate the references (see above, p. 30). The following toponyms appear in ostraka from Trimithis: Well Names θαν (...than) - appears in 14, a memorandum or account of oil. A[ (A[) - appears in well tag 99. Άλεξάνδρο(υ) (Alexandros) - appears in an account of hay, 50. Αμε[ (Ame[) - appears in well tag 118. Άραυ (Harau) - appears in various payment tags and labels from Areas 1, 2 and 4; see 21,133,134,135 and 256. Βερρι (Berri) - appears in three well tags, 114,146 and 158, twice as a hydreuma pmoun. Γαλμεν (Calmen) - appears in one well tag from Area 4.1, 97. χωρίο(υ) Ψεναμ(ούνιος) (Chorion of Psenamounis) - appears in an account, 10, described only as hydreuma, not as a pmoun. E[ (E[) - appears in 27, an account of people from Pmoun Horou, one of whom, Makarios, is from E[. Έμβωο(υ) / Ένβω(ου) (Embôo(u) / Enbô(ou)) - appears in an apparent account from

Area 1.3, 22, and in a surface find from Area 1.1, 101, as Embôo(u), and appears in well tag 137 from Area 2.1 as Enbô(ou). May also appear in well tag 143, also from Area 2.1, one of only two tags to use "moun" without the Egyptian definite article P; cf. 144. Έρμ( ) (Herm( )) - appears in a well tag unusual for the occurrence of two year dates and a calendar date, 149. Identical with one of the subsequent Pmouns? ΈρμοΟ (Hermes) - appears in an unusual well tag, 172, lacking the word "pmoun," a year number and a proper name, but indicating the χω(ρίου) Διοφάν(ους), or chorion of Diophanes. To be associated with preceding or subsequent pmoun? Έρμο.( ) (Hermo( )) - appears in well tag 159 as a hydreuma pmoun. Perhaps to be 18 The same toponym appears in an unpublished ostrakon from Ain es-Sabil found in 2009.

INTRODUCTION

35

resolved as Hermod(orou). Identical to preceding Pmouns? Ήσε (Ese) - appears in a list, 77, with Pmoun Po.itou. Θατ( ) (That( )) - appears in well tag 167 as a hydreuma pmoun. Θαυτ (Thaut) - appears in 105, an ostrakon from Area 4 naming a vineyard of Pal( ) son of Pete( ). This well may also appear in Trimithis ostrakon inv. 13004 from the 2008 season, not published herein. Identical to preceding Pmoun? Θοτ( ) (Thot(eus?)) - appears in well tag 163 as a hydreuma pmoun. Ίβυθο( ) (Ibutho( )) - appears in well tag 162 as a hydreuma pmoun. Μαρκ( ) (Mark( )) - appears in an account of days, 39, with Pmouns Phib and Psoin. όνηλ( ) - appears in 22.3: presumably "of the donkey-driver(s)." Όσειρε/ Όσιρε (Oseire / Osire) - appears in well tag 139 as Oseire, and appears as Osire in a well tag from Area 2.1,119; in an account naming a Pmoun Emb( ) from Area 1, 22; in a tag from Area 4.1,171; and as a hydreuma pmoun in an unusual double-sided well tag also naming Pmoun Pso, 155. Also attested as Ousire in unpublished Trimithis ostrakon inv. 13223 from the 2008 season, appearing with Pselthis and pmoun Tsalem (see Tsamen below). Όστρακ( ) (Ostrak( )) - appears in well tag 152 as a hydreuma pmoun. Π ττφ ( )- appears in well tag 169 as a hydreuma pmoun. Πακ( ) (Pak( )) - appears in well tag 154 as a hydreuma pmoun. Perhaps identical to subsequent Pmoun. Πακειλ (Pakeil) - appears in well tag 145. Παλου (Palou) - appears over two lines of well tag 108 as Πα Ι λου. Παμώ(νθου) (Pamo(nthes)) - appears in well tag 166 as a hydreuma pmoun. Perhaps identical with Pmoun Pam( ) and Pmoun Pamo from Kellis texts; therefore perhaps also related to Pmoun Pem( ) mentioned in Deir el-Hagar 1 (see note 17 above). ΠατώφιΚς) (Patoos) - appears in well tag 131. Πιαμ (Piam) - appears in well tag 112 associated with a so-called self-operated chdrion. Πισήχθις (Pisechthis) - appears in well tags 113 and 121. The name "Pisechthis" also appears in 183 and in the unpublished Trimithis ostraka inv. 13070 and 13249, from the 2008 season. Πκης (Pkes) - appears in well tag 157. Πω ITOU (Po.itou) - appears in a list with Pmoun Ese, 77. Σαραττί(ωνος) (Sarapion) - appears in well tag 160 as a hydreuma pmoun. Tcc[ (Ta[) - appears in well tag 126; perhaps to be restored as Taesis, Taeteim, Talebe, or Tametra. The latter two Pmouns are known from Kellis texts. Ταήσεω(ς) (Taesis) - appears in well tag 124. Ταητειμ (Taeteim) - appears in well tag 147. Ταοσιρα (Taosira) - appears in well tag 174. Τκηλε (Tkele) - appears in an account listing shipments of wood and chaff, 17. Also appears in the Kellis agricultural account book. Τπακ(ε) (Tpake) - appears in multiple well tags, 125, 156 and 165, including as a hydreuma pmoun. Also appears as a village alone in texts from Kellis. Τσαμεν (Tsamen) - appears in well tag 98. Tsamen seems likely to be the same as the Pmoun Tsalem which appears with Pmoun Osire in the unpublished Trimithis ostrakon inv. 13223, from the 2008 season. Τσαταρσε (Tsatarse) - appears in well tag 109. Τσοα (Tsoa) - appears in well tag 110.

36

OSTRAKA FROM TRIMITHIS

Ταύρου (Tsurou) - appears in a well tag with an unusual format, 138. No personal name is given, only a note for 30 desmai. Τχαλεμσαψε (Tchalemsapse) - appears in tag 223, which shares the format of the standard well tags, but lacks the terms hydreuma or pmoun. Based on the Coptic root 2αλμε, for "spring." Φιβ (Phib) - appears in an account of days, 39, with Pmouns Psoin and Mark (?), and in a well tag 161 as a hydreuma pmoun. It is also mentioned in 56. Reference to hydreuma pmoun Pheb appears in Deir al-Hagar ostrakon 2 (see note 17 above). ΧΘ traces - appears in what is perhaps a list of wells, ostrakon 426. Identical to the subsequent Pmouns? Χθαυ (Chthau...) - appears in the well tag 106. Χθυα( ) (Chthua) - appears in the well tag 227 associated with the vineyard called Moun, which is perhaps a vineyard self-referentially named for the well itself. Ψω (Pso) - appears in various chits from Area 2.1 and Area 4. Also appears in an unusual double-sided well tag naming Pmoun Osir[ios], 155. Hydreuma is perhaps to be restored to the left of that attestation. Identical with subsequent Pmoun? Pmoun Pso may also appear in the unpublished Trimithis ostrakon inv. 13251 from the 2008 season. Ψωι (Psoi) - appears in various tags, including some attestations as a hydreuma pmoun. Also appears on the reverse of an account mentioning an Aphrodisieion in 13. Identical to preceding Pmoun? 144, an attestation of Psoi, is one of only two well tags to use the word 'moun' lacking the Egyptian definite article T'; cf. 143. Ψωιν (Psoin) - appears in 39, an account of days with Pmouns Phib and Mark( ? ). 'G0p( ) (HorQ) - appears in well tag 111 naming the vineyard of Petosiris. Identical with subsequent Pmoun? "Qjpoy (Horou) - appears in list or account 27 followed by five personal names, including one, Makarios, affiliated with Pmoun E[. Identical with preceding Pmoun? The following ostraka also contain partially read, unread or lacunose pmoun names: 14,106,129,151,162,164,168,170,175, 211, 212, 290. Other Place Names Aphrodisieion - appears in an account. See commentary to 13. Chorion of Ammonianos - oil from this site entered into a storeroom as recorded in a memorandum, 257. May also appear in a similar tag, 255. Chorion of Ammonios son of Claudius - appears in three tags, 191, 192, and 225. Note also the appearance of υδρ(ευμα) το Άμμώνιον in the unpublished Trimithis ostrakon inv. 13086, from the 2008 season. Chorion of Ammonios son of Onnophris - appears in a tag, 242. Chorion of Diophanes - appears in a well tag, 172, with the hydreuma pmoun formula. Chorion of Horigenes the smith - appears in a tag, 233. Chorion of Moun - appears in a well tag, 227, associated with the Pmoun Chthua( ). Chorion of Pal( ) son of Pete( ) - appears in a tag, 105, associated with Pmoun Thaut. Chorion of Petosiris son of Harmais - associated with Pmoun Hor( ), 111. Chorion of Petosiris son of Hermokles - appears in a tag, 203. Chorion of Psenamounis - appears in a grain account, 10. Chorion of Sarapion - associated with both Pmoun Berri (114) and Pmoun Tsoa (110).

INTRODUCTION

37

Magd(ol-) - appears in a tag, 200. Perhaps identical with the next item. Magdoliou? - appears in a list, 77, including other places names, Pmoun Ese and Pmoun Po.itou. Pselthis - this toponym appears in the correspondence of Serenos to Philippos as the origin of a marión of oil, 299. With the exception of the unpublished Trimithis ostraka inv. 13223 and 13249, from the 2008 season, it is not attested, and is presumably a nearby Oasis village.

Managing Wells and Estates One of the striking discoveries from the ostraka in Area 2.1 has been that the structure of estate management on which the Kellis Agricultural Account Book (KAB) is founded is identifiable here as well in most details through the appearance of its characteristic terminology, even if we can say far less about it from these ostraka than a long account naturally allows us to do. It is noteworthy that the finds from House B2 also show signs of referring to similar forms of estate management. The center of the estate to which House Bl was connected was the House (οίκία), which we see mainly in the role of receiving goods. It is referred to as the house of Mr. Serenos (ό κύριος Σερήνος) in 300, and as Serenos is the second best-attested individual in the ostraka published in this volume, it is hard to doubt that in fact this house was his during some part of its existence.19 The other individual similarly identified is Domnion, this volume's best attested individual from the house, referred to as "landlord" (γεούχος) in 253. He may have been the head of household before or after Serenos. His association with a subordinate named Nikokles, himself securely attested after 351 via pricing data, suggests a similar dating for Domnion, but the picture is complicated by stratigraphy, as we have seen. As in the KAB, there was also a lady of the house, not identified by name but, as in the KAB, referred to as the οικοδέσποινα in 249 and 250. Demetria's mention in 302 of Gelasios, another apparent subordinate of Domnion, suggests her as a potential lady of the house.20 Explicit reference to the oikodespoina in a letter from Antinoos to Domnion further strengthens this connection. Mention in an unpublished 2008 ostrakon, inv. 13232, of four matia for the oikodespoina in Mothis, shows that the lady of the house was active in both cities in this period. At the next level down were managers. In the KAB these seem to have been called pronoetai, but in the ostraka from Area 2.1 the only reference (4) is to a phrontistes, a more-or-less synonymous term well known from other contexts for the managers of production units of aristocratic estates.21 These managers had storehouses, for which the standard term was celia (κέλλα), attested in 257. They addressed one another as "colleague", literally "brother" (αδελφός), although this term was used so broadly in correspondence that we cannot be confident that all of those using it were in fact 19 The Trimithis ostraka from the 2008 season, not published in this volume, produce a considerable number of references a man named Serenos. Many, perhaps all, of these likely concern the same man here, the head of the house. 20 Gelasios is not a common name in the Oases or in Egypt generally; cf. Onomasticon Oasiticum for references in the Kellis papyri. P.Kellis 1.16, a "business note" directing the Kellis Gelasios to deliver four arfabas of dates, is comparable in context to the appearances of our Gelasios in Trimithis and may well refer to the same person. But two of the other instances (P.Kellis 1.29 and SB 18.13852) refer to a former logistes and an exaktor respectively, perhaps too high status to be the same man. 21 See D. W. Rathbone, Economic Rationalism and Rural Society in Third-Century A.D. Egypt (Cambridge 1991) 71-82.

38

OSTRAKA FROM TRIMITHIS 22

colleagues at the same level. A particular curiosity in this respect is that Serenos uses this term in 297 and 299 addressing Philippos; in one of these the letter is about a minor matter of sending goods, and we might tend to think of the two men as mid-level 23 managers. But in the other it is about a political matter, which we have argued (see below on Political Community) suggests that they were city councillors, or at least that Serenos was. In 298 Serenos also uses the term "brother" in addressing Paesis on a petty business matter, directing him to make a delivery. Although some element of breadth and even condescension should be allowed for, it is by no means inconceivable that Serenos himself was a manager. The example of Aurelius Alypios, who managed the Arsinoite estates of the equestrian Aurelius Appianus although himself a bouleutes, is a useful comparison. Texts from the 2007 season showing both Serenos and Domnion in various intermediary business roles strengthen this possibility. If the Serenos in 290 is the Trimithite councilman, he is there doing no more than issuing a receipt for delivery from an unknown well. A Paesis issued two receipts, 292 and 293, for delivery through Domnion of 10 matia "to the scribes" and 4 matia ώστε τω κυρίω μου τω στατιωναρίω respectively. If this Paesis is the same man whom Serenos addressed in 298, we have a tidy picture of Domnion and Serenos conducting their managerial responsibilities through the same subordinate. The estate was clearly not, as one would anticipate in the Nile valley, simply a matter of land and its plantings and infrastructure. It was above all, as the tags indicate, a matter of wells, the most important source of wealth in the oases. One account (19), which as indicated above must date from the period 352-360, shows that water was measured in units of a day and charged on a cash basis, at that date of 1,000 talents per day. The solidus was according to the same text trading at some 7,511 talents, which means that a well would earn a solidus a week if fully engaged at that rate, or 2/3 pound of gold per year. No doubt there were maintenance expenses, but it is unlikely that they took more than a modest fraction of that income. The capital value of the well is unlikely to have been less than 6 pounds of gold, depending on what allowance for expenses and rate of return on investment one adopts.24 As the previous section showed, on the assumption that the well names on the ostraka indicate rents paid from the use of wells, the house's occupants seem to have had an interest in many wells. They may not, however, have been the sole proprietors, and it is of course possible that their investments at some well locations were in land or plantings rather than in wells themselves. Any attempt to calculate total wealth would thus be in vain. The cultivators who paid for the water were, as everywhere, called γεωργοί. One other term found in the ostraka needs discussion here. That is "share", μέρος (11, 53, 250), made explicit as "the share of Domnion" (μέρος Δομνίωνος) in 279. In papyri from the oases mews can refer to a portion of jointly held property. P.Kellis 1.30 (363 CE), an exchange of ownership rights involving land at Aphrodito, describes a 22 Examples occur in 295, 298, 299, 301, and 302. 23 As he might also be in the unpublished Trimithis ostrakon inv. 13009, from the 2008 season, where Serenos's name is perhaps to be restored. 24 There is no reason to suppose that the well would not have been in use consistently throughout the year. If expenses ran to 10 solidi, net income would have been 42 solidi. At a 10 percent rate of return, the capital value would be 420 solidi, or 5 5/6 pounds. Most accounts of ancient agricultural investment assume rates of return lower than 10 percent, which would yield a correspondingly higher capital value.

INTRODUCTION

39

"sixth part" (1. 10: μέρος έκτον) of land under joint ownership (11. 10-11: κατ]ά κοινω| νίαν). The term is also used to refer to portions of a house involved in inheritance or sale (e.g. P.Kellis 1.37, 320 CE), and portions of a harvest (e.g., P.Kellis 1.74, mid-4th c.). In the region's ostraka, the word appears to refer more typically to shares of outgoing payments, particularly to one's landlord (e.g., O.Kellis 1.73-74). In the Trimithis texts, several of these uses may be at stake. The exact context of the expense account 11 is unclear, but the entry μέρους κερ(άμια) μ may indicate expenditure of a portion of a recent harvest. Other ostraka more clearly suggest divisions in a large estate. The payments made by Horos in 250 are divided evenly between those sent to Trimithis and those sent to the mews. The account in 53 lists α μέρος, the first meros, implying the existence of additional numbered divisions. We may then suppose that the estate run by the House itself had a number of administrative divisions, distinct from its home base in Trimithis, perhaps allotted to the various high-profile individuals we see in the ostraka. Another element of house activities comes from a series of accounts detailing the distribution of rations. Eight of these ration accounts survive, almost all from room 14; a potentially related memorandum was found in Room 13. These texts most commonly include entries for artoi (loaves) and tiphagia, an as yet unidentified product.25 Entries for barley and wine appear in four of the ration accounts as well. The amount of bread usually 20 loaves, although 30 appear in 65 - and wine - 2 maria, nearly 22 liters, in 69 suggests that we are dealing with rations for a substantial number of people. Precisely how many is impossible to say, but it is suggestive that the most complete ration accounts indicate a specific month and date. Are these daily rations for some part of the estate work-force? The chief hint of the intended recipients for these rations comes from 64, the memorandum from room 13, which mentions 20 loaves τοίς παιδίοις, "for the slaves." The only other reference to slaves in Trimithis texts so far comes from another text from room 13, 289, a fragmentary receipt mentioning a mation and της παιδίσκ[ης, "the slavegirl." These two references in tandem with the ration ostraka may introduce us to a new element of estate management in the Dakhla Oasis. Slavery as an institution appears absent from the Kellis account book, and the term used here in 64, παιδίον, appears in 26 other Kellis texts only in reference to children. If 20 to 30 loaves and just over 20 liters of wine did amount to a daily ration, we are perhaps not dealing with a very large number of slaves even here. Whether the role of the slaves in this estate was primarily urban or in some way related to the wells under estate management is not possible to determine on present evidence. Domnion appears in two of these ration accounts, once (66) addressing his subordinate, Gelasios, whom we have already mentioned. In 295, Gelasios issues an order for payment of 2 maria of wine to an otherwise unknown Erabios. The coincidence in personnel between the orders and the ration accounts may suggest that a number of other Trimithite memoranda and receipts deal with the same transactions as the eight ration accounts themselves. 322, for instance, a memo written in Choiak, describes payment to the camel-driver through Domnion of 10 matia and 1 keramion for 25 For tiphagia see Bagnall, Thanheiser, and Worp, ZPE 122 (1998) 173-88, with response at O.Douch 5.535 (continuing to argue for a connection with cereals), and further speculation at O.Kellis 1.92. See also 30, 34, 42 and 301 herein. In the ration accounts 63 and 65-68, tiphagia always appear with loaves (άρτ(οι)) in the same ratio of 1:1. 26 P.Kellis 1.26, a fragmentary report of judicial proceedings, and P.Kellis 1.71, a private letter.

OSTRAKA FROM TRIMITHIS

40

the arrival of the centurion. (Compare for instance 68, also from Choiak, a ration account for 5 matia and 1 keramion.) Returning to estate management more generally, we see that four men act as signatories in texts mentioning Domnion: Nikokles, Gelasios, Paesis, and an unread ....es. Nikokles is the most central figure here, appearing in five ostraka, three times as a signatory.27 All four men issue receipts, write letters, give orders for distribution of goods, and conduct other business relating to estate management. Perhaps these men were all phrontistai for the share of Domnion. If both Domnion and Serenos were landlords connected to the house in Area 2.1, the personal connections between them are unclear. No prosopographical links between Domnion and Serenos can be demonstrated in these texts with any certainty (although see Paesis on p. 38). Business operations involving so many people might, however, be expected to appear in evidence from elsewhere at Trimithis. If the Nikokles signing a receipt in 286 is the same as Domnion's subordinate, then this text from Area 4.1 demonstrates prosopographical links between the activity at the villa and activity documented by finds on the temple hill. The following table lists the find-spots for the texts attesting to Serenos, Domnion and Nikokles, the most frequently attested figures in the business of estate management in Area 2.1. (Regnal years have not been read in any of these texts.) Serenos Room (DSU) 2(13) 2(13) 2(13) 4(23) 5 (213) 10 (119) 17 (153) 17 (155)

O.Trim. 297 298 299 300 271? 268 310 311

Domnion Room (DSU) 4(23) 4(15) 4(23) 14 (221) 14 (233) 14 (233) 14 (233) 14 (233) 14 (233) 14 (233)

O. Trim. 279 287 253 66 292 293 324 322 323 69

Nikokles Room (DSU) 4(23) 10 (220) 13 (208) 15 (backfill) 16 (161) 16 (161) 16 (158)

O.Trim. 279 226 288 291 313 282 312

Table 3. Find-Spots for the Three Most Frequently Attested Trimithites Bold indicates texts providing links between the three people. The three Serenos ostraka from Room 2 and the one from Room 4 were found at or above the house's top floor level, indicating that they belong to the last occupation phase. Room 5's 271 may have come from the roof of the house, or may have been used as a chinking sherd; in either event, it may not refer to the same Serenos as the landlord. Rooms 16 and 17, which provide two of the Serenos ostraka and three of the Nikokles ostraka, are a corridor and room under a staircase, and should be considered together. All five of these ostraka belong to the last phase of the villa. Of the remaining four Nikokles ostraka, one comes from backfill, and another comes from the last occupation phase of the so-called Purple Room (Room 13). Seven of the Domnion ostraka come 27 Nikokles also appears in the unpublished Trimithis ostrakon inv. 13086, from the 2008 season.

