This work is available at the URI http://dlib.nyu.edu/awdl/isaw/amheida-i-otrim-1 as part of the NYU Library's Ancient World Digital Library and in partnership with the Institute for the Study of the Ancient World (ISAW). Further information about ISAW's publication program is available on the ISAW website. Please note that while the base URI of this document is stable, the exact content available is still undergoing development. Additionally, the Greek and Demotic characters will look best with the "New Athena Unicode" and "Umschrift_TTn" fonts installed.
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 B1, 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 (B1), 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.
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 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 B1, 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 well-known 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 B1 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 B1 (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.
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 B1, 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 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 (ἔτους), ‘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 (ἥμισυ). 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: LιεS/. 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 ἥμισυ, half. It is not probable that our tags record only amounts ending in ½. 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 ὕδ(ρευμα) Πμο(υν) found in many of the tags. “Yr 1st/last” indicates whether the year appears at the start of the ostrakon’s text or at its end.
Year | Ostrakon | Room | Strat. Unit | Pmoun or Hyd. Pm.? | Yr first/ last | Year symbol |
1 | 244 | 12 | 245 | X | first | αS/ |
1 | 235 | 12 | 245 | X | first | αS/ |
2 | 146 | 15 | 157 | Hyd. Pm. | last | Lβ |
2 | 147 | 4 | 214 | Hyd. Pm. | last | βL |
3 | 100 | 1 | 53 | Pm. | last | γL |
3? | 213 | 9 | 212 | Pm.? | last | γS |
4 | 102 | 8 | F74 | Pm. | last | δL |
4 | 103 | 9 | 129 | Pm. | first | δS |
4 | 32 | 15 | 157 | X | first | δS/ |
5 | 104 | 4 | 29 | Pm. | last | εS |
5 | 270 | 14 | 211 | X | last | εS |
6 | 243 | 12 | 245 | X | first | ϛS// |
7 | 241 | 10 | 191 | X | first | ζS |
7 | 196 | 8 | 102 | X | first | Lζ |
7 | 175 | 4 | 29 | X | first | Lζ |
7 | 202 | 10 | 142 | X | last | ζS/ |
7 | 106 | 9 | 132 | Pm. | first | ζS/ |
7 | 142 | 4 | 214 | X | first | Lζ |
7 | 357 | 3 | 22 | ? | first | ζS |
9 | 148 | 3 | 93 | Hyd. Pm. | last | θS |
9 | 108 | 8 | 102 | Pm. | last | θL |
10, 11 | 149 | 4 | 35D+G | Hyd. Pm. | both | ιL, ιαL |
11 | 110 | 9 | 132 | Pm. | first | ιαS |
11 | 111 | 10 | 142 | Pm. | first | ιαS/ |
13 | 178 | 4 | 35A | X | last | ιγL |
13 | 179 | 4 | 35A | X | last | ιγL |
13 | 150 | 9 | 167 | Hyd. Pm. | last | ιγL |
13 | 223 | 9 | 200 | X | last | ιγL |
14 | 151 | 9 | 151 | Hyd. Pm. | last | ιδL |
14? | 152 | 9 | 127 | Hyd. Pm. | last | ιδL |
14? | 154 | 14 | 194 | Hyd. Pm. | last | ιδL? |
14 | 96 | 9 | 218 | Pm. | first | Lιδ |
14 | 112 | 9 | 200 | Pm. | last | ιδL |
14 | 233 | 11 | 235 | X | first | Sιδ/ |
15 | 115 | 8 | 98 | Pm. | last | ιεS |
15 | 113 | 4 | 35C | Pm. | first | Lιε |
15 | 114 | 3 | 36, F38 | Pm. | first | LιεS/ |
15 | 116 | 9 | 212 | Pm. | first | Lιε |
15 | 215 | 9 | 212 | Pm. | first | ιεL |
15 | 242 | 12 | 245 | X | first | ιεS/ |
16 | 117 | 8 | 102 | Pm. | first | Lιϛ |
16 | 183 | 4 | 35D+G | X | first | Lιϛ |
16 | 118 | 1 | 168 | Pm. | last | ιϛ[ |
16 | 119 | 1 | 168 | Pm. | last | ιϛL |
16 | 121 | 9 | 218 | Pm. | first | Lιϛ |
16 | 122 | 4 | 214 | Pm. | first | Lιϛ |
17 | 123 | 4 | 214 | Pm. | last | ιζL |
20 | 155 | 9 | 104 | Hyd. Pm. | last | κL |
20 | 156 | 9 | 104 | Hyd. Pm. | last | κL |
20 | 125 | 9 | 107 | Pm. | last | κL |
20 | 124 | 4 | 23 | Pm. | first? | κL |
20 | 157 | 9 | 128 | Hyd. Pm. | last | κL |
20 | 158 | 10 | 191 | Hyd. Pm. | last | κL |
21 | 160 | 11 | 79 | Hyd. Pm. | last | καL |
21 | 159 | 4 | 9 | Hyd. Pm. | last | καL |
22 | 161 | 4 | 12 | Hyd. Pm. | last | κβL |
22 | 228 | 10 | 220 | X | last | Lκβ |
23 | 162 | 4 | 9 | Hyd. Pm. | last | κγL |
23 | 194 | 3 | 95 | X | first | κγS/ |
25 | 163 | 1 | 4 | Hyd. Pm. | last | κεL |
25 | 164 | 1 | 4 | Hyd. Pm. | last | κεL |
25 | 231 | 9 | 232 | X | first | κεL |
27 | 165 | 8 | 63 | Hyd. Pm. | last | κζL |
28 | 166 | 1 | 4 | Hyd. Pm. | last | κηL |
30 | 126? | 4 | 35 | Pm. | last | λ[ (uncertain if year) |
30? | 214 | 13 | 207 | X | first | λ̣S |
32 | 167 | 9 | 151 | Hyd. Pm. | last | Lλβ |
In order to have the material from the other areas available for comparative purposes, we provide a table also with these:
Year | Ostrakon | Area | Strat. Unit | Pmoun or Hyd. Pm.? | Yr first/ last | Year symbol |
1 | 200 | 1.3 | 19 | X | first | αS |
1 | 203 | 4.1 | 64 | X | first | αS/ |
2 | 97 | 4.1 | 61 | Pm. | first | β// |
2 | 98 | 4.1 | 77 | Pm. | first | β// |
2 | 99 | 1.4 | 3 | Pm. ? | last | Lβ |
4 | 101 | 1.1 | 4 | Pm. | last | δL |
5 | 105 | 4.1 | 61 | Pm. | first | εS/ |
8 | 107 | 4.1 | 43 | Pm. | first | ηL |
10 | 109 | 4.1 | 99 | Pm. | first | ιS/ |
10 | 195 | 4.1 | 11 | X | first | ιS/ |
10 | 199 | 4.1 | 43 | X | first | ιS// |
14? | 153 | 4.2 | 13 | Hyd. Pm. ?? | first | ιδS/ |
30 | 230 | 1.4 | 1 | X | first | λS |
33 | 127 | 4.1 | 18 | Pm. | first | λγS |
We will first look at these data from the point of view of formal criteria and patterns, then passing on to a stratigraphic 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 1 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) | Possible senior emperor(s) |
1 | Diocletian (or earlier) |
2–9 | Diocletian; Valentinian I |
10–13 | Diocletian; Constantine I; Valentinian I |
14 | Diocletian; Constantine I |
15–16 | Diocletian; Galerius; Constantine I |
17–21 | Diocletian; Galerius; Constantine I; Constantius II |
22–23 | Galerius; Constantine I; Constantine II; Constantius II |
24 | Constantine I; Constantine II; Constantius II |
25–31 | Constantine I; Constantius II |
32– | 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 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 mid-century, but each stratigraphic 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 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 stratigraphic 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 ἰνδικτίων. (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 | L sign | S or / sign following |
1–5 | 6 | 12 |
6–10 | 6 | 9 |
11–15 | 14 | 5 |
16–20 | 12 | 0 |
21–25 | 9 | 1 |
26–30 | 2 | 1 (+ 1?) |