INTRODUCTION

41

from the so-called Red Room (Room 14), and were either likely or certainly at the level just above the floor, on the room's last occupational phase. Two of the remaining Domnion pieces came from the third and final floor level of Room 4 in chronological sequence (DSU 23) and the level just above it respectively (DSU 15). This leaves 279, the prosopographical link between Nikokles and Domnion. This ostrakon also came from the layer constituting the final floor layer (Room 4, DSU 23). In sum, Serenos is attested with certainty only during the house's last occupation phase, and likewise Domnion is attested in both the last phase of the Red Room and final occupation levels elsewhere in the house. This leaves us no obvious way to determine whether the Domnion period predates the Serenos period or vice versa. The one ostrakon from any of these texts datable via pricing data, 282, is from the batch of Nikokles and Serenos texts under the staircase in Rooms 16 and 17, and dates after 351. Both groups belong generally to the period from 350 to 370 CE.

Commodities and Measures Unsurprisingly these ostraka attest many commodities and measures. Without exception, the measures are those already known from the Kellis documentation, particularly from the Kellis Agricultural Account Book (discussed in the introduction to that volume, section 4D, pp. 47-51), and there is nothing to suggest that they have different usage or values at Trimithis from those in use at Kellis at the same period. As the same is true in the ostraka found in excavations at Ain el-Gedida, a fourth-century site not far from Kellis, it may well be that the oasis formed a natural unit in this respect. Similarities in the ostraka from Douch and Ain Waqfa may suggest that the Great Oasis as a whole shared a number of metrological characteristics somewhat distinctive from those of the valley. The reader is referred to the introduction to the KAB for details and discussion. References to occurrences can be found in the index. Apart from the artaba, which occurs throughout, we enumerate the following measures and the commodities with which they occur: For wine, must, and other liquids, the keramion of 18 sextarii (κεράμιον), the marión of 20 sextarii (μάριον), and the chous of 6 sextarii (χους). The sextarius (ξέστης), the standard unit of measurement for liquid in the Kellis account book, has so far appeared at Trimithis at least three times, including in 69, one of the estate ration accounts discussed in more detail above (p. 38). For wheat and barley, the modius, at 10 modii to 3 arfabas (μόδιος) and mation (μάτιον), which can have varying relationships to the artaba, mostly either 1/10 or 1/23; the mation is in one case qualified with τοπικφ (μετρώ), "by the local (measure)," a term denoting a measure 1/23 or 1/24 of the artaba. The precise reference of "local" in these texts is unknown. For chaff, the moion (μώιον) appears four times in Trimithis ostraka, in 17, 35, 48 and 298. 35 and 48 are accounts of chaff distributed through individuals listed by name, but no indication of the size of the moion is given beyond that already known from other sources. For hay, the bundle (δέσμη) appears many times in Trimithis ostraka, in 32, 41, 50, 55, 59, 63, 70, 138, 254, 256, 266, 285, 286 and 302. Unlike the KAB, where the unit is simply abbreviated δ/, δέσ( )and δέσμ( )are more standard here, confirming the supposition in the KAB that desme is the resolution to be prefered over dema.

42

OSTRAKA FROM TRIMITHIS

For cotton, the Trimithis ostraka provide more information about the nature of the λιθ( ), the ambiguous measure present in the KAB and two Kellis ostraka. Two cotton accounts from room 13 of House Bl provide a sample of amounts measured as λιθ( ), 38 and 44. These attestations of the term add further weight to the argument put forth in the KAB that the measurement at stake is λίθος (or a diminutive),28 not a phonetic error for λίτρα. The first of these accounts, 38, includes extremely small fractions of a lithos which allow us to determine that the measure's fractional system was based on a unit at least 10 Roman pounds in size (see note ad loc.), thus further confirming the tentative conclusions put forth about the size of the lithos in the KAB. The second cotton account, 44, contains cotton payments to individuals listed by name. What is striking about these payments is their range, from a fraction of a lithos in one case to over 26 lithoi in another. Two more payments are 11 and 12 lithoi respectively. These larger payments would hardly have been light loads, and may suggest a larger place for cotton in the agriculture of the oases than previously suspected. With a lith( ) at a minimum about 3.23 kg of cotton, we would be dealing with a total of more than 200 kg of cotton—a very large amount of cotton, when considered in terms of volume. (On cotton in the oases see Bagnall, BASF 45 (2008) 21-30.) Other commodities mentioned in the ostraka also present no novelties compared to the rest of the oasis documentation. These are counted rather than measured: loaves of bread, dom fruits, tiphagia (above, p. 39), and chickens.

Trimithis as a Political Community29 As we have indicated above, one of the ostraka involving Serenos (297) suggests that he was, along with Philippos, a bouleutes, a member of the city council. By chance, before the excavations at Amheida began, the only two personal names in published texts that could be identified with fourth-century Trimithis were a Horigenes from the KAB30 and a Serenos from P.Kellis IG. 27.31 That text is a letter addressed to Serenos in his capacity as pmepositus pagi Trimitheos. Its editor remarked that "it remains mysterious how this document (which was probably sent to Trimithis!) arrived at Ismant-el-Kharab. Did Serenus retire to Kellis and did he take the document with him?" Since the letter tells Serenos that he has wrongly infringed the rights of Mothites, it is likely that this copy belonged to a resident of Kellis, which was part of the territory of Mothis. Serenos is not a particularly common name for the Oasis in this period: the KAB gives only two examples, and there is only one other instance in the volumes of Greek 32 and Coptic documents from Kellis. (Some unpublished evidence will be discussed 28 Cf. O.Douch 5 634, an order to furnish 1.5 λίθιν of cotton. In the note, Wagner remarks, "C'est la seule fois que la mesure cotonnière est écrite en toutes lettres: ce serait donc le lithion, diminutif rarissime de lithos (LSJ 1048, S.O.), clairement une unité pondérale standard, et non une mesure volumétrique." Regrettably, there is no plate to allow the reader to check this proposed text, of which every letter is dotted, and to verify whether the reading of the second iota is in fact compelling. 29 Most of the contents of this section first appeared in somewhat different form in R. S. Bagnall and G. R. Ruffini, "Civic Life in Fourth-Century Trimithis: Two Ostraka from the 2004 Excavations," ZPE 149 (2004) 143-52. 30 A graffito in the temple area found in 2005 was written by Horigenes son of loannes. 31 We (like J. D. Thomas, cf. BL 11.100) accept the interpretation of the text by R. P. Salomons quoted in the first edition but not accepted by the editor. 32 This is a patronymic on a jar sealing in O.Kellis 198. A Serenos in the Kharga Oasis appears in O.Waqf. 15, where an identification with a Serenos in O.Douch 107 is proposed.

INTRODUCTION

43

below.) No date is given in the Kellis papyrus. The editor cites (note to line 2) an unpublished Kellis papyrus dated to 309 which might be addressed to the same Valerius Herculanus who is the author of P.Kellis 27. That would give an early date for Serenos. But this is very insecure, as only the end of the name survives in the unpublished text. The date of the papyrus can, however, hardly be later than 324/5, as the editor observes, because the use of the nomen Valerius for officials disappears, in favor of Flavius, once Constantine gets control of Egypt in 324. Nor could the date be earlier than 307/8, on present evidence, because the pagus system is not known in Egypt before that time. Before we consider further the possibility of identifying the Serenos of our ostraka with the pmepositus pagi, we must look more closely at 297, where a reference to a psephisma concerning liturgies forms the the basis of our view that he was a councilor. Three documents from approximately the same period help make this case. (1) SB 16 12754 is dated to the fourth century33 Although described as being of unknown provenance, it was surely excavated clandestinely at Kellis34 and is directly relevant to our case in Trimithis. Its first full lines read: Μ]ωθιτών πόλεως της μεγάλης Όάσ[εως ]ς του βουλευτηρίου ψηφίσμασι ούτως . [,35 (2) P.Oxy. 12.1417 is dated by the editors to the early fourth century. It is a report of a trial concerning the attempt of Nilos and others to avoid their duties as eutheniarchs after appointment by psephisma of the boule of Oxyrhynchos. (3) P.Oxy. 49.3507 is dated by the editors to the late third or early fourth century. It is a personal letter from Apollonios to Sarapodoros, in which the former asks that the psephisma regarding the gymnasiarchy be enforced in the case of Andromachos, who had been compelled to serve additional days in the previous year and now hopes to be credited for that service in the current year. It is evident that as in the case of the Trimithis ostrakon these decrees often deal with issues of liturgical obligation and exemption. Bowman drew a similar picture in his work on town councils.36 In his formulation, it "is therefore clear that the task of the boule was simply to confirm that the person claiming [liturgical] exemption fulfilled the required conditions, and to grant the exemption."37 Our psephisma περί της λειτουργίας could well be just this sort of decree. The process through which psephismata were drafted, passed, publicized, and enforced remains somewhat unclear. Here, Serenos says he wrote the decree himself. (One may wonder why then he would need it sent to him. The answer may be that he had drafted it in a single copy, now in the possession of Philippos.) Presumably, members of the boule would vote yes or no on a psephisma once it had been proposed: Bowman points to P.Oxy. 44.3171 (cited in Town Councils as P.Oxy.ined. 1), a list of signatures of boule members followed by the word έ'δοξε, as evidence of this voting 33 First published inAegyptus 63 (1983) 149-50, with discussion of the identification of Mothis. 34 There are other papyri in the collection of the Université Cattolica di Milano from the Dakhla Oasis, acquired in 1968 (see SB 16 12229 and 24 15903; cf. also 15902? and K.A. Worp, Tyche 15 [2000] 189-90). Around the same time the University of Genova and Duke University also acquired papyri from Kellis on the antiquities market. One of the Duke papyri is published by J. F. Dates in BASF 25 (1988) 129-35, and several more appeared in the memorial volume for P. J. Sijpesteijn as P.Sijp. lla-c. P.Genova 1.20 and 21 are from Kellis; the first of these is republished with one of the Milanese fragments in P.Genova 2 Appendix. 35 Restore perhaps εν TOÏ]Ç (suggestion of Klaas Worp). 36 A. K. Bowman, Town Councils of Roman Egypt (Toronto 1971) 83-87,110,112,114. 37 Bowman (above, n. 36) 86.

44

OSTRAKA FROM TRIMITHIS 38

practice from the early third century. According to Bowman, the γραμματεύς of the 39 boule was "responsible for the publication of the ψήφισμα." Could this have been why Philippos was involved in this exchange? Faced at Trimithis with a house abandoned in the 360s, evidently belonging then or earlier to a member of the city council named Serenos, who was concerned with liturgical appointments, is it plausible to propose an identification with the praepositus pagi named Serenos in P.Kellis 1G.27? The likelihood of this identification is strengthened by the nature of the role of the praepositus pagi. In the early fourth century, the traditional role of the stratèges in the appointive process for liturgies had been largely transferred to the praepositus.40 The likelihood of this identification grows still further when we consider two unpublished Trimithis ostraka found in Serenos's house in the 2008 season.41 They are brief memos concerning annona collection made through a man named Serenos and recorded by an exaktor named Sarapion. The exaktofs position as a superior to the praepositus pagi with authority over certain tax collections strengthens the impression we have of Serenos's place in the Oasis.42 And the same group of ostraka from the courtyard has yielded another piece of evidence perhaps linked to the same role of praepositus, ostrakon 252. This is a cryptic note naming Mouses son of Psirais as "warden of foreigners," τηρητής ξένων. This office is not elsewhere attested, and it is not easy to interpret. In general, references show teretai acting as private watchmen. Only two of our references appear to concern a teretes acting in a public capacity. The first is P.Oxy. 12.1507, a third century CE order from the nome irenarchs to send the teretai of the village of Teruthis. The second is P.Oxy. 34.2714, which its editors describe as a "straightforward document... submitting to the strategos the names of the liturgists designated by the villagers for the year just beginning." Included among those liturgists are the τηρηται κώμης και καρπών ("Watchmen for village and crops"), four men for the hamlet of Paneui. Their role is thus public, but in other respects closely analogous to the private watchman's role attested in other papyri. Nothing in these texts thus provides an obvious parallel to the mention of ξένων. It is not immediately evident who is meant by xenoi in this context, or why they would have needed a warden. Presumably these are people currently at Trimithis but domiciled elsewhere. In the municipal structure of the fourth century, that should mean people whose idia was in another city or its territory. Of course that might mean people whose registered domicile was no more than a few hours away, in another part of the oasis that was part of the Mothite nome. The crucial question is whether the role of the teretes was to protect the xenoi and their interests, or to protect the state's interest in keeping track of them and making sure that they fulfilled their obligations. Anyone familiar with the administration of the later Roman Empire, certainly of the Egypt we know from the papyri of the period of Diocletian and Constantine, will instinctively suppose the latter to be more likely, even though the instances of teretes cited above would incline one more to think of protection. It is worth looking at the contemporary evidence in the papyrological corpus concerning 38 Bowman (above, n. 36) 37. 39 Bowman (above, n. 36) 12. 40 N. Lewis, The Compulsory Public Services of Roman Egypt2 (Pap.Flor. 28,1997) 86. For bibliography on the office, Lewis (p. 42) cites his article in J/P 2 (1948) 53-57 and J. Lallemand, L'administration civile de l'Egypte (Brussels 1964) 131-34. 41 Trimithis ostraka inv. 13004 and 13014, both found in Room 22 of Area 2.1. 42 For the office of the exactor, see J.D. Thomas CdE 34 (1959) 124-40.

INTRODUCTION

45

"foreigners." There is not much from an explicitly official, civic context. One exception is P.Oxy. 17.2106, an early fourth century letter from a prefect to the Oxyrhynchite civic officials. The letter outlines the provisions for a compulsory purchase of gold, but specifies that "no burden be laid on strangers unless they have established homes and have not yet been councillors and happen to be well-to-do." The prefect did not find it necessary to explain what strangers he meant. Some strangers were fugitives, and apprehending these was a perennial preoccupation, as we see in the archive of Aurelius Isidoros. In P.Cair.Isid. 128, a group of village officials from the Memphite nome declare that they have apprehended some fugitives (ανθρώπους εν φυγή) from their village who were staying in Karanis. "We may safely assume," the editors wrote, that these fugitives "were impelled by the desire to escape from the burden of taxation or liturgies." P.Cair.Isid. 126 provides a still more direct comparison. It is addressed from Herakleides, praepositus pagi of the fifth pagus of the Arsinoite nome, to another praepositus pagi, and refers to "the order... to turn over to the most sacred fiscus at the rate of five folles per head all strangers [τους ξένους] who are found in the villages."43 The editors date the text, and the otherwise unknown imperial edict to which it refers, to c. 308/309. It is not difficult to imagine that such persons might need a watchman, especially if financial gain were at stake. With Herakleides the praepositus in mind, the possible identification of Serenos with the praepositus pagi of the same name seems even more plausible. But we may wonder if xenoi in our ostrakon would be likely to refer to such fugitives in the absence of any qualification. It is now time to return to a striking aspect of P.Kellis G. 27, namely the fact that it is addressed to a praepositus pagi of Trimithis. What can this mean? The pagi were, as far as we know the system elsewhere, numbered rather than named (as the toparchies had been). It is of course conceivable that we have the name of the old toparchy transferred to the new pagus through custom (in the way that the new pagi were sometimes called toparchies in the Hermopolite nome, see P.Herm.LandL, pp. 9-10). But this curt message44 to the praepositus pagi certainly comes from a high official, most likely the praeses, and such informality is not as likely in this context as it would be in a text generated locally. Moreover, the notion of Trimithis as only the chief village of a pagus at a date that is perhaps most likely ca. 309 (and is certainly not before 307/8) runs into the fact that 45 Trimithis was a polis by 304. Cities, in our experience elsewhere, do not have praepositi, they have a civic government and a lo gistes. And yet it must be said that so far we have no evidence of a lo gistes for Hibis, Mothis, or Trimithis. Rather, there is a logistes "of the Great Oasis" mentioned in P.Kellis 1 G. 25. This document, unfortunately, lacks a date, but it certainly cannot be earlier than the introduction of the office of logistes into the Egyptian metropoleis (303, on present 43 Lines 3-8, trans, by eds.: το κελευσ[θέ]ν | υπό τή$ θείας κ[α'ι ο]ύρανίου αύτ[ών τύ]χηξ των δεσποτών ήμ[ών] βασιλέων πάντ[α$] του$ ξένους του$ εύρισκωμένουξ εν ταΐ$ κώμαΐξ .piç[.].yq)Eç κατά [α]νδρεΐ[ον λ]όγον προ σ[ε]ισμοο έπανεγκίν τω ίερωτάτω ταμι[εί]ω προς φόλλ'ι$ πέντε. The editors cite RÀpolLAno. 9 for an eighth-century parallel to the use of xenoi to refer to fugitives. On fugitives in fourth-century villages see D. Rathbone, "Villages and Patronage in Fourth-Century Egypt: The Case of P.Ross.Georg. 3.8," BASF 45 (2008) 189-207. 44 It is to be translated, "Valerius Herculanus to Serenos, praepositus pagi of Trimithis, greetings. You require without good reason that people coming from the city of the Mothites be made subject against their will to provision of camels and beasts." As J. D. Thomas remarks (JEA 84 [1998] 262), it is a reprimand. 45 P.Kellis I G. 49.

46

OSTRAKA FROM TRIMITHIS

evidence). In his notes to the edition, the editor pointed to the appearance of an Aurelius Kleoboulos, the same name as the logistes here, as a politeuomenos in M.Chrest. 78, a communication from this Kleoboulos to the praeses Thebaidis in 376-378. It is worth quoting the full prescript of P.Kellis 25, because the distinctions embedded in it are important for our purposes: [Α]ύρηλίοις Κλεοβο[ύλ]ω λογιστή Όάσεως [Με]γάλη$ και Φιλοσαρά [πι] δι τω και Μικκάλω [αρ]ξαντι προέδρου Μω[θι]τών πόλεως και [Άν]δρομάχω Άπόλλων[ο]ς αρξ(αντι) συνδίκο[ις] One notices immediately that the circumscription of the logistes is the entire Great Oasis, not one of its cities. The ex-magistrate and current proedros, by contrast, has only Mothis as his circumscription. Unfortunately, the papyrus breaks off with only the tops of the letters of line 5 visible, and we cannot tell what the circumscription of the syndikoi was. It looks, in other words, as if there was a logistes for the Great Oasis as a whole, consisting of three cities, but each city had its own municipal officials and presumably council. In a structure of this sort, it could well be that each city also constituted a pagus with a pmepositus. Exactly what his relationship to the civic officials was, we cannot say, but he may have functioned as a kind of mini-logistes on the spot.46 It is a natural assumption that the ostraka found in the courtyard and, as we suppose, fallen from the shelf in the niche, belong to the last phase of occupation of the house, which should belong roughly in the 360s. There is additional, although inexact, support for placing this Serenos in the middle of the century in two unpublished Kellis papyri, one from Serenos to Alexandros and one from Alexandros to Serenos. The first of these refers to a payment formerly made "to the deceased Philippos." Given the connection with Serenos, it is very likely that this is the Philippos of our first ostrakon. The archaeological context of these letters, like that of most of the published Kellis letters, is to be assigned to the middle or third quarter of the fourth century. It is not impossible that the Serenos of P.Kellis 27, dated ca. 309-324, could be the same as that of the other texts, particularly if the identification of Valerius Herculanus as a praeses of ca. 309 is discarded. Equally, however, a homonymous son or grandson could be at stake. It may also be noted that the Serenos known from the KAB (line 801) was an officialis. One could imagine that Serenos, after a civic career in Trimithis, had entered the imperial civil service. But this speculation would take us well beyond the evidence. In any event, these ostraka give us a sense of the operations of the civic government of Trimithis as a polis, a status that it and Mothis may well have acquired at the same time under Diocletian, probably ca. 303 as part of a general restructuring of the Great Oasis into the domain of a single logistes presiding over three cities, each with a council and under the supervision of a praepositus pagi. We do not mean to minimize the thinness of the evidence on the basis of which we propose this reconstruction, but it make sense of several features otherwise difficult to explain. A recent study has suggested that almost every city in Egypt ever designated by the place name -itôn polis, 46 It may be added that there is one published mention of a Mothite nome, in P.Sijp. lib, dated to 350 or 351, and presumably coming from Kellis (cf. above, n. 34; the village name is only partly preserved). In that instance, however, [rfjc μεγάλης] Oaoecoç is added after the nome name. Another instance occurs in an unpublished papyrus from Kellis.