31+ | 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 stratigraphic discussion below.
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 stratigraphic 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 B1, including the many stratigraphic 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.
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 quem 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 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 2-meter 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 B1. None of the ostraka from Room 2, however, contains a regnal year.
The other stratigraphic 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 B1. Unfortunately, these were mostly poorly preserved and uninformative. Only 177 is perhaps of interest, as one 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 4,9 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.
DSU 214 is a foundation fill, thus consisting entirely of material from before the construction of House B1. 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: Selles, 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 pre-construction 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.
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 B1.
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.
With Room 11 we begin the series of rooms on the west side of House B1. It is the room painted with a green decorative pattern. It produced just 8 ostraka. DSU 79 (sand filling 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 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 (ιεS// ἰνδ(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.
This room is the northernmost of the series of rooms on the west side of House B1. 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 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 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 quem.
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.
The courtyards numbered 9 and 10, and Rooms 15 and 16, belong to a structure that was originally not part of House B1. The bulk of this area was occupied by a multi-room 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.
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.
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 (presbyteros) 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.
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 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 B1. 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 quem 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.)
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 Mersis in 241 (dating to year 7) may be the same man appearing in P.Kellis (see below, link) 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). 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.
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 quem. 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.
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. 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 ὑδροκαθαρτής 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 topography.14
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 ϣωι, “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 B1 of Area 2.1, the temple hill, and Area 1.3. The last of these (21) is an account including payment to a donkey-driver, perhaps for transportation of goods to and from the well. Pmoun Tpake appears three times in ostraka from House B1. 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 1 G. 35 as “of the city of the Mothites” (τῆς Μωθειτῶν| [πόλεως]), that is, modern Mut, may help place the well near there, but equally may refer simply to the Oasis in general.
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 double-sided tag (155) from a year 10, the reverse of which mentions a Pmoun Ὀσιρε. 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 Ἐμβ(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 an ostrakon from Deir el-Hagar, due west of Trimithis.17 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 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 Pmoun Beri appears in P.Kellis 1 G 5.12, and may refer to the same locale.18 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) | α̣ὐτουργ( ), “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 Mersis.” 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, link).
The following toponyms appear in ostraka from Trimithis:
. . . .θ̣αν̣ (…than) – appears in 14, a memorandum or account of oil.
Α[ (A[) – appears in well tag 99.
Ἀλεξάνδρο̣(υ) (Alexandros) – appears in an account of hay, 50.
Ἀμε̣[ (Ame[) – appears in well tag 118.
Ἁραυ (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 (Galmen) – appears in one well tag from Area 4.1, 97.
χω̣ρίο(υ) Ψενα̣μ̣(ούνιος) (Chorion of Psenamounis) – appears in an account, 10, described only as hydreuma, not as a pmoun.
Ε[ (E[) – appears in 27, an account of people from Pmoun Horou, one of whom, Makarios, is from E[.
Ἐμβωο(υ) / Ἐνβω̣(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.
Ἑ̣ρμ̣( ) (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?
Ἑρμοῦ (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?
Ἑρμο.( ) (Hermo( )) – appears in well tag 159 as a hydreuma pmoun. Perhaps to be resolved as Hermod(orou). Identical to preceding Pmouns?
Ἠσε (Ese) – appears in a list, 77, with Pmoun Po.itou.
Θα̣τ( ) (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.
Ἰβυθο( ) (Ibutho( )) – appears in well tag 162 as a hydreuma pmoun.
Μαρκ( ) (Mark( )) – appears in an account of days, 39, with Pmouns Phib and Psoin.
ὀνηλ( ) – appears in 22.3: presumably “of the donkey-driver(s).”
Ὀ̣σειρε/ Ὀσιρε (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).
Ὀστρακ( ) (Ostrak( )) – appears in well tag 152 as a hydreuma pmoun.
Π. .πο̣. . .( )– 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|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).
Π̣α̣τ̣ώ̣ο̣υ̣(ς) (Patoos) – appears in well tag 131.
Πιαμ (Piam) – appears in well tag 112 associated with a so-called self-operated chôrion.
Πισῆχθις (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.ιτο̣υ (Po.itou) – appears in a list with Pmoun Ese, 77.
Σαραπί(ωνος) (Sarapion) – appears in well tag 160 as a hydreuma pmoun.
Τα[ (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.
Τα̣οσιρα (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.
Τσυρου (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 ϩαλμε, 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 ‘P’; cf. 143.
Ψωιν (Psoin) – appears in 39, an account of days with Pmouns Phib and Mark( ? ).
Ὡρ( ) (Hor()) – appears in well tag 111 naming the vineyard of Petosiris. Identical with subsequent Pmoun?
Ὥρου̣ (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.
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 ὕδρ(ευμα) τὸ Ἀμμώνιον 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).
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 marion 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.
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 B1 was connected was the House (οἰκία), which we see mainly in the role of receiving goods. It is referred to as the house of Mr. Serenos (ὁ κύριος Σερῆνος) 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” (γεοῦχος) 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 οἰκοδέσποινα 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 cella (κέλλα), attested in 257. They addressed one another as “colleague”, literally “brother” (ἀδελφός), although this term was used so broadly in correspondence that we cannot be confident that all of those using it were in fact colleagues at the same level.22 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 managers.23 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 ὥστε τῷ κυρίῳ μου τῷ στατιωναρίῳ 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 meros 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 “sixth part” (l. 10: μέρος ἕκτον) of land under joint ownership (ll. 10–11: κατ]ὰ 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 meros. The account in 53 lists α̣ 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 τοῖς παιδίο̣ις, “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 τῆς παιδίσκ[ης, “the slave-girl.” 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 other Kellis texts only in reference to children.26 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 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 link). 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 | Domnion | Nikokles | |||
Room (DSU) | O.Trim. | Room (DSU) | O.Trim. | Room (DSU) | O.Trim. |
2 (13) | 297 | 4 (23) | 279 | 4 (23) | 279 |
2 (13) | 298 | 4 (15) | 287 | 10 (220) | 226 |
2 (13) | 299 | 4 (23) | 253 | 13 (208) | 288 |
4 (23) | 300 | 14 (221) | 66 | 15 (backfill) | 291 |
5 (213) | 271? | 14 (233) | 292 | 16 (161) | 313 |
10 (119) | 268 | 14 (233) | 293 | 16 (161) | 282 |
17 (153) | 310 | 14 (233) | 324 | 16 (158) | 312 |
17 (155) | 311 | 14 (233) | 322 | ||
14 (233) | 323 | ||||
14 (233) | 69 |
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 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.
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 marion of 20 sextarii (μάριον), and the chous of 6 sextarii (χοῦς). 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 (link).
For wheat and barley, the modius, at 10 modii to 3 artabas (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 τοπικῷ (μέτρῳ), “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.
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 B1 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, BASP 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, link), and chickens.
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 1G. 27.31 That text is a letter addressed to Serenos in his capacity as praepositus 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 and Coptic documents from Kellis.32 (Some unpublished evidence will be discussed 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 praepositus 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 century.33 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]ω̣θιτῶν πόλεως τῆς μεγάλης Ὀάσ[εως | ]ς τοῦ βουλευτηρίου ψηφίσμασι οὕτως . [.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 περὶ τῆς λειτ̣ο̣υ̣ργία̣ς 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 ἔδοξε, as evidence of this voting practice from the early third century.38 According to Bowman, the γραμματεύς of the boule was “responsible for the publication of the ψήφισμα.”39 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 strategos 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 exaktor’s 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,” τηρητὴς 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 τηρηταὶ κώμης καὶ καρπῶν (“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 “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 (ἀνθρώπους ἐν φυγῇ) 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 [τοὺς 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 Trimithis was a polis by 304.45 Cities, in our experience elsewhere, do not have praepositi, they have a civic government and a logistes.
And yet it must be said that so far we have no evidence of a logistes 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 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]ὐρηλίοις Κλεοϐο[ύλ]ῳ λογιστῇ Ὀάσεως̣
[Με]γάλης καὶ Φιλοσ̣α̣ρά[πι]δι τῷ καὶ Μικκάλῳ̣
[ἄρ]ξαντι προέδρῳ Μω[θι]τῶν πόλεως κα̣ὶ̣
[Ἀν]δρομάχῳ Ἀπόλλ̣ω̣ν[ο]ς ἄρξ(αντι) συνδίκ̣ο[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 praepositus. 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, 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.
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 (εἴλη) and to a decurion (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.
As well as House B1, 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 B1.49)
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 Iona, 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 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 B1 and the Area 4 temple site. Pmoun Osire appears in an ostrakon found in House B1 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 respectively.51 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 (B1) 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 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.
These ostraka from Area 1 attest perhaps as many as 57 different people.52 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 people have names of this variety.53 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 Egyptian deities and more standard Greek names (Korax,54 Syros). The demotic texts include a proper name with the element Ws|r, 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 ἔσχον ἀπὸ Ἀκύλα, “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” (ὑπ(ὲρ) 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:
Σαμως Πατχωτ( ) κε(ράμια) ιη καὶ ὑπ(ὲρ) προσ-
θήκ(ης) κε(ράμιον) α, κ̣α̣ὶ̣ ὑπ(ὲρ) προσθ(ήκης) κερ(άμιον) α
4 Βῆκι̣ς̣ Ψάιτ(ος) ἐλ(αίου) χό(ες) λβκαὶὑπ(ὲρ) προσθήκ(ης)
χό(ες) λβ.
In the first entry, Samos makes two payments hyper prosthêkês amounting to one-ninth 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 Aurelia Charite 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 artabas of grain, precisely one-third an initial payment of 225 artaba 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 ὑπ(ὲρ) 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 1/9th, one whole, 1/5th, 1/10th 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, 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-to-day 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.
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 of Psais the deacon (known from two other ostraka of that period) in 383 and the co-occurrence of Nikokles and Philippos in 286. Both of these are well known from the final phase of occupation of House B1.
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 quantity.59
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 included.(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.
Amoun/Ammon
Names formed from the stem Ἀμμων-, representing the Libyan version of Amoun, are very common. +Ἀμμώνιος dominates; there are smaller numbers of Ἀμμωνιανός, +Ἄμμων (not always easy to distinguish from Ammonios, given truncation and abbreviation), and *Ἀμώνιος. All of these are fully Hellenized in form. Cf. also under Hephaistos for Ἡφαιστάμμων, Hermes for Ἑρμάμμων, 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, *Παμοῦνις, Σεναμοῦνις, *Ταμουν, and +Ψεναμοῦνις. 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 *Παχοῦμις occurs; it is found moderately often at Kellis.