INTRODUCTION

47

as Trimithis was, can be shown to have been a nome capital.47 If true, an alternative reconstruction could be that Trimithis enjoyed a brief period as a nome capital before ceding the honor to nearby Mothis, perhaps in 310.48 But the pattern discerned in the valley reflects a situation in which civic status and position and capital of a nome always coexisted; if, as we have suggested, this was not the case with the oasis, then the argument from the form of the name loses all force.

Figure 4. Plan of Area 1.

47 J.H.F. Dijkstra and K.A. Worp, ZPE 155 (2006) 183-87, with P.Kellis I G. 49 (304) in which a certain Piperismi is described as apo Trimitheitôn poleos. 48 For Mothis, see P.Kellis 1, p. 124. The place appears in Trimithis texts so far only in the unpublished inv. 13085,13090, and 13232, from the 2008 season.

48

OSTRAKA FROM TRIMITHIS

Figure 5. Plan of Area 1.3 (House B2). One other aspect of the public life of Trimithis should be mentioned here, its garrison. One ostrakon found in Area 2.1 (73), refers to an ala (εϊλη) and to a decurión (δεκαδάρχης). We know from the Notitia Dignitatum that Trimithis (mistakenly placed in that text in the Small Oasis) was the home of the Ala I Quadorum, and this is surely the unit referred to. So far, however, the archaeological exploration of Amheida has found no sign of structures identifiable as a military camp, and we accept the view that the camp in fact lay at El-Qasr, a few kilometers to the north of Amheida, the oldest part of which is laid out in a fashion strongly reminiscent of a Roman camp and in which observations by the Dakhleh Oasis Project in 1980 of the old well found at its bottom signs of Roman occupation. Explorations by F. Leemhuis in 2006 and 2007 and by the Supreme Council of Antiquities have revealed standing remains of massive Roman brickwork structures, which are to be associated with this camp, and Coptic ostraka referring to the imperial kastron, now become a settlement with headmen.

INTRODUCTION

49

House B2 (Area 1.3) As well as House Bl, a smaller house in Area 1 was excavated, designated B2. (Its area is referred to as Area 1.3. Area 1.1 is an area across the street from 1.3, not yet excavated; Area 1.2 is the street; Area 1.4 is the courtyard outside House Bl.49) Most of the ostraka from Area 1.3 are short accounts, listing payments of wine, wheat, doum palm fruits, and other items now lost or unclear from the context. These accounts are discussed in more detail below. One ostrakon from occupational material in Area 1.3 (200) and another from the surface of the unexcavated building across the street (101) are jar tags, giving a personal name, a date, and a place. The first of these tags was found intact in a mud jar-stopper, presumably serving as a label for the goods contained in the jar itself. This jar stopper was found underneath roof collapse on a floor layer, which indicates that it derives from a secure context. The other tag, if correctly restored, is similar to the well tags found in abundance in Area 2.1. Those texts indicated the origin of goods, probably as a payment from someone at a well or agricultural area near that well to someone in Trimithis, presumably the owner of the well or property nearby. The second tag is somewhat different: it records the name of a misthotês or contractor instead of that of what we may suppose was a tenant farmer, perhaps in this case someone handling the well on behalf of a large-scale landowner. Finally, a noteworthy exception to the dominance of ostraka among surviving Trimithis texts comes in the form of an incised unbaked clay tablet (24, which is, with 62, the second such tablet from Trimithis) recording a account of doum fruits dated Hathyr 9 of an unknown year. This was found in occupational debris with brick collapse in Room 7 of the house. The texts themselves provide few more specific clues to their own dating. 36, found in the street just outside the house, has a ratio of 20 drachmas per artaba, which indicates a date prior to 275 CE, showing that the street was in use by that time. 51, an account of wheat and barley found in Area 1.4, the courtyard to the south of the house, most likely dates to before ca. 244. From an archaeological standpoint, the house plan and associated ceramics from Area 1.3 suggest a third-century CE occupation, perhaps extending into the fourth century, but more precision is difficult. Year 1 appears in one text already mentioned, 200; this does little to narrow the options, but it is unlikely that it is to be attributed to any regnal year later than that of Diocletian (284/5); an indiction, however, is possible. The reference appears in an ostrakon set in a mud jar-stopper found in a level of occupational debris. (The mention of year 4 in 101 is of no use, coming from surface finds across the street.) Nor do we have many prosopographical clues, with one possible exception discussed below. Certainly, the apparent lack of Biblical names in this small sample supports our sense of a third- or early fourth-century context in preference to any date much later. By contrast, the fourth-century material from Area 2.1 is richer in Christian names, including lona, Mouses and Makarios, and attests to several deacons and priests. One contrary indication, however, is the appearance in one Area 1.3 account (26) of a Bêkis διάκων, or deacon, implying the presence of ecclesiastical hierarchy at Trimithis while this house was still occupied. Such use of Christian titles as identifiers is not found in documentary papyri before the early fourth century; the period of Licinius' rule 49 This building and its environs will be published in a forthcoming volume of Amheida, edited by Anna L. Boozer.

50

OSTRAKA FROM TRIMITHIS

between 313 and 324 is the earliest horizon where this is found.50 The text in question comes from a secure context on the final floor level of the house, giving our most secure indication of the period of house's final occupation phase. As with the ostraka from elsewhere in Trimithis, the accounts from Area 1.3 feature several payments to and perhaps from outlying agricultural sites named after regional water-wells, indicated by the formula "pmoun + place-name." We know two of the wells named in Area 1.3 ostraka from other Trimithis texts. Pmoun Harau appears both in texts from House Bl and the Area 4 temple site. Pmoun Osire appears in an ostrakon found in House Bl in a locus of probable dump material under windblown sand. The ostraka from Area 1 also name Pmoun Embôo(u), a well attested in Area 2.1's well tag 137. Pmoun Osire, which appears with Pmoun Emb(ôou) in 22, also appears in Area 2.1 ostrakon 155 with Pmoun Pso, one of the most frequently attested well names in Trimithite texts. This may suggest that these two Pmouns were near Pso, itself in turn perhaps a major well close to Trimithis. Pmoun Osire appears in an unpublished 2008 ostrakon from Trimithis, inv. 13223, with Pselthis, a toponym appearing in 191 below, which reinforces the impression that Pselthis, a previously unknown toponym, is a small village nearby. In short, the ostraka from Area 1 not only add to our knowledge of Oasis toponyms, but also confirm that the inhabitants of this house had ties to sites appearing elsewhere in the city's economic networks. The units of measure used in the ostraka from Area 1.3 are familiar from the Kellis Agricultural Account Book and elsewhere. An account of loads, 22, delivered from (?) the well of Pmoun Emb(ôou) and from Pmoun Osire lists loads of 4, 15 and 22 gomoi. If the loads were hay or chaff, they would have weighed 80, 300 and 440 kilograms respectively. If the loads were turnips, as the measure is used in the Kellis account book, we do not know how large they may have been. Another house account, 20, mention payments in maria, a unit of liquid measure attested in the ostraka from Area 2.1 and from texts elsewhere in the Oasis. A surface find account from Area 1 (1) mentions payments in maria as well. The payments in maria amount to a range from 10.8 to 540 liters of wine. Another account, 26, the largest from the house itself, records payments from seven different people, six of whom pay more than once. The payments are varied, from a single keramion to 18 keramia and from 1 to 22 choes of oil. (The standard rate is 3 choes to a keramion.) These figures are not inconsiderable. No ostraka from Area 2.1 use gomoi as a unit of measurement, but 440 kilograms of hay or chaff is a heavy load. The amounts recorded in the account from Area 1.3 would have been one donkey-load, four donkey-loads and six donkey-loads respectively51 This level of scale raises interesting questions. Who in this house needed to transport 440 kilograms of wheat, and on what sort of business? The physical remains of House B2 suggest inhabitants of rather lower social status than those of the large house (Bl) at Area 2.1. The occupations attested in the Area 1.3 ostraka confirm that impression. Two texts in very similar hands from Area 1.3 have linen-weavers in them (20 and 21); those from area 2.1 have none. (The presence in the house's Room 3 of archaeological material connected with textile/clothing manufacture may be related to this textual evidence.) Area 1.3 has two camel-drivers (20 and 21); Area 2.1 has only one, Syros, on an ostrakon found in backfill. Area 1.3 appears to have two donkey-drivers (22 and 200); Area 2.1 has none. Whether the figure (Pse( )?) in Area 50 See M. Choat, Belief and Cult in Fourth-Century Papyri (Turnhout 2006) 57-73. 51 KAB 49-50.

INTRODUCTION

51

1.3 is a tailor or a teamster (21), Area 2.1 has neither. These gaps in occupation attestation levels are all the more striking when we consider how many more ostraka have been found from Area 2.1 than Area 1.3. 52 These ostraka from Area 1 attest perhaps as many as 57 different people. As with the names attested from Area 2.1, those we find here typically derive from Amoun, Horus, and Shai, all gods popular in the Oasis in the Roman period. At least 18 of the 57 53 people have names of this variety. Thoth, the central Trimithite deity does better here than in Area 2.1, appearing at least three times in the form of Pathotes and twice as Hermesias, a related form. The rest of the onomastic corpus includes variants for other 54 Egyptian deities and more standard Greek names (Korax, Syros). The demotic texts include a proper name with the element Wslr, for Osiris, and a daughter of Tephnachthis (NB Dem. 1232). Akula, for the Latin Aquila ("eagle"), is one onomastic stand-out in the ostraka from Area 1, appearing at the end of 1, an account found during surface clearing. As a Roman proper name, Aquila would seem to be more at home with Valerius or Claudia, names attested in Area 2.1. In our context, the name appears in an account of wine payments, the longest text from Area 1, which ends with the phrase εσχον από Ακύλα, "I got from Aquila." The entry is followed immediately by a single line with a date different from the other entries in the account, which may mean that Aquila is unrelated to the rest of the text. Alternatively, we may propose that the author is an estate manager making payments of wine on behalf of a large estate owner, and Aquila himself is the ultimate source of the wine he pays out. The sums involved suggest a large-estate context. 50 maria of wine is the largest amount listed here. (No amount of maria higher than 19 appears in the ostraka from Area 2.1; none higher than 33 appears in the Kellis account book.55) This entry shows that someone associated with Aquila was paying 540 liters of wine to Geron son of Pamoun, and listing the payment as only one, albeit the largest by far, among three others that day. We know an Aquila from elsewhere: the father of the estate owner in the Kellis Agricultural Account Book had the same name. However, as the Kellis account book dates to the third quarter of the fourth century (at which time Aquila is only a patronymic and may not have been alive) and our text from an undatable surface find, we do not have a sufficient basis to hazard an identification between the two men named Aquila. The other large account from Area 1, 26, also suggests the presence of a large estate of some kind. The seven payments - presumably of oil, although specified as such in only one entry - in keramia and choes have already been mentioned. Six of the seven people paid more than once, in entries typically recorded as "for extra" (ύπ(έρ) προσθήκ(ης)), a formula recorded a mere dozen times in the Duke Databank. The first two such entries from the Area 1 account can be taken as representative: 52 This count indudes patronymics. The Bekis listed without a patronymic may be one of the others appearing herein. Likewise for the Psais and Horos appearing without patronymics. Psen[ may be a Psenamounis from elsewhere in these texts. Psenamounis son of Pa[ may be Psenamounis son of Pathotes. 53 Perhaps add Plout( ) (Ploutammon?). 54 One is tempted to suppose this a translation of the Egyptian name Αβωκ (NB Dem. 96), but this name has not to date appeared in any documents from the Great Oasis. 55 Most of the entries in the Kellis account book are for a single marión. The two largest, for 33 maria and 20 maria, appear in lines 916 and 1700 respectively.

52

4

OSTRAKA FROM TRIMITHIS

Σαμως Πατχωτ( ) κε(ράμια) ιη και ύπ(έρ) προσθήκ(ης) κε(ράμιον) α, και ύπ(έρ) προσθ(ήκης) κερ(άμιον) α Βήκις Ψάιτ(ος) έλ(αίου) χό(ες) λβκαιύττ(έρ) προσθήκ(ης) χό(ες) λβ.

In the first entry, Samos makes two payments hyper prosthêkês amounting to oneninth the size of his original payment. By contrast, in the second entry, Bekis makes one payment hyper prosthêkês the same size as his original payment. From entries this sparse, it is not easy to tell what the "extra" payment is for. On its own, prosthêkê can refer to any addition or increase. Exact parallels to the usage in our text are hard to find. According to a fourth century Hermopolite tax receipt, P.Charite 14, a tenant of Aurélia Charité in arrears for the kanonika commuted the amount in cash with an extra 25% described as prosthêkê. Johnson and West took this to refer to a tax supercharge, "probably a superindiction," and provided further parallels in sixth century Hermopolis and Aphrodito.56 The closest parallel to our text seems to be P.Prag. 2.139 (V), a Hermopolite tax document in which the phrase records what the editor took to be a series of rent increases.57 This would be particularly attractive if lines 4 and 5 of that text refer to the same georgos; the prosthêkê in that case would be 75 arfabas of grain, precisely one-third an initial payment of 225 arfaba of grain. Nonetheless, the context is not entirely clear, and the editor's commentary on prosthêkê notes the unusual presence of a term better suited for a tax account. In our account, there is no indication that taxes of any kind are at stake. The primary payments may be for rent, but the supplemental payments are less likely to be for rent increases than rent arrears or future payments. The most suggestive example in this regard is the payment noted above by Bêkis son of Psais, in which the payment ύπ(έρ) προσθήκ(ης) is the same size as the original payment itself, 22 choes. The possibility that a rent increase is at stake would seem to be eliminated by the inconsistency in the size of the prosthêkê from one entry to the next. No apparent pattern in the ratio between payments and supplements can be seen to clarify the matter; the prosthêkê payments are l/9th, one whole, l/5th, 1/10* and 2/15ths the size of the initial payments respectively. Neither this account nor any of the other texts can help with the exact identity of the owner of House B2. Nonetheless, two hypotheses emerge regarding the identity of the house owner. Given the nature of the texts, property management is clearly at stake. Perhaps we are dealing with an estate manager. The fact that someone in this house could read and/or write supports this possibility. But the Kellis Agricultural Account Book alone suggests that an estate manager would produce more written material than we find here. Perhaps we are dealing with some sort of sub-manager or assistant (boêthos) instead. Alternatively, we may be dealing with someone in the transportation industry. The courtyard next door wraps around the house, which may suggest the need for ample space. The reference in one text to delivered shipments (the items παραδοθ(έντα) in 22) reinforces the impression of our house as a transportation destination. Reference to oil, 56 A. C. Johnson and L. C. West, Byzantine Egypt: Economic Studies (Princeton 1949) 241-42, discussing SPP 20.94 = P.Charite 14 (326-327 CE?). Worp's commentary to the re-edition in P.Charite follows Johnson and West on prosthêkê. 57 The phrase also appears in a series of eighth-century tax receipt ostraka, O.Petr. 464, 465 and 467.

INTRODUCTION

53

the classic estate product, in the same text, again suggests that property management is at stake. If these texts are contemporary - and stratigraphy suggests they are - then the presence of multiple toponyms suggests that we are dealing with an estate of some size, in multiple parcels. If agriculture in the Oasis was profitable in this period, levels of production must have been high. A large proportion of Trimithis was likely populated with people serving the traffic of goods from the pmouns to the city and then from the city on to the Nile Valley. The house from Area 1.3 may thus represent what we can expect to find throughout the city in structures of comparable social register. To summarize: the names of the people attested in Area 1.3 blend in well with those attested in Area 2.1 and in the rest of the Oasis. The occupations attested in their texts are typically manual labor, related to transportation and the manufacture of clothing. This might suggest that the inhabitants of this house were among the lower social strata of Trimithis, themselves active in one of these industries. At the very least, their day-today business put them in frequent contact with the donkey-drivers and camel-drivers of Trimithis. But the amount of goods appearing in their accounts suggests that they were involved in rather large transactions, perhaps receiving shipments and making payments from those shipments on behalf of local large estates. It is thus more likely that they should be seen as middle managers rather than transporters or craftsmen.

Ostraka from the Temple Hill No secure contexts were excavated during the 2004-2007 seasons on the hill where the Temple of Thoth once stood. The reasons for this situation are the extensive process of stone robbery and demolition that took place here. The more recent stages of this process go back to the sixteenth or seventeenth century, as stones from Amheida were taken for reuse in houses in El-Qasr, but the digging of pits in search of treasure may have continued until much later. At the same time, this very exposed spot has suffered extremely high deflation from wind erosion, resulting in a loss of perhaps 1-2 meters of deposit. As a result, although about 800 decorated or inscribed blocks from the temple have been recovered, and even more undecorated ones, none were in situ; all were found in a jumble, mostly in one or another of the dozens of pits, some several meters deep, dug by treasure-hunters. Nothing remains of the foundations of the temple of the early Roman period, decorated under Titus and Domitian. Only in the 2008 season was an undisturbed context, dating to Dynasty 25 or Dynasty 26, found. As a result, all finds published here come from thoroughly mixed contexts. In Area 4.2, where the foundations of a structure, perhaps a gateway, were investigated, the situation is only marginally better, and in any case most of the eight ostraka recovered in that area have little or no remaining content. One of them, 315, is a Demotic text assignable to the later Ptolemaic period. From Area 4.1, the area of the temple itself, 31 ostraka were found. Five of these (280, 305, 306, 378, 384) were in Demotic or Abnormal Hieratic. These undoubtedly all come from the period of the temple's operation. What is perhaps more surprising is the presence of a considerable number of well tags of the type found in Area 2.1. The bulk of these date from relatively low-numbered years, as can be seen in the table given above, with 8 of the 9 coming from years 1-10 (97 and 98 [both for Apollos son of Sarapion], 105, 107, 109, 195, 199, 203). The lone exception is 127, of year 33. There are two other indications that activity on the hill was continuing in the fourth century, the occurrence

54

OSTRAKA FROM TRIMITHIS

of Psais the deacon (known from two other ostraka of that period) in 383 and the cooccurrence of Nikokles and Philippos in 286. Both of these are well known from the final phase of occupation of House Bl.

Personal Names It would be premature to assess the onomastic character of Trimithis solely from the evidence of four seasons, dominated by a single house. The notes given here are simply an attempt to assemble some of the evidence that will eventually contribute to such an assessment and to a potential wider study of the names of the Great Oasis.58 It will be obvious that the repertory is dominated by names coming from three cults of great popularity in the Oasis during the Roman period, Amoun (in both Theban and Libyan variants), Horus, and Shai. Thoth, the central god of Trimithis, is only modestly represented, and Tutu, the god of Kellis, appears only a few times. Quite a number of other Egyptian gods appear, but in general only in a few instances each. It is striking how feeble the presence of Isis is; and even more noteworthy compared to Kellis is the almost total absence of names formed from Bes. No Greek cult does any better, and the remaining names are largely scattered over a large number with a few attestations each. Hardly any of the cults represented on the Roman temples of the oasis, then, were still vital sources of onomastic practice in fourth-century Trimithis. By comparison to Kellis, where Bes, Tutu, and Tapsais all make respectable showings, Trimithis is not well supplied with a range of theophoric names. Even its hometown deity Thoth/Hermes, is not abundantly present. And yet it is also striking how little sign of the Christianization of the onomastic repertory there is at this date. That is true of the Kellis Greek papyri as well, although one finds more Christian names at Kellis. Overall, however, Trimithis seems to have a profile similar to that of Kellis apart from this theophoric impoverishment and the relative lack of the common but locally distinctive names characteristic of Kellis like Gena/Iena, Kome, Loior, and the like. The one exception is names formed on Loui, which occur in considerable variety at Trimithis, even if not in great quantity59 Note: only names of individuals mentioned in the ostraka are included (with fathers and grandfathers); in the case of toponyms, only recognizable personal names are in eluded. (Other toponyms are discussed above.) Names that are too fragmentary to be recognizable are also not discussed. References to occurrences of names can be found in the index of names at the end of this volume. Names with more than ten references are marked with a plus (+); those occurring only once have an asterisk (*). Some names could evidently be included in more than one category.

58 It is regrettable that the onomastic appendix of Guy Wagner's thesis on the oases, which was not included in his Les oasis (1987), has never been published. It would now of course be much out of date, given the mass of material from Kellis, but it would have been of great use. Fortunately, the compilation of onomástica for the Kharga and Dakhla Oases by Robert Salomons and Klaas Worp have made it relatively easy for us to check occurrences of names outside Trimithis. The Onomasticon Oasiticum is available online at http://www.media.leidenuniv.nl/legacy/onomas_final.pdf.. 59 On the likelihood of local onomastic particularism manifest in rare names, see G. R. Ruffini, "The Commonality of Rare Names in Byzantine Egypt," ZPE 158 (2006) 213-25.

INTRODUCTION 1. Names formed on Egyptian gods

55

60

Amoun/Ammon Names formed from the stem Άμμων-, representing the Libyan version of Amoun, are very common. +Άμμώνιος dominates; there are smaller numbers of Άμμωνιανός, +Άμμων (not always easy to distinguish from Ammonios, given truncation and abbreviation), and *Άμώνιος. All of these are fully Hellenized in form. Cf. also under Hephaistos for Ήφαιστάμμων, Hermes for Έρμάμμων, Nile for Νειλάμμων, and Sarapis for Σαραπάμμων. Note also Πλουτάμμων and Φιλάμμων. Kellis in general exhibits a similar picture. From the Theban version, Amoun, we get Παμουν, *ΠαμοΟνις, ΣεναμοΟνις, *Ταμουν, and +ΨεναμοΟνις. Of these, only the last is at all common, with just 6 instances of all the others combined. Psenamounis is also very common at Kellis. Achem Only *ΠαχοΟμις occurs; it is found moderately often at Kellis. Anoubis Only *Άνουβ and *Πανουβας are found formed on the divine name. *Πουώρις ("the dog", NBDem. 181) may also refer to Anoubis. Apis The only names certainly formed on Apis are the Hellenized *Άττίων and the Egyptian *Πααττις. It is absent from the Kellis material. *Παφις could be Pa-h ρ (NBDem. 400), "the one of Apis," but pa-h'pj, "the one of the Nile" (NBDem. 399) and P3-hf, "the snake" (NBDem. 204) are also possible. Atum Only *ΨενετΟμις occurs.

Bes The only names formed on Bes are *Βησας and *Βήσις. The contrast with the prevalence of Bes names at Kellis is striking. Chonsu The only name formed on Chonsu, the son of Amoun and Mut and third member of the Theban triad, is Πετεχών, "the gift of Chonsu." It is rare at Kellis. Herakles Chonsu's Greek equivalent is represented by his name prefixed with the Egyptian article, Φηρακλής, found in three texts. Other forms are the article-less Ηρακλής, *Ήρακλας, ""Ηράκλειος, and ""Ηρακλείδης, which do not collectively make much of an impression, and even adding a couple of names where the resolution is unclear we do not get much. The compound *Ήρακλάμμων is also to be noted. The situation at Kellis is similar.

60 See also discussion above at pp. 53-54.

56

OSTRAKA FROM TRIMITHIS

Harpochrates Άρποκρατίων occurs twice. Horus Horus is very well represented in the names, mostly with the simple divine name +rGOpo$, but also with the compounds Άρπαήσις ("Horus the one of Isis"), *Άρσιήσις ("Horus the son of Isis"), *Άρμάις (divinized Horemheb), Άρυώτης ("Horus is well"), and ΠετεΟρις ("gift of Horus"), along with probably feminine derivatives of these, *Σεναρσιήσις and *Σεναρυώτης. In addition, there are the Greek derivatives *'ϋΟριγένης and 'COpicov, and apparently *'ϋΟριαίνα in an aberrant spelling. *Παχράτης ("the one of the child", NBDem. 411) is probably also to be referred to Horus. Βήκις (the falcon) appears three times at Trimithis. The name is also rare at Kellis.61 Isis Isis appears in two feminine versions of "the one (fern.) belonging to Isis," *Θαήσις and Ταήσις (only in a well name). The masculine Παήσις ("the one (mase.) belonging to Isis") appears four times. Commoner is Ψεννήσις (8 examples, one with single nu, compared to 5 at Kellis), "the son of Isis". *Πετήσις and the undeclined *Ψενταησα appear once each. There are also two instances of the Greek formation Ίσιδώρα, "gift of Isis".

Min *Παμιν is the only name attested so far derived from the god of Panopolis. Paminis occurs a few times at Kellis. Montu This Theban god is represented only by *Παμώνθης in a well name. A couple of Παμnames not completely preserved or written out could be additional instances. Nephotes This god, known at Thebes and at Diospolis Parva (and whose name is also used for Chonsu) is represented only by Πετενεφώτης ("the one (mase.) given by Nephotes"), also found once at Kellis. The name appears at Trimithis also in an inscription in the temple complex. Nile The river appears moderately often, with Νείλος/Νίλος in 6 cases and the compounds Νειλάμμων in 3 texts and Νειλοφέρων in one. The derivative *Νιλίων also appears. The former appears at Kellis in a few cases, the latter not. Cf. also under Apis for a possible Nile name in Παφις. Osiris +Πετοσΐρις is moderately common, as at Kellis; we also find *Ψενοσΐρις. Cf. also the epithet *Όννώφρις (Wn-nfr, NBDem. 118). Rait *Παραίθης ("the one of [the goddess] Rait (?)") is the one clear derivative. Possibly *Ψιράις is also to be placed here. 61 O.Kellis 30.1 and 30.6.

INTRODUCTION

57

Renenutet ΠατερμοΟθις/ΠατερμοΟτις is found twice. Sarapis +Σαραπίων is well attested, better than it is at Kellis. The compounds Σαραπάμμων and Φιλοσάραπις (with its local variant *Φιλοσερπι) appear, and the hypocoristic Σαραπας. The divine name ""Σεραπις is found once.

Shaï The Greek equivalent of Shaï, ""'Αγαθός Δαίμων, occurring once at Trimithis, is absent at Kellis. The typical form, however, is +Ψάις, "the Shaï", the single commonest name at Trimithis as at Kellis. "ΤΤαψάις and Σάις (4 instances) are rare. The feminine Tapsais, well attested at Kellis (where she was Tutu's consort), is absent at Trimithis. Thoth Trimithis's patron deity makes a modest showing. We find Παθώτης 9 times, Ψενθώτης twice, and 4 attestations of one of the theriomorphic manifestations of Thoth, the ibis, as Φίβις (strikingly absent at Kellis, although Phibion appears once there) and "ΤΤαήβις. The incompletely read ""Παν. .νίβις may also be an ibis name. Hermes Thoth's Greek counterpart appears in several forms: Έρμης (twice) and Έρμησίας (4 times) and the compound Έρμάμμων, also 4 times. *Έρμίας, *Έρμόδωρος, and *Έρμοκλής are found once each. A slightly larger range of Hermes names occurs at Kellis. Tutu Kellis's patron makes a respectable showing (15 times) as +Τιθοής, albeit a small fraction of what we find in his own village. Crocodile gods *Τριάδελφος is the only representative of a crocodile cult we have noted. 2. Names formed from Greek gods not listed above Apollo is represented by ""Απόλλων, Απολλώνιος, and Άπολλώς. The picture at Kellis is similar. Cf. also section 5 below on Πολλουν. Ares: ""Άρης Demeter: ""Δημήτρια. Eros: Έρως. Hephaistos: the compound Ήφαιστάμμων. Heron: "Ηρών appears three times (once at Kellis). 3. Other Greek names No Greek name is really common at Trimithis, as those occurring more than once or twice are mainly multiple appearances of the same individual. The Macedonian dynastic names are not common: Φίλιππος 4 times; ""Αλέξανδρος once in a place name; Βερενίκη twice. Names from the mythological or historical past occur with some regularity, but none is found for many individuals: ""Αγησίλαος, ""Άχιλλεύς, "Έϋμηλος, Νικοκλής (a prominent signer), and ""Πολύβιος. For the rest, most of the Greek (or

58

OSTRAKA FROM TRIMITHIS

probably Greek; some doubtful cases are included) names appear infrequently (and in most cases where they occur more than once, refer to a single individual): ""Αθηνόδωρος, ""'Άνητος (for Άνυτος?), Αντίνοος (4, some perhaps referring to Antinoopolis), Ασκλας (2), Γελάσιος (4), ""Γέρων, ""Δίδυμος, *Διόσκορος, *Διοφάνης, ""Δούλος, ""Ευτύχιος, ""Έπαφρόδιτος, "Επίμαχος, ""Ζηνόβιος, ""Θεόδωρος, "Ίον (= "Ιων?), "Ιώνιος (?), ""Κερδών, Κόραξ (cf. Πεβωκ, section 5 below), ""Κύλλος, ""Λυκαρίων, ""Μάρων, ""Μέμνων, ""Μύρων, Νικάων, Νίνος, "Όλύμπιος, ""Παΐων, Πλουτογένης (2), Σαρμάτης (3), ""Σκύλαξ, ""Στέφανος, Συρίων (2), Σύρος (5), ""Τρόφιμος, ""Τύραννος, ""Υγιεινός, ""Φιλίσκος, ""Φιλομ( ), and ""Χάρις. That some of these may be caiques of Egyptian names or have Egyptian reference is clear. 4. Roman names (and derivatives) The situation is much as with Greek names: a couple of names appearing multiple times because of the individual's prominence in the documentation from Area 2.1, but for the rest a scattering of names occurring mostly just once: ""Αιλιανός, ""Ακύλας, ""Άμάτιος, ""Αντώνιος, ""Βίκτωρ, "Γαίων, +Δομνίων, "Ιουλιανός, Καπίτων, ""Κλαυδία, Κλαύδιος, Λογγϊνος, Μακρινά, ""Μαξιμι.( ), "Όπτατος, ""Ούαλέριος, "Όύαλου( ), Περπέριος, +Σερήνος, ""Σιλβανός, ""Φαυστιανός. 5. Egyptian names other than theophoric Where identified, Demotic equivalents are cited from NBDem. Coptic parallels are cited from M. Hasitzka, Namen in koptischen dokumentarischen Texten, version of 22.i.2007 at http://www.onb.ac.at/files/kopt_namen.pdf. ""Ακης ""Ακυσις (= Έκυσις, "Nubian", NBDem. 80) ""Βανίπις ("iron", cf. Hasitzka 19, Crum CD 41a) Βελλής ("blind", NBDem. 143) ""Καλέ (prob, a variant spelling from Gti, "lame" [NBDem. 1034]) *Καρουρις (Qrwr, "frog", NBDem. 982) *Κολαβϊνις ("axe", Crum, CD 102b) Μέρσις ("red", NBDem. 602) *Νουβης (= Νοβς, cf. KAB 68 n. 36: jujube) *Ούαφ. .[ (perhaps Ούαφρής, Wlh-ib-f, NBDem. 113) "ΤΤαβώς (variant of Πεβώς?) Παλεβ Παμούρις (the one of mour) Παναμεύς Πανύρις (cf. Pinuris, "the dog of Horus", Pj-iwiw-n-hr, NBDem. 156—Vittmann) "ΤΓαούς (prob. Pa-hr, "the one of the face", NBDemAOl)62 ""Παστωούς/Πεστωούς/Παστώς (Pa-stJ.fcw and Plj-stJ.fcw, cf. NBDem., fase. 18, p. 180 ad pp. 559 and 567: "He whom they have redeemed", cf. Enchoria24 [1997/98]92-93Vittmann) ""Πατχωτ( ) ""Παχάλμιος ("the one of the spring; see 109.2-3n.) Ίίεβωκ (perhaps *Pl-cbq, "the raven", cf. NBDem. 96, Αβωκ) 62 Citing J. Quaegebeur, "Tithoes, dieu oraculaire," Enchoria 7 (1977) 103-08 at 104; reference to Tutu in this context seems entirely plausible.

INTRODUCTION

59

Πεβώς (Pï-m, NBDem. 154) ΠεκΟσις (PJ-igs, "the Nubian", NBDem. 160, cf. ΆκΟσις) *ΠιαΟς

Πικ[ (Πικώς?, P#-fö, NBDem. 442-43) Πιναχθις, -ης (from «A/stem, "strength", perhaps with reference to Amon-Nakht?) *Πλήνις (PJ-ljn, "the smith", NBDem. 199) 63 ""Πάλλουν (or form of Πολλών?) ΠρεμεμοΟρις (perhaps *P)-rmt-)mwr, "the man of Crocodilopolis" [Gebelein]— Vittmann) *Πσιρις Σαμουν Σαμώς (T3j-n.lm*w, NBDem. 1348) Σελλής (Σελλαευς) Σίρις

*ΣιΟγχις (possibly *Sj-cnh, *Sj-cnh "living son", cf. Hr-cnh "living face", NBDem 793 — Vittmann). ""Τα OUT -Τερητ( ) Χράτη 5 (Hrt, "child", NBDem. 890) *Ψενελυταλ( ) ^Ψενθηθ *Ψενταλε( ) *Ψιράι5 (cf. above, Rait) Ψλύις 6. Hebrew Old Testament names Μουσής (Moses) is the most common of these, with 4 instances; we also find once each *Έφρεμ, *Ίωνας (Jonah), and *Ίωσηφ. 7. Other Christian names These are relatively uncommon. We find Μακάριος twice, ""Μαρτυρίας, Παύλος (4 times), Τιμόθεος 4 times, and Ψενπνούθης ("the son of God") 6 times. 8. Semitic names Σαμβας (twice) and the diminutive Σαμβάτιον. 9. Thracian (royal) name *Ροιμητάλκας 10. Iranian name ""Κύρος (found in the Old Testament) 11. Dakhla Oasis epichoric names64 Γενα

θατ *Κελε 63 Cf. P.Kellis 5, p. 40, for thé suggestion that this is a shortened version of Απόλλων. 64 We do not intend to suggest any particular ethnic or linguistic origin with this classification, which includes names we cannot confidently place in one of the above categories; some of these are certainly theophoric. For discussion of the onomastic peculiarities of the Oases see Wagner, Les Oasis 242-249.

60

OSTRAKA FROM TRIMITHIS

*Kopccu Λουι *Λουια *Λουιάμμων *Λουιατο[ *Λουιερμαμουν *Λουισαι *Λουισάι$ *Λουισεν *Λουισΐρι$ *Λωινο[υτε?] *Μώνις *Παγενα$

Παρά Πισήχθις65

Του For Loui- names, cf. KAB, p. 68 n. 33; P.Kellis5, p. 31. It seems likely that Loui(a) was the name of a local god, but no direct evidence of the cult has yet been discovered. We take Pagenas to be Pa + Gena (var. lena; cf. KAB, p. 67). That, Kele, Korau, Moni, Para, and Tou are all distinctively oasite names found in KABand to a large extent also in the Kellis ostraka and papyri. We suspect that some may reflect a local predilection for truncation of names: Κελε for Κελεπιν (also found in KAB), Μωνι(ς) for Αμμώνιος (extremely common in the oasis), Παρά perhaps for Παράμμων, not so far found in Dakhla but attested at Douch, in the Small Oasis, in Siwa, and in the El-Arag Oasis (130 km south of Siwa, where Parammon was the god worshipped along with Ammon).66 Cf. also above, section 5, on Πολλουν. 12. Names of uncertain origin *Είραμου

*Έράβιος (perhaps = Άράβιος, but that is a doubtful name itself) "Ασμα Νοθη

Ούαει/Ούαϊ (also O.KeUis 114.conv. 2; Egyptian whj, "oasite"? Or Greek ούαί, "woe"?) and Πουαει. *Σαττίων (variant of Σαραττίων?)

Archaeological Contexts of the Ostraka, with Assigned Dates Ostraka are organized in this table by their archaeological context, Area, Room (where applicable) and S(tratigraphic) U(nit); the SU is a D(eposition) SU unless indicated to be a F(eature). The brief descriptions of context in the fourth column are given for all areas 65 See KAB, p. 69, n. 37. 66 Douch: O.Douch 3.225 and 5.521; Small Oasis: P.Oxy. 60.4071 (other Oxyrhynchos instances may also reflect connections with Bahariya); El-Arag: SB 10.10551; Siwa: P.Oxy. 43.3126. Note also instances in P.Marm., perhaps influenced by Siwa. See the discussion by Wagner, Les oasis, 231-32 and 334-35 (on the cult), who calls the god "une divinité spécifique de tout le désert occidental de l'Egypte." Wagner also suggests that Παλάμμων (now also attested several times at Kellis) might be the same name with interchange of liquids. We are doubtful that this is correct.

INTRODUCTION

61

except 4.1 and 4.2. All contexts in those areas were thoroughly disturbed and without stratigraphie value. Room 3 belonged to the house to the south of Bl. In the list of rooms, 0 indicates a surface find. Dates are discussed in the discussion of stratigraphy above, p. 23, and in the notes to individual texts. Precise dates are generally based on regnal years or indictions, coupled with archaeological context. If internal evidence other than years helps indicate a date, this is noted in the edition of the individual texts. Texts without internal evidence of date are dated according to their archaeological context; it will be evident that some texts are probably older than these dates would indicate, particularly in the case of layers of material dumped in preparation for construction. Occupation layers, however, appear to have little or no older material except where a chinking sherd has come out of vault or wall collapse and mixed with occupational debris. Area

Room

SU

Context

1.1 1.1 1.2

0 0 0

0 4 3

1.2 1.2 1.3 1.3 1.3 1.3 1.3

0 0 1 1 1 1 2

5 6 12 16 16 16 7

Surface find Surface of unexcavated area Mud brick collapse in street from north wall of House B2 Secondary trash deposit in street Trash compacted into street surface Clean wind-blown sand Occupational debris above floor Occupational debris above floor Occupational debris above floor Occupational debris at floor level

1.3 1.3

2 2

15 15

Room fill Room fill

1.3

2

20

Occupational debris? Or foundation fill?

1.3

2

20

Occupational debris? Or foundation fill?

1.3

2

20

Occupational debris? Or foundation fill?

1.3 1.3 1.3 1.3 1.4

3 5 7 9 0

19 40 48 73 1

Occupational debris above floor Wind-blown sand below surface Wall and ceiling collapse Top floor level, compacted mud Wind-blown sand fill

1.4 1.4 2 2.1

0 0 0 0(W

2 3 1 (near 2.1) 88

Room fill, occupation debris Room fill, occupation debris Surface find Wind-blown sand covering collapse

Date Attributed 4c. 287/8? 285/6?

O.Trim.

ca 240-275 ca 240-350 ca 240-350 ca 300-350 ca 300-350 ca 300-350 PtolemaicEarly Rom. ca 300-350 PtolemaicEarly Rom. PtolemaicEarly Rom. PtolemaicEarly Rom. PtolemaicEarly Rom. 284/5? Unknown ca 300-350 ca 315-350 335/6 or 353/4 Unknown ca 200-250 ca 350-370 ca 350-370

36 141 21 20 23 388 390

1 101 99

22 389 391 392 393 200 395 24 26 230 445 51 294 191

62 Area

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

OSTRAKA FROM TRIMITHIS Room alley) 0(W alley) 0(W alley) 14/18 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3

SU

Context

Date Attributed

O.Trim.

88

Wind-blown sand covering collapse

ca 350-370

192

190

Wind-blown sand covering collapse

ca 350-370

316

177 1 1 4 4 4 4 4 53 168 168 13 13 13 13 13 13 13 13 13 13 14 32 34 34 34 34 34 34 34 38 40 40 41 41 41 22

Wind-blown sand just below surface Wind-blown sand, top fill Wind-blown sand, top fill Dome collapse and occupation Dome collapse and occupation Dome collapse and occupation Dome collapse and occupation Dome collapse and occupation Wall collapse and occupation debris Floor preparation layer Floor preparation layer Deposit on floor, sealed by sand Deposit on floor, sealed by sand Deposit on floor, sealed by sand Deposit on floor, sealed by sand Deposit on floor, sealed by sand Deposit on floor, sealed by sand Deposit on floor, sealed by sand Deposit on floor, sealed by sand Deposit on floor, sealed by sand Deposit on floor, sealed by sand Fill of niche in wall F3 Final floor level in room Second floor level from top Second floor level from top Second floor level from top Second floor level from top Second floor level from top Second floor level from top Second floor level from top Part of sequence of occupation levels Compacted layer below DSU 38 Compacted layer below DSU 38 Refuse layer below floor DSU 40 Refuse layer below floor DSU 40 Refuse layer below floor DSU 40 Sand and debris above mudbrick collapse

4c. 4c. 4c. 348/9 348/9 351/2 ca 350-370 ca 350-370 286/7? 321/2? 321/2? ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 200-270 ca 350-370 ca 335-350 ca 335-350 ca 335-350 ca 335-350 ca 335-350 ca 335-350 ca 335-350 ca 335-350 ca 335-350 ca 335-350 ca 335-350 ca 335-350 ca 335-350 290/1? or 318/9 or

259 71 340 163 164 166 174 295 100 118 119 4 5 252 297 298 299 356 358 359 360 2 366 176 177 303 345 367 368 369 180 189 371 132 188 376 357

INTRODUCTION Area

Room

SU

Context

2.1 2.1 2.1 2.1 2.1 2.1 2.1

3 3 3 3 3 3 3

33 33 33 36 36 36 93

Debris deposit on top of last floor Debris deposit on top of last floor Debris deposit on top of last floor Part of sequence of floor layers Part of sequence of floor layers Part of sequence of floor layers Pit fill or covering, above earliest floor?

2.1

3

95

Pit fill, above earliest floor?