Anoubis
Only *Ἀνουβ and *Πανουβᾶς are found formed on the divine name. *Πουῶρις (“the dog”, NBDem. 181) may also refer to Anoubis.
Apis
The only names certainly formed on Apis are the Hellenized *Ἀπίων and the Egyptian *Παᾶπις. It is absent from the Kellis material. *Πᾶφις could be Pa-ḥp (NBDem. 400), “the one of Apis,” but pa-ḥ῾pj, “the one of the Nile” (NBDem. 399) and P3-ḥf, “the snake” (NBDem. 204) are also possible.
Atum
Only *Ψενετῦμις occurs.
Bes
The only names formed on Bes are *Βησᾶς and *Βῆσις. 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 Πετεχῶν, “the gift of Chonsu.” It is rare at Kellis.
Herakles
Chonsu’s Greek equivalent is represented by his name prefixed with the Egyptian article, Φηρακλῆς, found in three texts. Other forms are the article-less Ἡρακλῆς, *Ἡρακλᾶς, *Ἡράκλειος, and *Ἡρακλείδης, 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 *Ἡρακλάμμων is also to be noted. The situation at Kellis is similar.
Harpochrates
Ἁρποκρατίων occurs twice.
Horus
Horus is very well represented in the names, mostly with the simple divine name +Ὧρος, but also with the compounds Ἁρπαῆσις (“Horus the one of Isis”), *Ἁρσιῆσις (“Horus the son of Isis”), *Ἁρμάις (divinized Horemheb), Ἁρυώτης (“Horus is well”), and Πετεῦρις (“gift of Horus”), along with probably feminine derivatives of these, *Σεναρσιῆσις and *Σεναρυώτης. In addition, there are the Greek derivatives *Ὡριγένης and Ὡρίων, and apparently *Ὡριαίνα in an aberrant spelling. *Παχράτης (“the one of the child”, NBDem. 411) is probably also to be referred to Horus. Βῆκις (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 (fem.) belonging to Isis,” *Θαῆσις and Ταῆσις (only in a well name). The masculine Παῆσις (“the one (masc.) belonging to Isis”) appears four times. Commoner is Ψεννῆσις (8 examples, one with single nu, compared to 5 at Kellis), “the son of Isis”. *Πετῆσις and the undeclined *Ψενταησα appear once each. There are also two instances of the Greek formation Ἰσιδώρα, “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 Παμ-names 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 (masc.) 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 Νεῖλος/Νῖλος 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 Πᾶφις.
Osiris
+Πετοσῖρις is moderately common, as at Kellis; we also find *Ψενοσῖρις. Cf. also the epithet *Ὀννῶφρις (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.
Renenutet
Πατερμοῦθις/Πατερμοῦτις 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 Σαραπᾶς. The divine name *Σερᾶπις is found once.
Shaï
The Greek equivalent of Shaï, *Ἄγαθος 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. *Παψάις 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 *Παῆβις. The incompletely read *Παν. .νῖβις may also be an ibis name.
Hermes
Thoth’s Greek counterpart appears in several forms: Ἑρμῆς (twice) and Ἑρμησίας (4 times) and the compound Ἐρμάμμων, also 4 times. *Ἑρμίας, *Ἑρμόδωρος, and *Ἑρμοκλῆς 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 +Τιθοῆς, 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.
Apollo is represented by *Ἀπόλλων, Ἀπολλώνιος, and Ἀπολλῶς. The picture at Kellis is similar. Cf. also section 5 below on Πολλουν.
Ares: *Ἄρης
Demeter: *Δη̣μη̣τρία.
Eros: Ἔρως.
Hephaistos: the compound Ἡφαιστάμμων.
Heron: Ἥρων appears three times (once at Kellis).
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; *Ἀλέξανδρος 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: *Ἀγησίλαος, *Ἀχιλλεύς, *Εὔμηλος, Νικοκλῆς (a prominent signer), and *Πολύβιος. For the rest, most of the Greek (or 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): *Ἀθηνόδωρος, *Ἄνητος (for Ἄνυτος?), Ἀντίνοος (4, some perhaps referring to Antinoopolis), Ἀσκλᾶς (2), Γελάσιος (4), *Γέρων, *Δίδυμος, *Διόσκορος, *Διοφάνης, *Δοῦλος, *Εὐτύχιος, *Ἐπαφρόδιτος, *Ἐπίμαχος, *Ζηνόβιος, *Θεόδωρος, *Ἰον (= Ἴων?), *Ἰώνιος (?), *Κέρδων, Κόραξ (cf. Πεβωκ, section 5 below), *Κύλλος, *Λυκαρίων, *Μάρων, *Μέμνων, *Μύρων, Νικάων, Νίνος, *Ὀλύμπιος, *Παΐων, Πλουτογένης (2), Σαρμάτης (3), *Σκύλαξ, *Στέφανος, Συρίων (2), Σύρος (5), *Τρόφιμος, *Τύραννος, *Ὑγιεῖνος, *Φιλίσκος, *Φιλομ( ), and *Χάρις. That some of these may be calques of Egyptian names or have Egyptian reference is clear.
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: *Αἰλιανός, *Ἀκύλας, *Ἀμάτιος, *Ἀντώνιος, *Βίκτωρ, *Γαίων, +Δομνίων, *Ἰουλιανός, Καπίτων, *Κλαυδία, Κλαύδιος, Λογγῖνος, Μακρίνα, *Μαξιμι.( ), *Ὀπτᾶτος, *Οὐαλέριος, *Οὐαλου( ), Περπέριος, +Σερῆνος, *Σιλβανός, *Φαυστιανός.
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.
*Ἀκῆς
*Ἀκῦσις (= Ἐκῦσις, “Nubian”, NBDem. 80)
*Βανίπις (“iron”, cf. Hasitzka 19, Crum CD 41a)
Βελλῆς (“blind”, NBDem. 143)
*Καλε (prob. a variant spelling from Gl#, “lame” [NBDem. 1034])
*Καροῦρις (Ärwr, “frog”, NBDem. 982)
*Κολαβῖνις (“axe”, Crum, CD 102b)
Μέρσις (“red”, NBDem. 602)
*Νουβης (= Νοβς, cf. KAB 68 n. 36: jujube)
*Οὐαφ. .[ (perhaps Οὐαφρῆς, W#H-|b-ro, NBDem. 113)
*Παβῶς (variant of Πεβῶς?)