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

95 7 9 9 9 9 9 9 9 9 9 10 12 15 15 21

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

4 4 4 4 4 4 4 4 4 4 4 4 4 4

23 23 23 23 23 23 23 23 23 23 23 23 23 23

Pit fill, above earliest floor? Wind-blown sand in top layer Mudbrick + debris layer over floor Mudbrick + debris layer over floor Mudbrick + debris layer over floor Mudbrick + debris layer over floor Mudbrick + debris layer over floor Mudbrick + debris layer over floor Mudbrick + debris layer over floor Mudbrick + debris layer over floor Mudbrick + debris layer over floor Wind-blown sand over floor deposit Ash layer in DSU 9 in part of room Wind-blown sand between floor and collapse Wind-blown sand between floor and collapse Mudbrick collapse debris above DSU 23 floor level Top floor level Top floor level Top floor level Top floor level Top floor level Top floor level Top floor level Top floor level Top floor level Top floor level Top floor level Top floor level Top floor level Top floor level

63 O.Trim.

Date Attributed later ca 350-370 ca 350-370 ca 350-370 4c. 320/1? 4c. 292/3? or 320/1 or later 328/9 or 346/7 Unknown ca 335-370 344/5 346/7 ca 345-370 ca 345-370 ca 345-370 ca 345-370 ca 345-370 ca 345-370 ca 345-370 ca 345-370 345/6 ca 350-370 ca 350-370 ca 350-370

381 248 159 162 173 250 296 344 354 355 454 249 161 6 287 3

ca 350-370 ca 350-370 ca 350-370 ca 350-370 343/4 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370

7 8 9 14 124 253 254 255 279 300 301 302 320 362

10 11 304 12 114 257 148

194

64

OSTRAKA FROM TRIMITHIS

Area

Room

SU

Context

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

23 23 23 29 29 29 29 35 35 35 35 35 35 35 35

Top floor level Top floor level Top floor level Fill of bin Fill of bin Fill of bin Fill of bin Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

4 4 4 4 4 4 4 4 4 4 4

35 35 35 35 35 35 35 35 35 35 35

Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

35 35 35 35 35 214 214 214 214 214 214 214 214 214 214

Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Fill under floor level DSU23 Foundation fill Foundation fill Foundation fill Foundation fill Foundation fill Foundation fill Foundation fill Foundation fill Foundation fill Foundation fill

Date Attributed ca 350-370 ca 350-370 ca 350-370 288/9? 3-4 c. 290/1? 3-4 c. ca 275-340 320/1? ca 275-340 ca 275-340 ca 275-340 ca 275-340 ca 275-340 293/4 or 315/6? ca 275-340 318/9? 318/9? ca 275-340 ca 275-340 321/2? ca 275-340 ca 275-340 ca 275-340 ca 275-340 PtolemaicEarly Rom. ca 275-340 ca 275-340 ca 275-340 ca 275-340 ca 275-340 321/2? 322/3? ca 275-340 290/1? 285/6? ca 275-340 ca 275-340 ca 275-340 ca 275-340 ca 275-340

O.Trim. 363 364 365 104 128 175 256 13 113 126 129 130 131 133 149 168 178 179 181 182 183 184 185 186 187 278 370 372 373 374 375 122 123 137 142 147 217 218 219 341 438

INTRODUCTION Area

Room

SU

Context

2.1 2.1 2.1 2.1 2.1 2.1 2.1

5 5 5 5 5 6 6

16 16 213 213 213 67 238

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

7 8 8 8 8 8 8 8 8 8 8 8 8

69 F74 63 63 63 63 74 74 74 74 74 74 77

Wind-blown sand as top fill Wind-blown sand as top fill Wall collapse at base of stairs Wall collapse at base of stairs Wall collapse at base of stairs Roof and wall collapse Compacted layer above one floor and below last floor Ceiling and wall collapse on floor Pit in floor Vault collapse on floor Vault collapse on floor Vault collapse on floor Vault collapse on floor Occupation debris above floor Occupation debris above floor Occupation debris above floor Occupation debris above floor Occupation debris above floor Occupation debris above floor Occupation debris above floor, interpenetrating floor level

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

8 8 8 8 8 8 8 9

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

9 9 9 9 9 9 9 9 9 9 9

85 98 102 102 102 102 182 F93 (north of) 104 104 107 107 114 117 117 127 127 127 127

65 Date Attributed ca 300-370 ca 300-370 ca 300-370 ca 325-350 ca 300-370 ca 300-370 ca 300-370

O.Trim. 251 361 271 318 435 379 172

72 102 165 169 377 380 15 16 17 18 63 73 198

Collapse and debris just above floor Floor debris Probably floor foundation deposit Probably floor foundation deposit Probably floor foundation deposit Probably floor foundation deposit Backfill and Wind-blown sand Backfill

ca 300-370 287/8? 350/1 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 Ptolemaic or Early Roman ca 350-370 320/1? 292/3? 321/2? 290/1? ca 275-350 ca 275-350 Unknown

Probably dump material layer Probably dump material layer Fill dumped under floor Fill dumped under floor Wind-blown sand fill Dump layer Dump layer Fill dumped under floor Fill dumped under floor Fill dumped under floor Fill dumped under floor

325/6? 325/6? 325/6? ca 275-350 ca 275-350 ca 275-350 ca 275-350 ca 315-370 319/20? ca 275-350 ca 275-350

155 156 125 387 414 308 396 76 152 334 398

19 115 108 117 196 197 208 349

66

OSTRAKA FROM TRIMITHIS

Area

Room

SU

Context

2.1 2.1 2.1 2.1 2.1 2.1

9 9 9 9 9 9

127 128 128 129 129 130

2.1

9

130

2.1

9

130

2.1

9

130

2.1 2.1

9 9

132 132

Fill dumped under floor Fill dumped under floor Fill dumped under floor Sandy dump layer in circular structure F93 Sandy dump layer in circular structure F93 Sandy/ashy debris layer in circular structure F93 Sandy/ashy debris layer in circular structure F93 Sandy/ashy debris layer in circular structure F93 Sandy/ashy debris layer in circular structure F93 Pit slightly later than fill under floor Pit slightly later than fill under floor

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9

132 132 133 137 138 146 146 148 151 151 151 151 151 151 151 151 151 167 167 167 167 167 167 197 200

Pit slightly later than fill under floor Pit slightly later than fill under floor Mud floor in corner of courtyard Early dump layer Dump under mud floor DSU 133 Sandy layer under mud-brick collapse Sandy layer under mud-brick collapse Preparation level for floor Dumped debris after closure of courtyard Dumped debris after closure of courtyard Dumped debris after closure of courtyard Dumped debris after closure of courtyard Dumped debris after closure of courtyard Dumped debris after closure of courtyard Dumped debris after closure of courtyard Dumped debris after closure of courtyard Dumped debris after closure of courtyard Dumped debris after closure of courtyard Dumped debris after closure of courtyard Dumped debris after closure of courtyard Dumped debris after closure of courtyard Dumped debris after closure of courtyard Dumped debris after closure of courtyard Pre-construction dumped material Pre-construction dumped material

O.Trim.

Date Attributed ca 275-350 325/6? ca 275-350 287/8? ca 275-350 ca 325-350

404 157 347 103 331 201a

ca 275-350

415

ca 275-350

416

ca 275-350

417

290/1? 294/5, 316/7, 322/3 or later ca 275-350 ca 275-350 ca 325-350 ca 275-350 ca 325-350 ca 275-350 ca 275-350 ca 340-370 319/20? 355/6 ca 275-350 ca 275-350 ca 275-350 ca 275-350 ca 275-350 ca 275-350 ca 275-350 318/9? ca 275-350 ca 275-350 ca 275-350 ca 275-350 ca 275-350 ca 275-350 ca 275-350

106 110

332 338 201 397 394 25 258 314 151 167 335 346 348 400 401 402 407 150 206 207 419 420 421 317 87

67

INTRODUCTION Area

Room

SU

Context

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9

200 200 200 200 200 200 200 200 200 200 203 203 212 212 212 212 212 212 212 212 212 212 212 212 212 212 212 212 212 215 215 215 218 218 218 218 230 230 232

2.1 2.1 2.1

10 10 10

119 119 142

Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Dumped debris Dumped debris Dumped debris Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Pre-construction dumped material Debris layer Debris layer Fill in foundation trench for North wall of house Debris layer above floor level Debris layer above floor level Dump layer under floor

O.Trim.

Date Attributed 319/20? ca 275-350 ca 275-350 ca 275-350 ca 325-350 318/9? ca 325-350 ca 325-350 ca 275-350 ca 275-350 ca 275-350 ca 275-350 ca. 275-340 ca 275-350 ca 275-350 ca 275-350 ca 275-350 ca 275-350 320/1? ca 275-350 ca 275-350 ca 275-350 318/9? 320/1? ca 275-350 ca 275-350 ca 275-350 ca 275-350 ca 275-350 ca 275-350 ca 275-350 ca 275-350 ca 275-350 319/20? 321/2 ca 275-350 ca 275-350 ca 275-350 330/1?

112 139 140 143 222 223 224 319 436 437 211 212 40 43 45 46 47 85 116 120 135 136 213 215 336 430 431 433 450 83 86 434 49 96 121 144 50 227 231

ca 350-370 ca 350-370 294/5 or

75 268 111

68

OSTRAKA FROM TRIMITHIS

Area

Room

SU

Context

2.1 2.1 2.1

10 10 10

142 191 191

Dump layer under floor Dump layer under floor Dump layer under floor

2.1 2.1 2.1 2.1 2.1 2.1

10 10 10 10 10 10

191 191 191 202 220 220

Dump layer under floor Dump layer under floor Dump layer under floor Possible dump directly on gebel Debris after or contemporary with floor F106 Debris after or contemporary with floor F106

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

10 10 10 10 10 10 10 11 11 11 11 11 11 11 11 12 12 12 12 12 12 12 12 12 12 12 12 12

220 220 220 241 241 241 255 0 (backfill) 79 229 235 235 235 235 235 226 226 226 226 245 245 245 245 245 245 245 245 245

Debris after or contemporary with floor F106 Debris after or contemporary with floor F106 Debris after or contemporary with floor F106 Channel fill from construction Channel fill from construction Channel fill from construction Channel fill from construction Backfill Wind-blown sand fill Wind-blown sand fill Collapse onto occupation debris Collapse onto occupation debris Collapse onto occupation debris Collapse onto occupation debris Collapse onto occupation debris Wind-blown sand fill Wind-blown sand fill Wind-blown sand fill Wind-blown sand fill Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault

Date Attributed 316/7, 322/3 or later? 290/1? ca 275-350 311/2? 325/6? 290/1? ca 275-350 ca 275-350 ca 275-350 ca 350-370 327/8, 337/8, or 345/6 ca 275-350 ca 275-350 ca 275-350 ca 250-325 ca 275-350 ca 275-350 ca 275-350 Unknown 344/5? ca 350-370 ca 350-370 319/20? ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370

O.Trim.

202 95 158 241 353 429 352 226 228

237 274 321 62 342 446 337 399 160 232 145 233 234 239 273 52 54 238 236 53 56 57 58 59 61 89 90 91

INTRODUCTION Area

Room

SU

Context

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 13 13 13 13 13 13 13 13 13 13

245 245 245 245 245 245 245 245 245 245 245 245 245 245 245 245 245 245 245 F150 83 185 186 186 186 186 187 199 199

Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Collapse of north wall and vault Top floor level Wind-blown sand Surface layer, Wind-blown sand Vault collapse in W. part of room Vault collapse in W. part of room Vault collapse in W. part of room Vault collapse in W. part of room Clean wind-blown sand fill Vault collapse outside room Vault collapse outside room

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

13 13 13 13 13 13 13 13 13 13 13 13 13

199 199 199 199 207 207 207 207 208 208 208 208 208

Vault collapse outside room Vault collapse outside room Vault collapse outside room Vault collapse outside room Vault collapse from above doorway Vault collapse from above doorway Vault collapse from above doorway Vault collapse from above doorway Fill of northwest bin Fill of northwest bin Fill of northwest bin Fill of northwest bin Fill of northwest bin

69 Date Attributed ca 350-370 ca 350-370 ca 350-370 357/8 ca 350-370 356/7 362/3 357/8 356/7 ca 350-370 ca 350-370 ca 350-370 ca 350-370 354/5? ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 (354/5?) ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 353/4? ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370

O.Trim. 92 93 94 235 240 242 243 244 245 275 276 277 325 326 327 328 447 448 449 55 193 261 262 263 265 266 225 34 35 48 79 138 272 80 81 214 289 38 39 82 84 288

OSTRAKA FROM TRIMITHIS

70 Area

Room

SU

Context

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

13 13 13 13 13 13 13 13 14 14 14 14 14 14 14 14

208 209 216 216 216 216 216 216 184 184 194 194 194 194 194 211

Fill of northwest bin Fill of northeast bin Occupation debris on floor Occupation debris on floor Occupation debris on floor Occupation debris on floor Occupation debris on floor Occupation debris on floor Wall and vault collapse Wall and vault collapse Wall and vault collapse Wall and vault collapse Wall and vault collapse Wall and vault collapse Wall and vault collapse Wall collapse and floor debris

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 15 15 15

221 221 221 221 233 233 233 233 233 233 233 233 233 233 233 233 233 233 233 233 233 233 233 F109 F78 0 (backfill)

Occupation debris with brick & sand Occupation debris with brick & sand Occupation debris with brick & sand Occupation debris with brick & sand Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Habitation layer just above floor Stack of baked brick North-South wall Backfill

Date Attributed ca 350-370 ca 350-370 358/9 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 361/2 or 367/8 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 275-350 ca 275-350 ca 350-370

O.Trim. 432 64 41 42 44 216 269 290 209 210 37 154 229 264 267 270

66 220 221 439 65 67 68 69 70 88 247 292 293 322 323 324 330 440 441 442 443 444 453 451 452 291

INTRODUCTION Area

Room

SU

Context

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1

15 15 15 15 15 15 15 15 15 15 15 15 15 15

109 145 145 145 148 152 152 152 152 157 157 157 157 157

Ash deposit in test pit Flat roof collapse on occupation debris Flat roof collapse on occupation debris Flat roof collapse on occupation debris Sand and brick debris under floor Flat roof collapse on occupation debris Flat roof collapse on occupation debris Flat roof collapse on occupation debris Flat roof collapse on occupation debris Occupational and mud brick debris Occupational and mud brick debris Occupational and mud brick debris Occupational and mud brick debris Occupational and mud brick debris

2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 2.1 4.1

15 15 15 15 15 15 15 15 15 15 15 16 16 16 16 16 17 17 17 17 17 17 0

157 157 157 157 157 157 157 157 261 261 152 158 158 158 161 161 153 153 153 155 155 161 1

Occupational and mud brick debris Occupational and mud brick debris Occupational and mud brick debris Occupational and mud brick debris Occupational and mud brick debris Occupational and mud brick debris Occupational and mud brick debris Occupational and mud brick debris Rubble layer below floor Rubble layer below floor Flat roof collapse on occupation debris Wind-blown sand below mud-brick collapse Wind-blown sand below mud-brick collapse Wind-blown sand below mud-brick collapse Sand and dust above floor Sand and dust above floor Wind-blown sand Wind-blown sand Wind-blown sand Sand and dust above floor Sand and dust above floor Sand and dust above floor

4.1

0

1

4.1

0

2

71 Date Attributed ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 360/1? ca 350-370 363/4 or 364/5? ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 275-350 ca 275-350 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 ca 350-370 PtolemaicEarly Rom. PtolemaicEarly Rom. PtolemaicEarly Rom.

O.Trim. 329 77 204 281 412 29 405 409 410 30 31 32 60 146 260 284 285 350 413 423 425 426 246 343 283 205 312 408 282 313 27 309 310 311 403 411 280 385 305

72

OSTRAKA FROM TRIMITHIS

Area

Room

su

4.1

0

2

4.1 4.1 4.1

0 0 0

8 11 11

4.1

0

11

4.1 4.1 4.1

0 0 0

11 11 11

4.1 4.1 4.1 4.1 4.1 4.1

0 0 0 0 0 0

14 18 18 18 43 43

4.1 4.1 4.1

0 0 0

60 60 61

4.1

0

61

4.1

0

64

4.1

0

77

4.1 4.1 4.1 4.1 4.1 4.1 4.1

0 0 0 0 0 0 0

79 81 86 86 90 99 99

Context

Date Attributed Dyn. 25 or earlier 3-4 c. 3-4 c. 293/4? or 321/2 or later PtolemaicEarly Rom. Unknown ca 315-370 PtolemaicEarly Rom. 3c? 356/7 4c. Unknown 291/2? 293/4? or 321/2 or later 3-4 c. 3-4 c. 285/6? or 328/9 or later 288/9? or 316/7 or later 284/5? or 327/8 or later 285/6? or 328/9 or later Unknown 3-4 c. Unknown Unknown ca 350-370 4c. 293/4? or 321/2 or later

O.Trim. 306 134 190 195

378 382 383 384 307 127 170 386 107 199

74 171 97

105

203

98

33 28 418 424 286 78 109

INTRODUCTION Area

Room

SU

4.1

0

99

4.2 4.2

0 0

1 6

4.2 4.2 4.2 4.2 4.2

0 0 0 0 0

7 8 10 13 13

4.2

0

17

Context

73 Date Attributed PtolemaicEarly Rom. Unknown PtolemaicEarly Rom. Unknown Unknown Unknown Unknown PtolemaicEarly Rom. Later Ptolemaic

O.Trim. 427 339 422 406 333 351 153 428 315

74

OSTRAKA FROM TRIMITHIS

Figure 6. Area 1.3, DSU 48, with tablet 24 in situ

Figure 7. Area 4.2, 315 in situ.

INTRODUCTION

Figure 8. Area 2.1 (House Bl), Room 9, DSU 127.

Figure 9. Area 2.1 (House Bl), Room 13, DSU 208.

75

76

OSTRAKA FROM TRIMITHIS

Figure 10. Area 2.1 (House Bl), Room 1, DSU 13 with 298, 299, and 360.

Figure 11. Area 2.1 (House Bl), Room 8, DSU 85.

TEXTS Note on information provided in the headers: The Inv(entory number) is a unique number in a sequence given to all objects found on the site. Numbers below 2000 come from the 2004 season, those from 2000-3311 from the 2005 season. Numbers above 11000 come from the 2006 season, those above 11700 from 2007. The find-spot designation Area 2.1 (for example) indicates Area 2, sub-area (building) 1. For the room numbers, see the plans of the houses. DSU is the Deposition Stratigraphie Unit (FSU is Feature Stratigraphie Unit); for fuller descriptions of these see the forthcoming report on the buildings in question. Object numbers (Field Number) assigned in the field within the Area (starting from 1 each season) are also given here to indicate proximity of discovery. Ostraka discovered in sieving do not have field numbers of this type. The TM# is the Trismegistos number, which can be used as the standard identifier for all texts published here (see http://www.trismegistos.org/). The same number will also identify each text in the Papyrological Navigator (http://papyri.info/). The main pottery fabrics are coded as follows: Ala = iron rich, red exterior, gray interior; Alb = iron rich, gray; All = Christian brittle, calcium rich; BIO = calcium rich, green. Other fabrics will be described in the report on the pottery of the appropriate context. Comments on the slip are added as needed. I. Accounts and Lists 1. Account of wine. 4th century. TM# 131049. Inv. 0175. Area 1.1, DSU 0, FN 1. From surface cleaning. 10.7 χ 8 cm; on concave side; broken at top. Al. Σαραπίω(νι) "QJpou [ "Qjpcp Παθώτου α όμ(οίως) δ κη Γέροντι Παμουν μάρια ν 4 "QJpcp Άκύσιος ιέ "Qjpcp όνηλάτη η Σαραπίωνι "QJpou είς υίφ Ήρακ( ) δ(ιά) υίοΟ Άμμω(νίου) Φιλίσκ(ου) δ 8 Έπειφ iy μέρο(ς) Ψεναμ(ούνιος) Παθώτ(ου)

ις-

έ'σχον από Ακύλα ] α// ΠαΟνι κ 6 υίόν

"To Sarapion son of Horos To Horos son of Pathotes 1, likewise 4 28th To Geron son of Pamoun 50 maria

78

OSTRAKA FROM TRIMITHIS To Horos son of Akysis 15 To Horos the donkey-driver 8 To Sarapion son of Horos for the son of Herak( ) through the son of Ammonios son of Philiskos 4 Epeiph 13 Share of Psenamounis son of Pathotes 16 I got from Aquila [- -] 1, Pauni 20"

3 For Γέρων cf. O.Kellis 128.5 (with note), where the reading now seems more likely. 50 maria: a large sum, 1000 sextarii or 540 liters of wine. See remarks in the introduction above, p. 51. 4 There are no other attestations of ΆκΟσις in Dakhla texts, but the form with the definite article, ΠακΟσις, occurs at Kellis. 7 Αμ is blotted. For Philiskos cf. O.Kellis 16. 11 One might imagine either [ίν]δ(ικτίωνο$) α// or [(έτους)] α// as possible restorations. 2. Account? After 350? TM# 131050. Inv. 0188. Area 2.1, Room 2, DSU 14, FN 33. In a jar set into the bottom of a wall niche, on top of the sand filling the jar. 4 χ 4.2 cm; on convex side; broken at left and below. Ala with cream slip.

[[ 3

]...CS/ ] Τιθ(ο-) eS/]] ] traces

1-2 These are presumably year or indiction numbers, more likely the latter. 2 This line has been cancelled with a line through it. 3. Account of wine. Ca. 350-370. TM# 131051. Inv. 0194. Area 2.1, Room 4, DSU 21, FN 44. 7.5 χ 8.2 cm; on concave side; complete. Faint ink on a dark fabric, poor contrast. Alb.

4

traces λ α μάρ(ιον) α ινω traces ιωνίορ μάρ(ιον) α

2 Perhaps a personal name at the start, as probably in lines 3 and 4. 4. Account of grain. Ca. 350-370. TM# 131052. Inv. 0199. Area 2.1, Room 2, DSU 13, FN 77. 9 χ 7.2 cm; on convex side; complete or perhaps broken at top, where there may be some

TEXTS

79

descending traces. Ala with cream slip.

4

δι(ά) Παμούνιος μό(διον) α δι(ά) φροντιστ(οθ) και Θατ και άλλων μο(δίους) β δι(ά) Ψάιτ(ος) Πετεχώντ(ος) μό(διον) α

There are faint and scattered traces of three or more additional lines. "Through Pamounis, 1 modius; through the steward and That and others, 2 modii; through Psais son of Petechon, 1 modius." 1 The abbreviation of μό(διος) here and in lines 3 and 4 is written with a dot over the center of mu. 2 For the role of phrontistai in estate management, see the above, p. 37. 1 For the name That, see KAB, p. 66. 5. Account? Ca. 350-370. TM# 131053. Inv. 0200. Area 2.1, Room 2, DSU 13, FN 78. 9 x 6 cm; on concave side; only faint traces preserved on convex. P37. 195 κριθ(ής) (άρτάβας) çr 2-3 vague traces 4 κερ(άμια) δ ..ß traces

6. Receipt or account? Ca. 350-370. TM# 131054. Inv. 0213. Area 2.1, Room 4, DSU 15, FN 124. 5.7 χ 8.9 cm; on convex side; dark discoloration with almost invisible text. Ala with cream slip. Lines 1-7 are only traces; in line 8, (γίνονται) (πύρου) μ ("Total, 40 of wheat"). 7. Account of wine. Ca. 350-370. TM# 131055. Inv. 0214. Area 2.1, Room 4, DSU 23, FN 128. 4.5 χ 4.4 cm; written on both sides; broken at top left. Alb with cream slip.

4

Convex side ]..[ [.].... μάρ(ιον) α ΨεναμοΟ(νις) μάρ(ιον) α Σάις Σάιτ(ος) traces Αμμώ(νιος) Πε

80

OSTRAKA FROM TRIMITHIS Λουισαι μάρ(ι-) (γίνεται) μάρ(ια) Concave side 3 lines, not readable. " . . . 1 marión Psenamounis, 1 marión Sais son of Sais . . . Ammonios son of Pe. . . Louisai, - maria total, - maria"

Convex 4 Presumably an amount in maria is to be read at right, as in the remaining lines on this side. 8. Account of wine. Ca. 350-370. TM# 131056. Inv. 0218. Area 2.1, Room 4, DSU 23, FN 142. 12.4 χ 8.6 cm; on convex side; complete but much faded. Ala. κ traces τοις α κερ(άμιον) α Μαξιμι( ) κερ(άμιον) α 4 Ίωνα κερ(άμια) β κε είς αριστ(ον) απ' οξ(ους) κ[ερ(άμια)] β κη είς δίπνον απ' οξ(ους) ccS' traces απ' οξ(ους) α 8 τω κερ(άμιον) α τω σκυτεί χό(ες) β/ "To the . . ., 1 keramion; to Maximi( ), 1 keramion; to Jonah, 2 keramia; on the 25th, for lunch, from vinegar, 2 keramia; on the 28th, for dinner, from vinegar, 1 Vi; . . . from vinegar, 1; to the . . ., 1 keramion; to the cobbler, 2 choes." This account of wine is difficult to read. It contains expenditures over several days, including for breakfast/lunch on the 25th and dinner on the 28th, as well as to a cobbler. 4 Ίωνας is the biblical name Jonah. The numeral might instead be the S symbol for Vï, but beta seems a better reading. 5-7 The form written here seems to be a xi with a line through it, suggesting ξ(έσται) (sextarii), but there is a small loop on the beginning of the horizontal line which leads us to think that the reading we print is superior; it also explains the presence of απ' before it. 6 είς is just the ghost of the writing.

TEXTS

81

9. Account of barley. Ca. 350-370. TM# 131057. Inv. 0232. Area 2.1, Room 4, DSU 23, FN 173. From a pottery lot. 3.9 χ 5.7 cm; on convex side; broken at left; extremely flaked surface, especially at right. P37, grayish fabric.

4

traces εγώ κριθ(ής) ε Νικοκλής κρ(ιθής) η α traces φα υθω traces τω traces

10. Account of grain. Ca. 350-370. TM# 131058. Inv. 0237. Area 2.1, Room 3, DSU 33, FN 190. 6.3 χ 8 cm; on convex side; complete. Ala.

4

χειμερινής υδ(ρευμα) χωρίο(υ) Ψεναμ(ούνιος) δι(ά) Ψάιτος Κόρακος φόρου σίτου δημ(οσίω se. μετρώ) (άρτάβας) ε y κριθής δημ(οσίω se. μετρώ) (άρτάβας) ßL

"For the winter period, Hydreuma of the Vineyard of Psenamounis, through Psais son of Korax for rent, 5 1/3 art. of wheat by public measure, 2 Vi art. of barley by public measure." 1 χειμερινής: see Wagner, Les oasis, 279-80, citing Olympiodorus for a six-day interval for watering crops during the winter, compared to three-day frequency in summer. 3-4 Neither wheat nor barley is attested described as δημόσιος. We instead resolve δημ(οσίω), seil, μετρώ. An instance is found in O.Douch 1.39, as corrected in O.Douch 2, p. 87. (The presentation there might lead the reader to think that an abbreviation for μέτρορ occurs in the text, but this is not the case.) The same term also occurs, more fully written out, in 312.3. 11. Account. Ca. 350-370. TM# 131059. Inv. 0245. Area 2.1, Room 3, DSU 33, FN 208. 8 χ 10 cm; on both sides; complete but very faded, with the convex side extremely faint. Ala.

4

Concave side τω γεωργ(ω) είς σπείσαι κερ(άμιον) α νης χό(ας) ς- είς κερ(άμια) β μέρους κερ(άμια) μ κερ(άμια) μγ faint traces faint traces

82 8

12

OSTRAKA FROM TRIMITHIS Ψάιτι traces traces δ κερ(άμια) ρκθ έγραψα σοι το ε [..]... ϊνα συ λ α λ ή σ α ι α ύ τ φ επί

τε traces μέρος Convex side λόγ(ος) άναλωμάτ(ων) δαπάνη χό(ας) λ είς κερ(άμια) [ι] μ traces

4

yáp9u traces άμαξηλ(άτη) traces Ψάιτ[ι] traces Faint traces of further writing

Concave 1-2 "To the farmer for libation, 1 keramion." Concave 3 At the start, perhaps δαπάνης, cf. Convex 2. The equation of measures here, at the standard rate of 3 choes to a keramion, is the basis for the restoration of Convex 2. Cf. KAB, pp. 48-49. Concave 4 For meros see the remarks in the introduction above, p. 38. Convex 1-2 "Account of expenses. Disbursement of 30 choes at 10 keramia." For dapanê

see KAB, pp. 32-35. 12. Account of grain. 4th century. TM# 131060. Inv. 0246. Area 2.1, Room 3, DSU 36, FN 210. 11.3 χ 9.2 cm; written on both sides, but only the concave is readable although very faded; complete. Alb with cream slip. Απόλλων Σαραπίω(νος)

4

8

(άρτάβας) δ ς' μά(τια) δ traces τιμ(ής) σίτ(ου) μ(ατίων) δ και μ(ατίων) ξ εν (άρτάβαις) y (άρτάβης) ιβ μ(άτια) η/ (γίνεται) μ(άτια) ι τοπ(ικφ) (άρτάβας) ε μ(άτια) ιβ τοπ(ικφ) (ων) τ(φ) Κόρακ(ος) λόγω μ(άτια) ιβ (άρτάβας) ε εν μοδ(ίοις) (άρτάβας) δ (lines 4-8) "Price of 4 matia of wheat, and 69 (65?) matia in 3 arfabas. 1/12 art., 8 matia, total 10 matia by the local standard. 5 art., 12 mat. by the local standard. From which for the account of Korax, 12 mat. . . . 5 art. in . . modii, 4 art."

TEXTS

83

1 There may be some traces at the end of the line. 2 At the start, one might think of reading γλευκ( ). But the entire account seems to concern grain, so this is unlikely. 4 At the expenditure or topikon rate, 3 art. would come to 69 matia; the number here is not clear enough to justify printing a theta. If 3 art. is the total of the two amounts mentioned in this line, 65 would be a plausible reading (4 + 65 = 69). 5 The calculation is correct: 1/12 of 23 matia is about 2 matia, and adding 8 matia gives a sum of 10. See KAB, pp. 47-48. 8 The number of modii is written on a little knob in the fabric, making the reading very doubtful. Possibly we have ιδ, which would give something close to the standard modius of 3 1/3 to the artaba. 13. Account. Ca. 275-340. TM# 131061. Inv. 0258. Area 2.1, Room 4, DSU 35, FN 223. 8.4 χ 14.4 cm; on both sides; complete but faded. Ala. Convex side τω Άφροδισίωι από Θωθ α έως ΤΟβι β α 4-5 traces ακ (άρτάβην) α

7-8

traces θατ

"Qjpcoi μά(τιον) α Πε έλαίο(υ) 12 ο μ(άτιον) α 13-14 traces α 16-19 faded traces covered by large dark discolorations

4

Concave side (turned 180 degrees) Πμουν Ψωι Ψ traces (άρτάβην) α Φ traces (άρτάβην) α traces (άρτάβην) α traces (άρτάβην) α traces

(Convex 1-2) "For the Aphrodisieion, from Thoth 1 to Tybi 2."

84

OSTRAKA FROM TRIMITHIS

Convex 1 The reading is uncertain in places, but the delta seems certain and sufficient to exclude reading the end as Πμουν Ψωι. A sanctuary of Aphrodite is not mentioned elsewhere in the documents from the oasis, as far as we know. An Aphrodisieion in the Small Oasis occurs frequently in the documents (see most recently Calderini-Daris, Dizionario Suppl. 3, 22 with earlier references; also Wagner, Les Oasis, 199-200). Convex 2 The reading is very conjectural. Convex 4 This line was squeezed in afterward. 14. Memorandum or account. Ca. 350-370. TM# 131062. Inv. 0271. Area 2.1, Room 4, DSU 23, FN 237. Found on wall F37. 8.5 χ 6.8 cm; on convex side; complete. Very dark color. Alb.

4

Ψάις Ψάις έλαίο(υ) Πμουν θαν κεράμιον ένα (γίνεται) κερ(άμιον) α

(lines 2-5) "Psais son of Psais, oil. Pmoun

than, one keramion. Total, 1 keramion."

1 Probably a numeral, perhaps alpha. 2 The second name should be Ψάιτος. 15. Account of wine. Ca. 350-370. TM# 131063. Inv. 2058+3061. Area 2.1, Room 8, DSU 74, FN 8+31. 10.3 χ 6.1 cm; on convex side; broken at lower right. Alb.

4

8

από μαρ(ίων) ιθ Κύρω μάρ(ιον) α δι' Ευτυχίου μάρ(ια) δ Άματίω μάρ(ιον) α 'COpícoví ταβουλαρ(ίω) μάρ(ια) δί(ά) Παίωνος [μάρ(ι-) ] δι' Εϋτυχίο[υ μάρ(ι-) ] ]..[ ' "From 19 maria: To Kyros, 1 marión Through Eutychios, 4 maria To Amatios, 1 marión To Horion the tabularius, - maria Through Paion, - maria Through Eutychios, - maria"

TEXTS

85

2-3 There is a mark between these lines at the left edge looking something like an opening round bracket. 5 The top of a vertical stroke is preserved where the number of maria should be. Iota (10) seems the most likely reading, although it does not leave very many maria for the remaining entries in the account. For the position of tabularius in this period, see e.g. P.Oxy. 50.3581, in which the tabularius processes a deed of divorce, with references given there. 16. Account of money. Ca. 350-370. TM# 131064. Inv. 2067. Area 2.1, Room 8, DSU 74, FN 7. 8.5 χ 8.5 cm; on convex side of a rim fragment; complete. Alb.

4

(ων) (?) Ψενα(μούνιος) ΠαχοΟμις Καρούρ(εως) τάλ(αντα) /Ετ δ(ιά) Έπιμάχ(ου) Παθώτης Ψενν(ήσεως) τάλ(αντα) /Β δ(ιά) Έπαφ(ροδίτου) επ Ψενττνούθ(ης) Λουια τάλ(αντα) /Β δ(ιά) Σάιτ(ος) Βησατ(ος)

"From which, from (?) Psenamounis. Pachoumis son of Karouris, 5,300 tal. through Epimachos; Pathotes son of Psennesis, 2,000 tal. through Epaphroditos ...; Psenpnouthes son of Louia, 2,000 tal. through Sais son of Besas." 2-4 The thousands of talents mentioned in this account suggest a date after 351. 17. Account. Ca. 350-370. TM# 131065. Inv. 3000. Area 2.1, Room 8, DSU 74, FN 15. 7.8 χ 10 cm; on convex side; complete.

4

Πμ(ουν) Τκηλε ξύλ(ων) κτή(νη) ς άχύρ(ου) μώ(ια) δ δ' υπό ( ) "Pmoun Tkele, 6 donkey-loads of wood; 4 moia of chaff; 4 ..."

1 For Pmoun Tkele, cf. KAB 1408. 4 Perhaps ΰποζύγ(ια). 18. Account. Ca. 350-370. TM# 131066. Inv. 3007. Area 2.1, Room 8, DSU 74, FN 12. 3.9 χ 5 cm; on convex side; broken at left, upper right, and bottom. ] όμοΚως)

86

4

8

OSTRAKA FROM TRIMITHIS ] της γυναικ(ός) Πουαε[ι ] Αμμωνίου ] τιμ(ής) όρνιθίο(υ) ] δι(ά) Δε μάτ(ια) ] δι(ά) Η faint traces ]δι(ά) ].[

5 This should be a name, but we cannot see one of the known figures in the Trimithis ostraka here. 6 Perhaps Ήρακλείδου, no doubt abbreviated; only the faintest traces remain. 19. Account of money. Ca. 350-370. TM# 131067. Inv. 3151. Area 2.1, Room 8, DSU 85, FN 62. 7 χ 6.5 cm; on both sides; complete. There are signs that earlier text has been washed out on both sides, but it is unclear how much. Concave 4-7 are fainter than the rest of the text and look somewhat washed out. Alb.

4

8

4

8

Convex side χρ(υσοθ) νο(μισμάτια) δ8 ταλ(άντων) (μυριάδας) y /Γω νacat I line (perhaps washed out) (ταλάντων) (μυριάδας) δ /Εω όμ(οίως) Τερητ( ) apy(upiou) traces [[και ]] είς τεσσάρας ήμισυ ápyup(íou) ταλ(άντων) (μυριάδας) β /Εχ αδη( ) νο(μισματίου) α και ταλ(άντων) Αχ Concave side (turned 180 degrees) Πανθρ(ις) ήμ(έρας) S/ Ψεναμοθν(ις) ήμ(έραν) α Σαμουν ήμ(έραν) α 'ϋΟρίων Είραμου ήμ(έρας) aSd' το ήμ(ερήσιον) (τάλαντα) /Α Ψάις Πανύρ(εως) ήμ(έραν) α Ψάις traces [] pu traces

(Convex) "Gold solidi, 4 Vi, 33,800 talents. 45,800 tal. Likewise, Teret( ), silver . . . for four and a half, 25,600 talents. Ade( ), 1 solidus and 1,600 talents." (Concave) "Panuris, Vi day. Psenamounis, 1 day. Samoun, 1 day. Horion son of Eiramos, 1 % day; the daily rate, 1,000 tal. Psais son of Panuris, 1 day. Psais . . ."

TEXTS

87

Convex 1-2 At a rate of 4 Vi solidi equaling 33,800 talents, the solidus would be 7,511 talents and the pound of gold would be 540,800 talents. That is a figure characteristic of the period 352-360; see Bagnall, Currency and Inflation in Fourth Century Egypt (1985) 4344. It is not clear how the mention of "four and a half" in line 6 is to be understood; the form is masculine/feminine and should not modify an understood νομισμάτια. In the absence of more information, speculation about scribal error of gender is risky. Convex 8 What αδη( ) refers to we do not know. Concave 5 If this is a daily rate for the use of a well, it would imply a weekly rate of roughly a solidus (7,511 talents, see note to Convex 1-2). 20. Account. Ca. 300-350. TM# 131068. Inv. 3405+3408. Area 1.3, Room l, DSU 16, FN 12+13+15. 13.6 χ 6.3 cm; on convex side; complete. The hand is very similar to that of 21, but they do not belong to the same sherd. 2

καμηλ(άτη) μάρ(ια) κ Έρμησία λινού(φω) ς

"To the camel-driver, 20 maria. To Hermesias the linenweaver, 6." 2 A Hermesias appears in 388, found in the same DSU; he is not there designated a linenweaver, but it is likely to be the same person. 21. Account of disbursements. Ca. 240-340. TM# 131069. Inv. 3406. Area 1.3, Room l, DSU 12, FN 21. 7.3 χ 9.5 cm; on convex side; broken at right and top; Ala.

4

8

].[ είς Πμο(υν) Άραυ 'Άμμω(νίω) Κόρακ(ος) καμηλίτη Πλου(τ- ) λινού(φω) Ψεναμ(ούνει) Πα[ είς Ψε( ) [ ρα[

"To Pmoun Harau. To Ammonios son of Korax, camel-driver - - . To Plout( linenweaver - -. To Psenamounis son of Pa[. To Pse( ) "

) the

2 Pmoun Harau also appears in 134 (Area 4.1) and 256 (Area 2.1). 4 The form καμηλάτης with alpha appears in 71 (Area 2.1). 5 Πλουτίων and Πλουτάμμων are both plausible resolutions. 7 With είς, Pse( ) is presumably a place name rather than person. Perhaps Pselthis (299.4) is intended. 8 Either ραβδούχος (teamster) or ράπτης (tailor) would be a possible restoration.

88

OSTRAKA FROM TRIMITHIS

22. Account. Ca. 300-350. TM# 131070. Inv. 3411. Area 1.3, Room 2, DSU 15, FN 8. 5.7 χ 3.7 cm; on convex side; broken at right; Ala.

4

υδ(ρευμα) Πμο(υν) Έμβ(ωου) γόμ(οι) ιέ Πμο(υν) Όσιρε γόμ(οι) κβ [Π]μο(υν) όνηλ( ) γόμ(οι) δ ] παραδοθ( )

"Hydreuma Pmoun Emb(ôou), 15 loads. Pmoun Osire, 22 loads. Pmoun Onel( loads . . . delivered"

), 4

1 For the resolution of this abbreviated well-name, see 101 below. 2 Pmoun Osire is attested as Pmoun Osirios in 171 (Area 4.1). 3 Probably to be resolved όνηλ(άτου) (or plural). 4 παραδοθ(έντα) or another case: "delivered". The partially preserved letter at left is probably alpha (τα?). 23. Account. Ca. 300-350. TM# 131071. Inv. 3412. Area 1.3, Room l, DSU 16, FN 10. 12.1 χ 8.2 cm; on convex side; broken at left; large surface chip lost at lower right before use; Ala.

4

] vacat p ] vacat νε ] vacat πδ όμ(οίως) κ$~ όμ(οίως) ν ] vacat λδ όμ(οίως) β ].θ( ) ιη ] ( ) άπολ( ) σα μβ η ]..( ) μβ [ό]μ(οίω$) [

2 ν ex μ. 3-4 These lines appear to list a sequence of unrelated amounts rather than any arithmetical operation. 4 At the start, kappa or lambda would be possible. What is transcribed here as omicron of όμ( ) seems to have a vertical stroke making it look like a d. 5 η ex ς. 6 Perhaps a name, Άπολ( ). 7 όμ(οίως) may have stood in the lacuna before the numeral 42. Presumably a numeral was written before it.

24. Account of doum-fruits on clay tablet. Ca. 300-350. TM# 131072. Inv. 11048. Area 1.3, Room 7, DSU 48, FN 17. 6.6 χ 8.4. Incised on one side of unbaked clay tablet. Complete.