Παλεβ
Παμοῦρις (the one of mour)
Παναμεύς
Πανῦρις (cf. Pinuris, “the dog of Horus”, P#-|w|w-n-Hr, NBDem. 156—Vittmann)
*Παοῦς (prob. Pa-Hr, “the one of the face”, NBDem.401)62
*Παστωοῦς/Πεστωοῦς/Παστῶς (Pa-sT#.ß=w and P#j-sT#.ß=w, cf. NBDem., fasc. 18, p. 180 ad pp. 559 and 567: “He whom they have redeemed”, cf. Enchoria24 [1997/98]92-93—Vittmann)
*Πατχωτ( )
*Παχάλμιος (“the one of the spring; see 109.2–3n.)
*Πεβωκ (perhaps *P#-obq, “the raven”, cf. NBDem. 96, Ἀβωκ)
Πεβῶς (P#-#b#, NBDem. 154)
Πεκῦσις (P#-|gS, “the Nubian”, NBDem. 160, cf. Ἀκῦσις)
*Πιαῦς
Πικ[ (Πικῶς?, P#j-k#, NBDem. 442–43)
Πινᾶχθις, -ης (from nXß stem, “strength”, perhaps with reference to Amon-Nakht?)
*Πλῆνις (P#-ljn, “the smith”, NBDem. 199)
*Πολλουν (or form of Πολλων?)63
Πρεμεμοῦρις (perhaps *P#-rmT-#mwr, “the man of Crocodilopolis” [Gebelein]— Vittmann)
*Πσιρις
Σαμουν
Σαμῶς (V#j-n.|m=w, NBDem. 1348)
Σελλῆς (Σελλαευς)
Σῖρις
*Σιῦγχις (possibly *c#-onX, *cj-onX “living son”, cf. Or-onX "living face", NBDem 793— Vittmann).
*Ταουτ
*Τερητ( )
Χράτης ($rß, “child”, NBDem. 890)
*Ψενελυταλ( )
*Ψενθηθ
*Ψενταλε( )
*Ψιράις (cf. above, Rait)
Ψλύις
Μουσῆς (Moses) is the most common of these, with 4 instances; we also find once each *Ἐφρεμ, *Ἰωνᾶς (Jonah), and *Ἰωσηφ.
These are relatively uncommon. We find Μακάριος twice, *Μαρτύριος, Παῦλος (4 times), Τιμόθεος 4 times, and Ψενπνούθης (“the son of God”) 6 times.
Σαμβᾶς (twice) and the diminutive Σαμβάτιον.
*Ῥοιμητάλκας
*Κύρος (found in the Old Testament)
Γενα
Θατ
*Κελε
*Κοραυ
Λουι
*Λουια
*Λουιάμμων
*Λουιατο[
*Λουιερμαμουν
*Λουισαι
*Λουισάις
*Λουισεν
*Λουισῖρις
*Λωινο[υτε?]
*Μῶνις
*Παγενας
Παρα
Πισῆχθις65
Του
For Loui- names, cf. KAB, p. 68 n. 33; P.Kellis 5, 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. Iena; 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 Ἀμμώνιος (extremely common in the oasis), Παρα 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.
*Εἰραμου
*Ἐράβιος (perhaps = Ἀράβιος, but that is a doubtful name itself)
*Λομα
Νοθη
Οὐαει/Οὐαϊ (also O.Kellis 114.conv. 2; Egyptian wHj, “oasite”? Or Greek οὐαί, “woe”?) and Πουαει.
*Σαπίων (variant of Σαραπίων?)
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 except 4.1 and 4.2. All contexts in those areas were thoroughly disturbed and without stratigraphic value. Room 3 belonged to the house to the south of B1. In the list of rooms, 0 indicates a surface find.
Dates are discussed in the discussion of stratigraphy above, link, 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 | Date Attributed | O.Trim. |
1.1 | 0 | 0 | Surface find | 4 c. | 1 |
1.1 | 0 | 4 | Surface of unexcavated area | 287/8? | 101 |
1.2 | 0 | 3 | Mud brick collapse in street from north wall of House B2 | 285/6? | 99 |
1.2 | 0 | 5 | Secondary trash deposit in street | ca 240–275 | 36 |
1.2 | 0 | 6 | Trash compacted into street surface | ca 240–350 | 141 |
1.3 | 1 | 12 | Clean wind-blown sand | ca 240–350 | 21 |
1.3 | 1 | 16 | Occupational debris above floor | ca 300–350 | 20 |
1.3 | 1 | 16 | Occupational debris above floor | ca 300–350 | 23 |
1.3 | 1 | 16 | Occupational debris above floor | ca 300–350 | 388 |
1.3 | 2 | 7 | Occupational debris at floor level | Ptolemaic-Early Rom. | 390 |
1.3 | 2 | 15 | Room fill | ca 300–350 | 22 |
1.3 | 2 | 15 | Room fill | Ptolemaic-Early Rom. | 389 |
1.3 | 2 | 20 | Occupational debris? Or foundation fill? | Ptolemaic-Early Rom. | 391 |
1.3 | 2 | 20 | Occupational debris? Or foundation fill? | Ptolemaic-Early Rom. | 392 |
1.3 | 2 | 20 | Occupational debris? Or foundation fill? | Ptolemaic-Early Rom. | 393 |
1.3 | 3 | 19 | Occupational debris above floor | 284/5? | 200 |
1.3 | 5 | 40 | Wind-blown sand below surface | Unknown | 395 |
1.3 | 7 | 48 | Wall and ceiling collapse | ca 300–350 | 24 |
1.3 | 9 | 73 | Top floor level, compacted mud | ca 315–350 | 26 |
1.4 | 0 | 1 | Wind-blown sand fill | 335/6 or 353/4 | 230 |
1.4 | 0 | 2 | Room fill, occupation debris | Unknown | 445 |
1.4 | 0 | 3 | Room fill, occupation debris | ca 200–250 | 51 |
2 | 0 | 1 (near 2.1) | Surface find | ca 350–370 | 294 |
2.1 | 0 (W alley) | 88 | Wind-blown sand covering collapse | ca 350–370 | 191 |
2.1 | 0 (W alley) | 88 | Wind-blown sand covering collapse | ca 350–370 | 192 |
2.1 | 0 (W alley) | 190 | Wind-blown sand covering collapse | ca 350–370 | 316 |
2.1 | 14 / 18 | 177 | Wind-blown sand just below surface | 4 c. | 259 |
2.1 | 1 | 1 | Wind-blown sand, top fill | 4 c. | 71 |
2.1 | 1 | 1 | Wind-blown sand, top fill | 4 c. | 340 |
2.1 | 1 | 4 | Dome collapse and occupation | 348/9 | 163 |
2.1 | 1 | 4 | Dome collapse and occupation | 348/9 | 164 |
2.1 | 1 | 4 | Dome collapse and occupation | 351/2 | 166 |
2.1 | 1 | 4 | Dome collapse and occupation | ca 350–370 | 174 |
2.1 | 1 | 4 | Dome collapse and occupation | ca 350–370 | 295 |
2.1 | 1 | 53 | Wall collapse and occupation debris | 286/7? | 100 |
2.1 | 1 | 168 | Floor preparation layer | 321/2? | 118 |
2.1 | 1 | 168 | Floor preparation layer | 321/2? | 119 |
2.1 | 2 | 13 | Deposit on floor, sealed by sand | ca 350–370 | 4 |
2.1 | 2 | 13 | Deposit on floor, sealed by sand | ca 350–370 | 5 |
2.1 | 2 | 13 | Deposit on floor, sealed by sand | ca 350–370 | 252 |
2.1 | 2 | 13 | Deposit on floor, sealed by sand | ca 350–370 | 297 |
2.1 | 2 | 13 | Deposit on floor, sealed by sand | ca 350–370 | 298 |
2.1 | 2 | 13 | Deposit on floor, sealed by sand | ca 350–370 | 299 |
2.1 | 2 | 13 | Deposit on floor, sealed by sand | ca 350–370 | 356 |
2.1 | 2 | 13 | Deposit on floor, sealed by sand | ca 350–370 | 358 |
2.1 | 2 | 13 | Deposit on floor, sealed by sand | ca 350–370 | 359 |
2.1 | 2 | 13 | Deposit on floor, sealed by sand | ca 350–370 | 360 |
2.1 | 2 | 14 | Fill of niche in wall F3 | ca 200–270 | 2 |
2.1 | 2 | 32 | Final floor level in room | ca 350–370 | 366 |
2.1 | 2 | 34 | Second floor level from top | ca 335–350 | 176 |
2.1 | 2 | 34 | Second floor level from top | ca 335–350 | 177 |
2.1 | 2 | 34 | Second floor level from top | ca 335–350 | 303 |
2.1 | 2 | 34 | Second floor level from top | ca 335–350 | 345 |
2.1 | 2 | 34 | Second floor level from top | ca 335–350 | 367 |
2.1 | 2 | 34 | Second floor level from top | ca 335–350 | 368 |
2.1 | 2 | 34 | Second floor level from top | ca 335–350 | 369 |
2.1 | 2 | 38 | Part of sequence of occupation levels | ca 335–350 | 180 |
2.1 | 2 | 40 | Compacted layer below DSU 38 | ca 335–350 | 189 |
2.1 | 2 | 40 | Compacted layer below DSU 38 | ca 335–350 | 371 |
2.1 | 2 | 41 | Refuse layer below floor DSU 40 | ca 335–350 | 132 |
2.1 | 2 | 41 | Refuse layer below floor DSU 40 | ca 335–350 | 188 |
2.1 | 2 | 41 | Refuse layer below floor DSU 40 | ca 335–350 | 376 |
2.1 | 3 | 22 | Sand and debris above mudbrick collapse | 290/1? or 318/9 or later | 357 |
2.1 | 3 | 33 | Debris deposit on top of last floor | ca 350–370 | 10 |
2.1 | 3 | 33 | Debris deposit on top of last floor | ca 350–370 | 11 |
2.1 | 3 | 33 | Debris deposit on top of last floor | ca 350–370 | 304 |
2.1 | 3 | 36 | Part of sequence of floor layers | 4 c. | 12 |
2.1 | 3 | 36 | Part of sequence of floor layers | 320/1? | 114 |
2.1 | 3 | 36 | Part of sequence of floor layers | 4 c. | 257 |
2.1 | 3 | 93 | Pit fill or covering, above earliest floor? | 292/3? or 320/1 or later | 148 |
2.1 | 3 | 95 | Pit fill, above earliest floor? | 328/9 or 346/7 | 194 |
2.1 | 3 | 95 | Pit fill, above earliest floor? | Unknown | 381 |
2.1 | 4 | 7 | Wind-blown sand in top layer | ca 335–370 | 248 |
2.1 | 4 | 9 | Mudbrick + debris layer over floor | 344/5 | 159 |
2.1 | 4 | 9 | Mudbrick + debris layer over floor | 346/7 | 162 |
2.1 | 4 | 9 | Mudbrick + debris layer over floor | ca 345–370 | 173 |
2.1 | 4 | 9 | Mudbrick + debris layer over floor | ca 345–370 | 250 |
2.1 | 4 | 9 | Mudbrick + debris layer over floor | ca 345–370 | 296 |
2.1 | 4 | 9 | Mudbrick + debris layer over floor | ca 345–370 | 344 |
2.1 | 4 | 9 | Mudbrick + debris layer over floor | ca 345–370 | 354 |
2.1 | 4 | 9 | Mudbrick + debris layer over floor | ca 345–370 | 355 |
2.1 | 4 | 9 | Mudbrick + debris layer over floor | ca 345–370 | 454 |
2.1 | 4 | 10 | Wind-blown sand over floor deposit | ca 345–370 | 249 |
2.1 | 4 | 12 | Ash layer in DSU 9 in part of room | 345/6 | 161 |
2.1 | 4 | 15 | Wind-blown sand between floor and collapse | ca 350–370 | 6 |
2.1 | 4 | 15 | Wind-blown sand between floor and collapse | ca 350–370 | 287 |
2.1 | 4 | 21 | Mudbrick collapse debris above DSU 23 floor level | ca 350–370 | 3 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 7 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 8 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 9 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 14 |
2.1 | 4 | 23 | Top floor level | 343/4 | 124 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 253 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 254 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 255 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 279 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 300 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 301 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 302 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 320 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 362 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 363 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 364 |
2.1 | 4 | 23 | Top floor level | ca 350–370 | 365 |
2.1 | 4 | 29 | Fill of bin | 288/9? | 104 |
2.1 | 4 | 29 | Fill of bin | 3–4 c. | 128 |
2.1 | 4 | 29 | Fill of bin | 290/1? | 175 |