TEXTS

4

89

Άθύρθ κού(κια) μ (ων) Σύρω Έρμα( ) ιςΨεναμ(ούνει) Νίλου ιδ Ψεναμ(ούνει) Άμμω(νίου) iy

"Hathyr 9. 40 doums. From which: To Syros son of Herma(mmon?), 16; to Psenamounis son of Nilos, 14; to Psenamounis son of Ammonios, 13." 2 The numbers below total 43, not 40. We take the numbers to refer to the large fruits of the doum-palm (similarly in O.Stras. 601 v.l), rather than to trees or other products from them. For the doum-palm in Dakhla, cf. P.Kellis 1 G. 52.4-5n. (where, however, we do not see why the editor describes the adjective κούκινος as "unattested" when it appears in LSJ). 3 Perhaps Έρμάμ(μωνος), not simply Έρμα. 25. Account of wine. Ca. 275-350. TM# 131073. Inv. 11054. Area 2.1, Room 9, DSU 146, FN 15. 10.5 χ 10.8 cm; on convex side; complete but badly abraded, especially in the middle. Ala.

4

Παβώς [ Ψάις και Άμ [ Ψενα(μοθνις) [ Ψε[ν]νήσις [ entirely abraded faint traces [ Αμμ traces [

] μάρ(ια) η ] μάρ(ια) η ] μάρ(ια) η ] μάρ(ια) η ] μάρ(ια) η

(Lines 2-4) "Psais and Am—, 8 maria; Psenamounis . . . 8 maria; Psennesis . . . 8 maña."

26. Account of oil. Ca. 315-350. TM# 131074. Inv. 11116. Area 1.3, Room 9, DSU 73. 8.5 χ 6.4 cm. Written on both sides. Complete.

4

8

Convex Πετοσΐ(ρις) Ψενθηθ κερ(άμια) içr Σαμώς Πατχωτ( ) κε(ράμια) ιη και ύπ(έρ) προσθήκ(ης) κε(ράμιον) α, και ύπ(έρ) προσθ(ήκης) κερ(άμιον) α Βήκις Ψάιτ(ος) έλ(αίου) χό(ες) λβ και ύπ(έρ) προσθήκ(ης) χό(ες) λβ. Βήσις "Q3p[o]u χό(ες) ιέ και ~χό(ες) β' Ψεννή(σις) ό κ(αι) Λυκαρίω(ν) χό(ες) ε και ύττ(έρ) προσθηκίης) χο(Ος) α Πετοσΐ(ρις) Ψενθώ(του) χό(ες) κ και ύπ(έρ) προσθήκης) χό(ες) β

90

11

OSTRAKA FROM TRIMITHIS Concave Βήκ(ις) διάκων χό(ες) ιέ και ύπ(έρ) προσθήκ(ης) Χό(ες) β

(Convex) "Petosiris son of Psentheth, 16 keramia. Samos son of Patchot( ), 18 keramia and 1 keramion for extra; and 1 keramion for extra. Bekis son of Psaios, 32 choes of oil and 32 for extra. Besis son of Horos 15 choes and 2 choes. Psennesis also called Lykarion, 5 choes and 1 chous for extra. Petosiris son of Psenthotes, 20 choes and 2 choes for extra." (Concave) "Bekis the deacon, 15 choes and 2 choes for extra." 1 The name Ψενθηθ is not otherwise attested; it is not given any sign of abbreviation. 2 Neither name appears elsewhere in texts from the Oasis. 3 προσθ(ήκης): For the term as a "supercharge," presumably as a superindiction, see A. C. Johnson and L. C. West, Byzantine Egypt: Economic Studies (Princeton 1949) 241-42. The word can hardly have the same meaning here, as one of the "extra" payments is 100% of the original. While arrears on rent are one possibility, the nature of the additional charge here ultimately remains unclear. See the discussion at p. 52 above. 6 It is striking that no Λυκ- names are found in the material from Kellis. Lykopolis was the terminus of the direct road from Dakhla to the valley. 10 διάκων: the appearance of the term here provides an indication for the date of the texts from Area 1.3. See further remarks on p. 49, above. 27. Account. Ca. 350-370. TM# 131075. Inv. 11117. Area 2.1, Room 17, DSU 153, AC 55. 10 χ 5.9 cm; on convex side; broken at lower right and faded at right. Ala.

4

υδ(ρευμα) Πμ(ουν) "QJpoy Αμμώνιος Πεβώς Κωμητ faded Πεβωκ αδελφός Μακάριος πρε(σβύτερος) εν Πμο(υν) Έ[ Αμμώνιος Άμμω[νίου

"Hydreuma Pmoun Horos. Ammonios son of Pebos. Komet—. Pebok his brother. Makarios, presbyteros in Pmoun —. Ammonios son of Ammonios." 5 Μακάριος: this Christian name is common at Kellis. See also 325. The name makes it more likely that presbyteros here refers to a Christian priest rather than to a village elder. 28. Account of wheat and barley? 3rd-4th century. TM# 131076. Inv. 11125. Area 4.1, DSU 81, AD 50. 3.4 χ 4 cm; on convex side; broken at right. Ala. Σαραπί[ων ττυροΟ-[

TEXTS

4

91

p( ) (πύρου) (άρτάβας) [ κρ(ιθής) (άρτάβας) [ κ.[

2 The horizontal line broken off is presumably the start of the sign for arfabas. 3 The rho may represent (έκατοσταί), but in the absence of more context we have not resolved it in the text. 29. Account. Ca. 350-370. TM# 131077. Inv. 11136. Area 2.1, Room 15, DSU 152, AD 55. 6 χ 6.1 cm; on convex and concave sides; complete. Alb.

4

Convex Πμο(υν) ε χόρτος traces οίνου μάρ(ια) β άργυρ traces traces οίνου μά[ρ(ια) ]

Concave 1-3 traces 4 traces ε traces μάρ(ιον) α traces β traces 30. Account of tiphagia. Ca. 350-370. TM# 131078. Inv. 11175. Area 2.1, Room 15, DSU 157, FN 94, AD 55. 4.3 χ 6.3 cm; on convex side; broken at left. Ala.

4

] τι(φάγια) λ ] τι(φάγια) λ ] τι(φάγια) λ ] τι(φάγια) κ ]ςτι(φάγια) μ

For tiphagia, an unidentified product, see Bagnall, Thanheiser, and Worp, ZPE 122 (1998) 173-88, and n. 25 above, with 34, 42 and 301 herein, an order for delivery and two accounts respectively. 31. Account or receipt concerning wine? Ca. 350-370. TM# 131079. Inv. 11193. Area 2.1, Room 15, DSU 157, FN 107, AD 55. 9.7 χ 7 cm; on concave side; complete but faded. Alb. There could be a line before line 1.

92

4

OSTRAKA FROM TRIMITHIS [. Ji.S δέκα οίνου κεράμιον εν και τρία δ// c traces

μάρια

32. Account? 360/1? ΤΜ# 131080. Ιην. 11227. Area 2.1, Room 15, DSU 157, AD 55. th 3.1 χ 4 cm; on convex side; broken at right. Iron-rich pre-4 cent, fabric.

3

δ (έτους) Έπειφ χόρτ(ου) δέσμ(αι) ι [ faint traces

"4th year, Epeiph. ![.] bundles of hay. . ." 1 It is possible that an iota stood at the start, making ιδ, but the traces are extremely faint. S/ ostr. The date is calculated on the assumption that the reference is to the in diction. 33. Account? Date unknown. TM# 131081. Inv. 11306. Area 4.1, DSU 79, AP 49. 5 χ 5.9 cm; on concave side; complete? 7 lines, entirely faded except for a few letters at the ends of lines, seemingly the ends of names in the dative. Probably an account. 34. Account of tiphagia. Ca. 350-370. TM# 131082. Inv. 11622. Area 2.1, Room 13, DSU 199, FN 11. 7.5 χ 11.4 cm; on convex side; complete but very faded. Alb.

4

8

ιδ' Τύραννος μάγειρ(ος) τι(φάγια) ρκ faint traces faint traces faint traces ι. Ψάιτι πελεκά τι(φάγια) λβ ιέ Π λινούφω τι(φάγια) ι ιέ Ψάις α άπα τι(φάγια) λη Γαίωνι κλιδοποειώ τι(φάγια) ιβ ρατι προφ(ήτη) τι(φάγια) ι

Line 1: "14th, Tyrannos the cook, 120 tiphagia"; Lines 5-9: "l-th, to Psais the axemaker, 32 tiphagia; 15th, to Ρ—, linen-weaver, 10 tiphagia; 15th, Psais . . ., 38 tiphagia; —th, to Gaion the locksmith, 12 tiphagia; —th, to -ras, prophètes, 10 tiphagia."

TEXTS

93

1 For tiphagia, see 30 above. 5 This word for axemaker is rare. It occurs in O.Wilck. 720.3, where Wilcken took it as a personal name but B. Olsson, APF 11 (1935) 219 recognized it as an occupation ("Das Wort πελεκάς ist meines Wissens neu, aber regelmäßig gebildet . . ."). LSJ Suppl. lists it

from this occurrence. 8 Not Παιωνι as in 15. The second omicron of κλιδοττοειφ (Ι. κλειδοττοιειφ; "locksmith") is a correction from iota. 9 Prophetai occur in a number of the Kellis ostraka; cf. O.Kellis 133, 211, 213, 215 and 280.

35. Account of chaff. Ca. 350-370 (354/5?) TM# 131083. Inv. 11624. Area 2.1, Room 13, DSU 199, FN 12. 13.2 χ 8.9 cm. On convex side. Broken at upper right, perhaps also lower left. Extremely faded.

"ïys [ traces traces

4

[

[ ] μώ(ιον) α μώ(ιον)] α μώ(ιον)] α μώ(ια)] β ] μώ(ι-). ] μώ(ιον) α ] μώ(ια) β ] μώ(ια) y ] μώ(ιον) α ]μώ(ι-)[

8

12

1 It is uncertain whether the sinusoidal curve should be resolved as (έτους) or taken as an indiction year, but the latter seems more likely given the context, which belongs to the last phase of occupation. 36. Account. Ca. 240-275. TM# 131084. Inv. 11626. Area 1.2, DSU 5, FN 5. 7.1 χ 3.7 cm; written on both sides; incomplete. Broken at upper left and bottom (convex surface), but uncertain if elsewhere. Convex surface turning to powder; concave surface cracking. Probably BIO.

3

Convex [ ]οςΣερήvov έ'σχον (ττυροΟ) (άρτάβας) β (δραχμάς) μ Concave ΠαΟνι

94

3

OSTRAKA FROM TRIMITHIS δια Μύρωνος μ[ατ( ) ] δια όνηλ(άτου) ματ( )

Convex: "ΝΝ (to?) Serenos: I received 2 art. of wheat, 40 drachmas." Concave: "Pauni. Through Myron, [.] matia. Through the donkey-driver, - matia." Convex 3 A wheat price of 20 dr. per art. indicates a date before 275 but later than the price rises of the later second century. See D. Rathbone, "Prices and Price Formation in Roman Egypt," in J. Andreau, Ρ Briant, and R. Descat, eds., Économie antique. Prix et formation des prix dans les économies antiques (Saint-Bertrand-de-Comminges 1997) 183-

244.

37. Account of oil. Ca. 350-370. TM# 131085. Inv. 11635. Area 2.1, Room 14, DSU 194, FN 23. 6.7 χ 9 cm; on convex side; complete. Alb. The surface has much dirt and other distracting material that cannot be removed without harm to the ink.

4

Ψεναμ(οθνις) άκτάρι(ος) Μαρτυρίας Άχιλλεύς ΠαμοΟρις Σιλβανός Σεράπις

μάρ(ια) η χό(ες) y χό(ες) ς χό(ες) δ χό(ε5) δ8 χό(ες) δ

"Psenamounios, actuarius, 8 maria Martyrios, 3 choes Achilleus, 6 choes Pamouris, 4 choes Silvanus, 4 Vi choes Serapis, 4 choes." 1 For the term actuarius, see I.-M. Cervenka-Ehrenstrasser and J. Diethart, Lexikon der lateinischen Lehnwörter in den griechischsprachigen dokumentarischen Texten Ägyptens l (Vienna 1996) 63-66. The rendering into Greek used here is that usual in the early Roman period. The actuarius was responsible for record-keeping related to military provisioning. See F. Mitthof, Annona Militaris: Die Heeresversorgung im spätantiken Ägypten (Florence 2001) 1: 152-56 (further references in the index, 2: 636) with bibliography. 38. Account of cotton. Ca. 350-370. TM# 131086. Inv. 11643. Area 2.1, Room 13, DSU 208, FN 28. 6.5 χ 4.5 cm; on convex side; complete. Alb. λιθ( ) d ιέ cpoçr λιθ( )ylX"f

λιθ( ) L y ü ρττ yí(v.) λιθ( ) α β'

TEXTS

4

λιθ( ) με ρπ

95

^

"1/4 1/15 1/576 lith( ); 1/3 1/10 1/30 1/90 lith( ); % 1/3 1/15 1/180 lith( ); 1/45 1/180 lith( ); total, 1 2/3 1/90 lith( )." The system of fractions in use here is visible as far as 1/576 in line 1, but even there, 1/15 requires 5 to be a factor, and the common denominator must also accommodate 1/180 in 8 2 lines 3 and 4. This must at a minimum be 2,880 (= 2 χ 3 χ 5). As 288 is the number of grams in the Roman pound, it is evident that the lith( ) uses a fraction system based on 10 Roman pounds (or, conceivably, a multiple thereof). See introduction, p. 41, and 44 below for cotton and cotton-measures in the oasis texts. 39. Account of days. Ca. 350-370. TM# 131087. Inv. 11645. Area 2.1, Room 13, DSU 208, FN 31. 5.2 χ 4.6 cm; on convex side; complete. Alb.

4

Πμ(ουν) Φιβ ήμ(έρα) α Πμ(ουν) Ψωιν ήμ(έρα) α των ήμ(ερών) β Πμ(ουν) Μαρκ( ) ήμ(έρας) S "Pmoun Phib, 1 day. Pmoun Psois, 1 day. For the two days. Pmoun Mark( ) (?), % day."

3 One would expect to find some indication of a charge for the two days, but nothing is written. 4 The well name may be Μα ρω ( ). At the end, S is written over alpha. 40. Account of grain, ca. 275-340. TM# 131088. Inv. 11654. Area 2.1, Room 9, DSU 212, FN 52. 4.9 χ 5.8 cm; on convex side; complete. Alb.

4

4

Χοίαχ y (άρτ.) ιη β μ όμοίω(ς) (άρτ.) ιβ η' Ήρακλ( ) (άρτ.) β εξ ων είσδοχ(ή) (άρτ.) ιβ η Άσκλοτος (άρτ.) ιζ β μ Έρμοδ(ωρ-) (άρτ.) α Ήρακ( ) (άρτ.) β "Choiak 3,18 art., 2 m(atia) (?) Likewise, 12 1/8 art. Herakl( ), 2 art. From which, for the receipt, 12 1/8 art.

96

OSTRAKA FROM TRIMITHIS Asklas, 17 art. 2 m(atia) (?) Hermodoros, 1 art. Herak( ), 2 art."

1-7 The total amount in lines 1-3 is the same as the total amount in lines 4-7, 32 1/8 artabas 2 m(atia?). The εξ ων in line 4 would suggest that lines 1-3 represented receipts, from which expenditures were made or accounts were credited as enumerated in lines 4-7. Perhaps the 1 artaba in line 6 indicates what is left from the total given in line 1 after the amount for Asklas has been subtracted. 1, 5 The numbers after the quantity of artabas are unclear. They appear to be a beta followed by mu, although the mu is not entirely clear in line 1. We have supposed that 2 matia are meant, although we would expect to find μ(άτια) β in that order if that were the case. The alternative would be to suppose a fraction of 1/40, which we have not found elsewhere, or that the scribe intended 1/48 but simply omitted the second digit. 4 Whether είσδοχ(ή) (in whatever case may have been intended) refers to a written receipt or to an act of receiving is not clear. The amount is the same as in line 2. 5 The traces seem to demand one more small letter in the ending, suggesting a reading of ΆσκλατοΟς, an otherwise unattested form. 41. Account of hay. 358/9 (?). TM# 131089. Inv. 11655. Area 2.1, Room 13, DSU 216, FN 74. 5 χ 18.5 cm; on both sides; complete. Alb.

4

8

12

16

20

Convex χόρτ(ος) β' ίνδ(ικτίονος) Πανουβα δέσ(μαι) ρ Παύλος Σαραπατ(ος) δέσ(μαι) ρ Παστωου δέσ(μαι) ρ Άμμων "QJpou δέσ(μαι) ρ Φιλάμμων δέσ(μαι) ρ Σαραπίων Σεναμού(νιος) δέσ(μαι) ρ r GOpoç Σαττίω(νος) δέσ(μαι) ρ Άττολλώς δέσ(μαι) ρ Σάϊς Άμμονι δέσ(μαι) ρ Παρά δέσ(μαι) ρ "Ηρών δέσ(μαι) ρ Λουιάμμονι δέσ(μαι) ρ Φιλοσερττι δέσ(μαι) ρ Άμμων Ταμουν δέσ(μαι) ρ Κελε δέσ(μαι) ρ εν αποθέτω π(αρά) Παρά δέσ(μαι) τ τοις σκρίβαις δέσ(μαι) σ Λογγίνω δέσ(μαι) ρ Σαμβατίω δέσ(μαι) ρ βσ

TEXTS

97

Concave 3

τοίς σκρίβαις δέσ(μαι) σ Aoyyívcp δέσ(μαι) ρ Σαμβατίω δέσ(μαι) ρ nd

(Convex) "Hay for the 2 indiction. Panoubas, 100 bundles. Paulos son of Sarapas, 100 bundles. Pastoou, 100 bundles. Ammon son of Horos, 100 bundles. Philammon, 100 bundles. Sarapion son of Senamounis, 100 bundles. Horos son of Sapion, 100 bundles. Apollos, 100 bundles. Sais son of Ammon, 100 bundles. Para, 100 bundles. Heron, 100 bundles. Louiammon, 100 bundles. Philoserpi, 100 bundles. Ammon son of Tamoun, 100 bundles. Kele, 100 bundles. In storage from Para, 300 bundles. To the scribes, 200 bundles. To Longinus, 100 bundles. To Sambation, 100 bundles. 2,200." (Concave) "To the scribes, 200 bundles. To Longinus, 100 bundles. To Sambation, 100 bundles." Conv. 17-18: "in storage from Para," perhaps the same Para as in line 11. The other entries in this account alternate between giving names in the nominative and the dative, perhaps indicating a distinction between payments of hay to and from the individuals in question. But it is also possible that the scribe's use of case is haphazard. The fact that the total in line Conv. 22 is an accurate sum of the figures in lines 2-21 supports that supposition. The explicit reference to storage in lines 17-18 presumably means that the other incoming payments in this accounts were not similarly set aside. Conv. 19: For another entry τοίς σκρίβαις, see 292, a receipt issued by Paesis for delivery of 10 matia through Domnion. Conv. 22: The final number here (2,200) is an accurate total for the number of bundles of hay listed in lines 2-21. Cone. 1-3: These lines repeat lines 19-21 on the convex side. 42. Account of tiphagia. Ca. 350-370. TM# 131090. Inv. 11665. Area 2.1, Room 13, DSU 216, FN 77. 8 χ 8.3 cm; on both sides; complete? or broken at top? A2b. Convex

4

8

τι( ) α εσχ( ) είς λάχανο(ν) είς είς ύπηρ traces traces είς α είς λάχαν(ον) τι(φάγια) υπέρ έλαί(ου) είς κελ

τι(φάγια) ιδ τι(φάγια) κ τι(φάγια) η τι(φάγια) α τι(φάγια) η τι(φάγια) δ τι(φάγια) δ τι(φάγια) ιέ τι(φάγια) ιέ τι(φάγια) ι

π χ( ) τι(φάγια) ι είς τι(φάγια) ι τι(φάγια) . είς τι(φάγια)

πβ

98

4

OSTRAKA FROM TRIMITHIS Concave (τάλαντα) /Γ [(τάλαντα)] /Γ [(τάλαντα)] τλ [(τάλαντα)] /Ατ βερ( ) [(τάλαντα)] /Αρ Φιλ( )

Convex 1 One might imagine a heading like λόγ(ος) τι(φαγίων) αφ' ων έ'σχον. But the traces are not sufficient to confirm or disprove this reading. For tiphagia, see 30 above. Convex 4 Presumably ύπηρεσίαν, but the degree of abbreviation is impossible to determine. Convex 11 κέλλαν seems likely, but we cannot read anything beyond the start. Concave 5 Alpha corrected. 43. Account of oil (?). Ca. 275-350. TM# 131091. Inv. 11671. Area 2.1, Room 9, DSU 212, FN 47. 3 χ 3.3 cm; on concave side; complete. Ala.