2.1 | 4 | 29 | Fill of bin | 3–4 c. | 256 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 13 |
2.1 | 4 | 35 | Fill under floor level DSU23 | 320/1? | 113 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 126 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 129 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 130 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 131 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 133 |
2.1 | 4 | 35 | Fill under floor level DSU23 | 293/4 or 315/6? | 149 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 168 |
2.1 | 4 | 35 | Fill under floor level DSU23 | 318/9? | 178 |
2.1 | 4 | 35 | Fill under floor level DSU23 | 318/9? | 179 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 181 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 182 |
2.1 | 4 | 35 | Fill under floor level DSU23 | 321/2? | 183 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 184 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 185 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 186 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 187 |
2.1 | 4 | 35 | Fill under floor level DSU23 | Ptolemaic-Early Rom. | 278 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 370 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 372 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 373 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 374 |
2.1 | 4 | 35 | Fill under floor level DSU23 | ca 275–340 | 375 |
2.1 | 4 | 214 | Foundation fill | 321/2? | 122 |
2.1 | 4 | 214 | Foundation fill | 322/3? | 123 |
2.1 | 4 | 214 | Foundation fill | ca 275–340 | 137 |
2.1 | 4 | 214 | Foundation fill | 290/1? | 142 |
2.1 | 4 | 214 | Foundation fill | 285/6? | 147 |
2.1 | 4 | 214 | Foundation fill | ca 275–340 | 217 |
2.1 | 4 | 214 | Foundation fill | ca 275–340 | 218 |
2.1 | 4 | 214 | Foundation fill | ca 275–340 | 219 |
2.1 | 4 | 214 | Foundation fill | ca 275–340 | 341 |
2.1 | 4 | 214 | Foundation fill | ca 275–340 | 438 |
2.1 | 5 | 16 | Wind-blown sand as top fill | ca 300–370 | 251 |
2.1 | 5 | 16 | Wind-blown sand as top fill | ca 300–370 | 361 |
2.1 | 5 | 213 | Wall collapse at base of stairs | ca 300–370 | 271 |
2.1 | 5 | 213 | Wall collapse at base of stairs | ca 325–350 | 318 |
2.1 | 5 | 213 | Wall collapse at base of stairs | ca 300–370 | 435 |
2.1 | 6 | 67 | Roof and wall collapse | ca 300–370 | 379 |
2.1 | 6 | 238 | Compacted layer above one floor and below last floor | ca 300–370 | 172 |
2.1 | 7 | 69 | Ceiling and wall collapse on floor | ca 300–370 | 72 |
2.1 | 8 | F74 | Pit in floor | 287/8? | 102 |
2.1 | 8 | 63 | Vault collapse on floor | 350/1 | 165 |
2.1 | 8 | 63 | Vault collapse on floor | ca 350–370 | 169 |
2.1 | 8 | 63 | Vault collapse on floor | ca 350–370 | 377 |
2.1 | 8 | 63 | Vault collapse on floor | ca 350–370 | 380 |
2.1 | 8 | 74 | Occupation debris above floor | ca 350–370 | 15 |
2.1 | 8 | 74 | Occupation debris above floor | ca 350–370 | 16 |
2.1 | 8 | 74 | Occupation debris above floor | ca 350–370 | 17 |
2.1 | 8 | 74 | Occupation debris above floor | ca 350–370 | 18 |
2.1 | 8 | 74 | Occupation debris above floor | ca 350–370 | 63 |
2.1 | 8 | 74 | Occupation debris above floor | ca 350–370 | 73 |
2.1 | 8 | 77 | Occupation debris above floor, interpenetrating floor level | Ptolemaic or Early Roman | 198 |
2.1 | 8 | 85 | Collapse and debris just above floor | ca 350–370 | 19 |
2.1 | 8 | 98 | Floor debris | 320/1? | 115 |
2.1 | 8 | 102 | Probably floor foundation deposit | 292/3? | 108 |
2.1 | 8 | 102 | Probably floor foundation deposit | 321/2? | 117 |
2.1 | 8 | 102 | Probably floor foundation deposit | 290/1? | 196 |
2.1 | 8 | 102 | Probably floor foundation deposit | ca 275–350 | 197 |
2.1 | 8 | 182 | Backfill and Wind-blown sand | ca 275–350 | 208 |
2.1 | 9 | F93 (north of) | Backfill | Unknown | 349 |
2.1 | 9 | 104 | Probably dump material layer | 325/6? | 155 |
2.1 | 9 | 104 | Probably dump material layer | 325/6? | 156 |
2.1 | 9 | 107 | Fill dumped under floor | 325/6? | 125 |
2.1 | 9 | 107 | Fill dumped under floor | ca 275–350 | 387 |
2.1 | 9 | 114 | Wind-blown sand fill | ca 275–350 | 414 |
2.1 | 9 | 117 | Dump layer | ca 275–350 | 308 |
2.1 | 9 | 117 | Dump layer | ca 275–350 | 396 |
2.1 | 9 | 127 | Fill dumped under floor | ca 315–370 | 76 |
2.1 | 9 | 127 | Fill dumped under floor | 319/20? | 152 |
2.1 | 9 | 127 | Fill dumped under floor | ca 275–350 | 334 |
2.1 | 9 | 127 | Fill dumped under floor | ca 275–350 | 398 |
2.1 | 9 | 127 | Fill dumped under floor | ca 275–350 | 404 |
2.1 | 9 | 128 | Fill dumped under floor | 325/6? | 157 |
2.1 | 9 | 128 | Fill dumped under floor | ca 275–350 | 347 |
2.1 | 9 | 129 | Sandy dump layer in circular structure F93 | 287/8? | 103 |
2.1 | 9 | 129 | Sandy dump layer in circular structure F93 | ca 275–350 | 331 |