3

έλ(αίου) Ψεναμο(Ονις) Ψάιτο(ς) κο(τύλη) α Αμώνιος Ροιμη(τάλκου) κο(τύλη) α

"Of oil: Psenamounis son of Psais, 1 kotyle; Amonios son of Rhoimetalkas, 1 kotyle." 2-3 κο( ): This can only be κο(τύλη), which measured a half of a sextarius. We have not found any previous use of the measure in material from the Dakhla Oasis, and it is extremely rare in general in the fourth century papyri. That may suggest an early date for this text. 3 This appears, surprisingly, to be the first secure attestation of this Thracian royal name in the papyri. See I.Syringes 292 for the only other Egyptian instance. 44. Account of cotton. Ca. 350-370. TM# 131092. Inv. 11668. Area 2.1, Room 13, DSU 216, FN 64. 8 χ 5.6 cm; on both sides; complete. Ala.

4

Convex Πετενεφ(ώτης) πρ(εσβύτερος) λιθ( ) KÇT Ly Ψάι 5 Ψάιτ(ο 5 ) λιθ( ) ία L íi Αμμώ(νιος) "QJpou λιθ( ) ιβ ξ Ψάι 5 "ÜJpou λιθ( ) α β' ε Αρ.. ( ) λιθ( ) α Ly yí(v.) λιθ( ) vy L y [d λ]

TEXTS

4

99

Concave (inverted) Πετενεφ(ώτης) Άμμω(νίου) λιθ( ) α Ly υμ( ) Άμμω(νίου) λιθ( ) c c y i Πετε ( ) και ΠφλύβΟος) λιθ( ) y! λ Ίσιδώρα Λεω( ) λιθ( ) IX yí(v.) λιθ( ) s ^ ϊΐ ομού λιθ( ) ξ L ιέ

Convex: "Petenephotes pr(esbyteros), 26 Vi 1/3 lith( ); Psais son of Psais, 11 Vi 1/15 lith( ); Ammonios son of Horos, 12 1/60 lith( ); Psais son of Horos, 1 2/3 1/5 lith( ); Ar. . ., 1 Vi 1/3 lith( ); total, 53 Vi 1/3 [1/4 1/30] lith( )." Concave: "Petenephotes son of Ammonios, 1 Vi 1/3 lith( ); . . . son of Ammonios, 1 1/3 1/10 lith( ); Pete. . .( ) and Polybios, 3 1/5 1/30 lith( )." For cotton measures, see the introduction, p. 41 and 38 above. Convex 6 This would be the correct sum of lines 1-5. There are perhaps traces of the last two fractions, but we cannot read them with any confidence. Concave 5 This is the correct sum of lines 1-4. Concave 6 This is the correct sum of Convex 6 and Concave 5. 45. Account. Ca. 275-350. TM# 131093. Inv. 11660. Area 2.1, Room 9, DSU 212, FN 44+43+46. 7.4 χ 4 cm; on convex side; complete. Ala.

4

Π .[..]..( ) ΨεναμοΟνις Πετοσί(ριος) . ßL Παθώτ(ης) Σαμβας ßL Σαμβας Άρσιήσιο(ς) ßL

"... Psenamounis son of Petosiris, 2 Vi; Pathotes son of Sambas, 2 Vi; Sambas son of Harsiesis, 2 Vi." 2 The sigma of Πετοσί(ριος) is crossed by a vertical stroke rising well above the line and descending some distance, as if in abbreviation. There is then another letter (or two?) immediately after this, the significance of which is obscure to us: part of the name still, a unit, or something else? 46. Account of bread (?). Ca. 275-350. TM# 131094. Inv. 11673. Area 2.1, Room 9, DSU 212, FN 42 + 39. 6.7 χ 7.3 cm; on convex side; broken at left and below. Alb.

4

]j|S Πετεχώντος δαρτ( ] vacat δαρτ( ]ις κελλαρι(τ- ) δαρτ( ] ήσιος δαρτ( ]ου δαρτ(

)β ) αL )α )α )α

100

OSTRAKA FROM TRIMITHIS ]δαρτ( ) [ ] vacat δα[ρ]τ( ) [

On the concave side are a few remains of letters, evidently from the ends of lines; they could be more of similar accounts. 1 Lambda or mu is likely, perhaps Σελ]λής. We have not found a parallel for δαρτ( ), the reading of which seems secure from the totality of the text, although uncertain in some lines. One might suppose it to be a form of the adjective τετράρταβος, found as τεταρτάβου5 in SB 6 9210.5. The note of J. Bingen in the ed.pr. (Cd'É 19 [1944] 273) offers a number of parallels for the formation. But by itself here it seems attractive to think of a double ζεύγος of άρτοι, or four loaves as a unit, to be resolved (τέτρ)αρτος, for which we have, however, no good parallel. 3 The kappa is not an easy reading, but it is not easy to see anything else. This is presumably a form of κελλαρίτης or another title from the same stem. 47. Account? Ca. 275-350. TM# 131095. Inv. 11674. Area 2.1, Room 9, DSU 212, FN 56. 3.9 χ 5 cm; on both sides; broken at lower right (convex), and the surface of the convex side is lost in places. Very faded. B group. Convex Ψεν χα( )β 4

α Ψεμ... μ traces ων αρ ση[

Concave 2-3 faint traces 4 Ψεν π α Πα Χ()α

Convex I It is possible that χα( ) is the abbreviation of a unit before the number 2, cf. Concave 7. But we do not see a good resolution for it here. Convex 5 There is a space between rho and sigma. The unread letter looks like sigma or epsilon rather than pi. 48. Account of chaff. Ca. 350-370. TM# 131096. Inv. 11693. Area 2.1, Room 13, DSU 199, FN 89 + 91. 6.1 χ 7.6 cm; on convex side; broken at upper left, upper and lower right, bottom. Ala. ]v.. [

TEXTS

4

8

101

].αις μώ(ια) .[ ]....μώ(ια) .[ ] μώ(ια).[ [δι(ά) Ή]φαιστάμμων(ο$) μώ(ια) [ δι(ά) Μέμνωνο(ς) μώ(ια) [ δι(ά) Παμουν μ[ώ(ια) δι(ά) Ήφαιστάμμω(νος) [μώ(ια) δι(ά) Μέρσιος [μώ(ια) [δι](ά)ϋ... [ '

(5-9) "through Hephaistammon, - moia; through Memnon, - moia; through Pamoun, moia; through Hephaistammon, - moia; through Mersis, - moia." 49. Account of vinegar. Ca. 275-350. TM# 131097. Inv. 11709. Area 2.1, Room 9, DSU 218, FN 80. 7.7 χ 10.8 cm; on concave side; broken at upper left, upper right, and right. Alb. In an unpracticed and inconsistent hand, with poor control of pen and ink. The readings are very uncertain.

4

]... πίστη .[ δι' Ήρακ( ) οξ(ους) [ δι(ά) οξ(ους) ξ(έσται) [ δι' Άρης ö£(ous) [

50. Accounts of hay and another commodity. Ca. 275-350. TM# 131098. Inv. 11722. Area 2.1, Room 9, DSU 230, FN 106. 6.5 χ 8 cm; on both sides; complete. A2b.

4

3

Concave Πμο(υν) Άλεξάνδρο(υ) υίός Στεφάνο(υ) δέσμ( ) Άμμων Μέρσ(ιος) δέσμ(η) α θος αρ( ) δέσμ( ) faint remains faint remains

σεση(μείωμαι) Κλαύδι(ος) υίός Κοράκι

Convex Βερρι Εύφρατ( ) [Άρ]ποκρ(ατίων) ξ(έσται) λβ Έρμησία ξ(έσται) λβ

Concave: "Pmoun Alexandros. Son of Stephanos, - bundles; Ammon son of Mersis, 1 bundle; , - bundles;. . . Claudius son of Korax signed." Convex: "Berri Euphrat( ): Harpokration, 32 sextarii; Hermesias, 32 sextarii."

102

OSTRAKA FROM TRIMITHIS

Concave 4 The first letter could be lambda or pi. Concave 6 Read Κόρακος. 51. Account of wheat and barley. Ca. 200-250. TM# 131099. Inv. 11723. Area 1.4, DSU 3, FN 2. 8.7 χ 8.2 cm; on convex and concave sides; broken at right. A2a. Convex ." ιθ Φιλομ( ) Άθυρ κθ (πύρου) (άρτάβας) β, κρ(ιθής) μ(άτια) β. Άδρ(ιανοΟ) ς(πυροΟ) (άρτάβας) ε [ μ( ) ιβ κρ(ιθής) μ(άτια) δ. ιέ (πυροΟ) (άρτάβας) y μά(τια) η. κ (πυροΟ) (άρτάβας) ε φ(ιθής) Φιλομ( ) κθ (πυροΟ) (άρτάβας) β. ΤΟβι ε (πυροΟ) (άρτάβας) β. vacat [ όμο(ίω5) μά(τια) γ

4

8

12

4

Πετο(σΐρις) yv( ) Άθύρ κε κρ(ιθής) (άρτάβας) β, ό(μοίως) (άρτάβην) α. Άδρ(ιανοΟ) γ [ θ (πυροΟ) (άρτάβας) ε. ιη (πύρου) (άρτάβας) ς· μά(τια) β. ΤΟβ(ι) y (πυροΟ) (άρτάβας) ε [ Ψενα(μοθνις) Άρπαή(σιος) Άδρ(ιανοΟ) ι (πύρου) (άρτάβας) δ μά(τιον) α. ι ς· (πυροΟ) (άρτάβας) [ (πύρου) (άρτάβας) δ. ΤΟβ(ι) ία κρ(ιθής) (άρτάβας) δ L vacat [ Πετο(σΐρις) Πε( ) Άδρ(ιανοΟ) ιβ (πύρου) (άρτάβας) ζ μά(τια) β κρ(ιθής) μά(τια) y [ Ky (πυροΟ) (άρτάβας) δ 'κρ(ιθής)' μά(τια) γ vacat [ ις (πύρου) (άρτάβας) ι μά(τια) δ [κρ(ιθής)] (άρτάβας) δ L [ πτ( ) (πύρου) (άρτάβας) iy L [ Concave Ά]θυ(ρ) κς" (πυροΟ) (άρτάβας) ς·. Άδρ(ιανοΟ) ιβ (πυροΟ) (άρτάβας) ς·. κς~ (πύρου) (άρτάβας) β κρ(ιθής) (άρτάβας) δ L ]. ιέ (πύρου) (άρτάβας) y. ΤϋΟρο(ς) Πετεχ(ώντος) (πύρου) (άρτάβας) δ κρ(ιθής) β L ]. κρ(ιθής) (άρτάβας) β L ] κρ(ιθής) μά(τια) δ όμ(οίως) ]. ] μά(τιον) α

Convex "... Philom( ): Hathyr 29, 2 art. of wheat, 2 mat. of barley. Hadrianos 6, 5 art. of wheat... 4 mat. of barley. (Hadrianos) 15, 3 art. of wheat, 8 mat. (Hadrianos) 20, 5 art. of wheat, ... of barley... (margin: Philom( )): (Hadrianos) 29, 2 art. of wheat. Tybi 5, 2 art. of wheat... likewise, 3 matia. Petosiris gn( ? ): Hathyr 25, 2 art. of barley, likewise 1 art. Hadrianos 3 ... [Hadrianos?] 9, 5 art. of wheat. (Hadrianos) 18, 6 art. of wheat, 2 mat. Tybi 3, 5 art. of wheat... "Psenamounis son of Harpaesis: Hadrianos 10, 4 art. of wheat, 1 mat. (Hadrianos) 16, ...

TEXTS

103

art. of wheat... 4 art. of wheat. Tybi 11, 4 Vï art. of barley. "Petosiris son of Pe( ). Hadrianos 12, 7 art. of wheat, 2 mat, 3 mat. of barley... (Hadrianos?) 23, 4 art. of wheat, 3 mat. of barley. (Month?) 16,10 art. of wheat, 4 mat., 4 % art. of barley... Pt( ): 13 l/2 art. of wheat." Concave "Hathyr 26, 6 art. of wheat. Hadrianos 12, 6 art. of wheat. (Hadrianos?) 26, 2 art. of wheat, 4 V^ art. of barley. [Month] 15, 3 art. of wheat. Horos son of Petechon, 4 art. of wheat, 2 Vï art. of barley... 2 V^ art. of barley... likewise 4 mat. of barley... 1 mat." Convex 1 and passim For Hadrianos, an honorific month replacing the Egyptian month Choiak in honor of Hadrian's regnal years, which began in December, see Kenneth Scott, YCS 2 (1931) 199-278, with discussion at 261-62. He collected attestations of the month over the span of 118/119 (?) to 244. Since then, no new attestation of the month can be dated after 244 with any certainty. The one possible exception, other than a handful of texts dated generally to III, is O.Kellis 1.48, which dates to a regnal year eleven taken by its editor as either 231/232 or 263/264. The former option is well within Scott's range of Hadrianos attestations. Our text here is likely therefore to date to before ca. 244 as well. If so, the Psenamounis son of Harpaesis attested in line 7 is probably not the same man as his homonym in 178 and 179, of a year 13 most likely to be 318/9. Convex 1, 3 (margin) We do not know what name Φιλομ( ) abbreviates. No name with this beginning has appeared in texts from Dakhla so far. Conceivably it is a misspelling of Φιλούμενος, well attested at Kellis, but we are reluctant to assume misspelling. Convex 4 ό(μοίως) presumably refers to the commodity, not the date, since multiple entries appear in the same date elsewhere in the text without such comment. So in the first instance, for example, "2 art. of wheat... likewise, 3 matia" of wheat is implied. Convex 4 yv( ) could represent a title, γνωστήρ, or a name as patronymic. Convex 12 πτ( ) could be the beginning of the common proper name Ptolemaios or the beginning of an occupational term like πτεριστής. The latter seems more likely. 52. Account? Ca. 350-370. TM# 131100. Inv. 11771. Area 2.1, Room 12, DSU 226, EN 162. 9.2 χ 8.8 cm; on convex side; complete but extremely faded. Alb. 12 lines; alpha can be read at the ends of lines 1, 2, 4, 9, and 10. A measure beginning mu appears to precede the numeral in some cases, perhaps μώ(ιον) or μ(άτιον). 53. Account of rushes. Ca. 350-370. TM# 131101. Inv. 11810. Area 2.1, Room 12, DSU 245, EN 176. 9.9 χ 7 cm; on convex side; complete. Al.

4

δ(ιά) Ψάιτ(ος) ΠεβώΟτος) δέμ(ατα) λ δ(ιά) Σίρις Ψάιτ(ος) σου δέμ(ατα) λ δ(ιά) ΣιΟγχις άττ(ελευθέρου) δέμ(ατα) ιέ δ(ιά) Σίρις υδροκ(αθαρτοθ) δέμ(ατα) ιέ α μέρος δέμ(ατα) ιέ δ(ιά) Ψεναμού(νιος) Άμμων(ίου)

104

8

OSTRAKA FROM TRIMITHIS Ήφαιστάμμ(ωνος) (γίνεται) σχοί(νων) δέμ(ατα) Ρε

[[δέμ(ατα) ιέ]]

"Through Psais son of Pebos, 30 bundles Through Siris son of your Psais, 30 bundles Through Siynchis, freedman, 15 bundles Through Siris the well-cleaner, 15 bundles 1st division 15 bundles Through Psenamounis son of Ammonios grandson of Hephaistammon [[15 bundles]] Total, 105 bundles of rushes." 2 What σου is intended to indicate here, we do not know. 3 απελεύθερος is normally abbreviated at least with the first four letters, but the twoletter abbreviation found here also occurs in O.Mich. 1.337.7. No other resolution appears more plausible to us; names beginning in Άπολ- are unlikely to have been abbreviated quite so drastically. The same form occurs in 106.3,155.3, and 220.1. 4 For a ύδροκαθαρτής as a "cleaner of the wells and the irrigation system" see O.Kellis 1.133 and RKellis 4.96.345-347n. The high number of wells attested in the Trimithis ostraka suggests considerable importance for such figures. 5 For the term meros or division, see above, p. 38. 7 It looks as if δέματα is washed out; it is not as clear that the numeral following it is, but possibly this was intended (see note to lines 8-10). It is conceivable, however, that the smear is accidental. 8 The first character in the line, which resembles an ornate psi, seems intended to represent (γίνεται). 9-10 The total is correct if the figure in line 7 was intended to be cancelled. If, on the other hand, the reference to the 1st division in line 5 is not to be included in the total because it is then explained in the following lines, then the total is correct including line 7. 54. Account in matia. Ca. 350-370. TM# 131102. Inv. 11816. Area 2.1, Room 12, DSU 226, FN 172. 6.4 χ 7.8 cm; on convex side; complete? Ala.

4

traces Κλαυδίου traces καθη [ ] traces Αρ[....] traces traces

μ(άτια) δ μ(άτια) ε μ(άτια) ε μ(άτια) ε μ(άτια)

TEXTS

105

55. Account of hay. Ca. 350-370. TM# 131103. Inv. 11820. Area 2.1, Room 13, FSU 150. 7.8 χ 7.8 cm; on concave side; complete but almost entirely faded. A2a. According to D. Ratzan, the ink was much clearer and brighter when found and was evidently an account. The ink essentially vanished within 24 hours except for line 1, which begins τω άδελφφ, followed by a name, and ends δέσ(μαι) ιέ. At the ends of lines 2-4 is also δέσ(μαι) followed by a number. 56. Account of barley and date stones. Ca. 350-370. TM# 131104. Inv. 11822. Area 2.1, Room 12, DSU 245, FN 204. 10.6 χ 5.4 cm; on convex side; broken left and below. Ala. ] κριθ(ής) μ(άτια) λγ ύπ(έρ) όστέω(ν) μ(ατίων) σξδ ]σ ύπ(έρ) ματ(ίων) σ ]." και αυτής ] ύττ(έρ) όστέω(ν) μ(ατίων) σξ ] Πμο(υν) Φιβ

4

". . . 33 matia of barley for 264 matia of date stones... for 200 matia... and the same ( ? ) . . . for 260 matia of date stones . . . Pmoun Phib." 1 A ratio of 8:1 in value for barley : date stones is thus given clearly. Barley prices in the KAB were problematic (pp. 52-53); date stones were valued in lines 1052 and 1663 at 100 talents per artaba, suggesting that barley would have been worth 800 talents per artaba. That is not far from the implied prices of 1,000 talents per artaba we find in some passages of that account. 4 Here stood another computation like that in line 1. The commodity at work may not have been barley as in line 1, however. 57. Account of hay. Ca. 350-370. TM# 131105. Inv. 11828. Area 2.1, Room 12, DSU 245, FN 198. 8.3 χ 4.8 cm; on convex side; broken at bottom. On concave side, faint traces of 5 lines. Alb. The surface slip is lost through much of the area.

4

faint traces δι(ά) Ψάιτ(ος) traces δι(ά) Ψάιτ(ος) Πεκύσι[ος δι(ά)

[

]δέσ(μαι)δ

58. Account of hay. Ca. 350-370. TM# 131106. Inv. 11829. Area 2.1, Room 12, DSU 245, FN 192. 6.2 χ 6.4 cm; on convex side; complete. Alb.

106

4

OSTRAKA FROM TRIMITHIS ζ ΠατερμοΟΘ(ις) χόρτ(ου) δέσμ(αι) ς ος Σαραπίω(νος) δέσμ(αι) y Άμμων Βερενίκ(ης) δέσμ(αι) y A traces δέσμ(αι) β traces δέσμ(αι). th

(1-4) "7 . Patermouthis, 6 bundles of hay. . . . son of Sarapion, 3 bundles. Ammon son of Berenike, 3 bundles. A- - -, 2 bundles." 59. Account of hay. Ca. 350-370. TM# 131107. Inv. 11833. Area 2.1, Room 12, DSU 245, FN 180. 8.7 χ 6.6 cm; on convex side; broken at upper left. A2a.

4

8

ώ]ν έ'σχον [y. . ]βουλος /Λουιερμαμουν /Ψάι 5 Πεβώτ(θ5) /Χάρις Πατερμούτ(ιος) /Σαρμάτης Λουισάιτ(ος) Πολλουν Ψάιτ(ος) Ψάις Σεναρσιήσ(ιος)

δέσμ(αι) ν δέσμ(αι) ν δέσμ(αι) ν δέσμ(αι) ν δέσμ(αι) ν δέσμ(αι) ν δέσμ(αι) κε

"From which I received: [ Jboulos, 50 bundles. Louiermamoun, 50 bundles. Psais son of Pebos, 50 bundles. Charis daughter of Patermoutis, 50 bundles. Sarmates son of Louisais, 50 bundles. Polloun son of Psais, 50 bundles. Psais son of Señarsiesis, 25 bundles." 2 Most likely [Εϋ]βουλο5. 3 The contrast between the form here, in which an Egyptian version of Έρμάμμων is joined without declension to Λουι-, and the formation of Λουισάις with a Greek termination in line 6 is striking. 7 One expects Πολλών, which is attested at Kellis, but that is not easily read here. 60. Account of oil (?). Ca. 350-370. TM# 131108. Inv. 11932. Area 2.1, Room 15, DSU 157. 5.9 χ 5.7 cm; on convex side; broken at left. Ala. The remains of an account, trimmed to become a lid set into a jar stopper.

Iß

TEXTS

107

]