2.1 | 9 | 130 | Sandy/ashy debris layer in circular structure F93 | ca 325–350 | 201a |
2.1 | 9 | 130 | Sandy/ashy debris layer in circular structure F93 | ca 275–350 | 415 |
2.1 | 9 | 130 | Sandy/ashy debris layer in circular structure F93 | ca 275–350 | 416 |
2.1 | 9 | 130 | Sandy/ashy debris layer in circular structure F93 | ca 275–350 | 417 |
2.1 | 9 | 132 | Pit slightly later than fill under floor | 290/1? | 106 |
2.1 | 9 | 132 | Pit slightly later than fill under floor | 294/5, 316/7, 322/3 or later | 110 |
2.1 | 9 | 132 | Pit slightly later than fill under floor | ca 275–350 | 332 |
2.1 | 9 | 132 | Pit slightly later than fill under floor | ca 275–350 | 338 |
2.1 | 9 | 133 | Mud floor in corner of courtyard | ca 325–350 | 201 |
2.1 | 9 | 137 | Early dump layer | ca 275–350 | 397 |
2.1 | 9 | 138 | Dump under mud floor DSU 133 | ca 325–350 | 394 |
2.1 | 9 | 146 | Sandy layer under mud-brick collapse | ca 275–350 | 25 |
2.1 | 9 | 146 | Sandy layer under mud-brick collapse | ca 275–350 | 258 |
2.1 | 9 | 148 | Preparation level for floor | ca 340–370 | 314 |
2.1 | 9 | 151 | Dumped debris after closure of courtyard | 319/20? | 151 |
2.1 | 9 | 151 | Dumped debris after closure of courtyard | 355/6 | 167 |
2.1 | 9 | 151 | Dumped debris after closure of courtyard | ca 275–350 | 335 |
2.1 | 9 | 151 | Dumped debris after closure of courtyard | ca 275–350 | 346 |
2.1 | 9 | 151 | Dumped debris after closure of courtyard | ca 275–350 | 348 |
2.1 | 9 | 151 | Dumped debris after closure of courtyard | ca 275–350 | 400 |
2.1 | 9 | 151 | Dumped debris after closure of courtyard | ca 275–350 | 401 |
2.1 | 9 | 151 | Dumped debris after closure of courtyard | ca 275–350 | 402 |
2.1 | 9 | 151 | Dumped debris after closure of courtyard | ca 275–350 | 407 |
2.1 | 9 | 167 | Dumped debris after closure of courtyard | 318/9? | 150 |
2.1 | 9 | 167 | Dumped debris after closure of courtyard | ca 275–350 | 206 |
2.1 | 9 | 167 | Dumped debris after closure of courtyard | ca 275–350 | 207 |
2.1 | 9 | 167 | Dumped debris after closure of courtyard | ca 275–350 | 419 |
2.1 | 9 | 167 | Dumped debris after closure of courtyard | ca 275–350 | 420 |
2.1 | 9 | 167 | Dumped debris after closure of courtyard | ca 275–350 | 421 |
2.1 | 9 | 197 | Pre-construction dumped material | ca 275–350 | 317 |
2.1 | 9 | 200 | Pre-construction dumped material | ca 275–350 | 87 |
2.1 | 9 | 200 | Pre-construction dumped material | 319/20? | 112 |
2.1 | 9 | 200 | Pre-construction dumped material | ca 275–350 | 139 |
2.1 | 9 | 200 | Pre-construction dumped material | ca 275–350 | 140 |
2.1 | 9 | 200 | Pre-construction dumped material | ca 275–350 | 143 |
2.1 | 9 | 200 | Pre-construction dumped material | ca 325–350 | 222 |
2.1 | 9 | 200 | Pre-construction dumped material | 318/9? | 223 |
2.1 | 9 | 200 | Pre-construction dumped material | ca 325–350 | 224 |
2.1 | 9 | 200 | Pre-construction dumped material | ca 325–350 | 319 |
2.1 | 9 | 200 | Pre-construction dumped material | ca 275–350 | 436 |
2.1 | 9 | 200 | Pre-construction dumped material | ca 275–350 | 437 |
2.1 | 9 | 203 | Pre-construction dumped material | ca 275–350 | 211 |
2.1 | 9 | 203 | Pre-construction dumped material | ca 275–350 | 212 |
2.1 | 9 | 212 | Pre-construction dumped material | ca. 275-340 | 40 |
2.1 | 9 | 212 | Pre-construction dumped material | ca 275–350 | 43 |
2.1 | 9 | 212 | Pre-construction dumped material | ca 275–350 | 45 |
2.1 | 9 | 212 | Pre-construction dumped material | ca 275–350 | 46 |
2.1 | 9 | 212 | Pre-construction dumped material | ca 275–350 | 47 |
2.1 | 9 | 212 | Pre-construction dumped material | ca 275–350 | 85 |
2.1 | 9 | 212 | Pre-construction dumped material | 320/1? | 116 |
2.1 | 9 | 212 | Pre-construction dumped material | ca 275–350 | 120 |
2.1 | 9 | 212 | Pre-construction dumped material | ca 275–350 | 135 |
2.1 | 9 | 212 | Pre-construction dumped material | ca 275–350 | 136 |
2.1 | 9 | 212 | Pre-construction dumped material | 318/9? | 213 |
2.1 | 9 | 212 | Pre-construction dumped material | 320/1? | 215 |
2.1 | 9 | 212 | Pre-construction dumped material | ca 275–350 | 336 |
2.1 | 9 | 212 | Pre-construction dumped material | ca 275–350 | 430 |
2.1 | 9 | 212 | Pre-construction dumped material | ca 275–350 | 431 |
2.1 | 9 | 212 | Pre-construction dumped material | ca 275–350 | 433 |
2.1 | 9 | 212 | Pre-construction dumped material | ca 275–350 | 450 |
2.1 | 9 | 215 | Dumped debris | ca 275–350 | 83 |
2.1 | 9 | 215 | Dumped debris | ca 275–350 | 86 |
2.1 | 9 | 215 | Dumped debris | ca 275–350 | 434 |
2.1 | 9 | 218 | Pre-construction dumped material | ca 275–350 | 49 |
2.1 | 9 | 218 | Pre-construction dumped material | 319/20? | 96 |
2.1 | 9 | 218 | Pre-construction dumped material | 321/2 | 121 |
2.1 | 9 | 218 | Pre-construction dumped material | ca 275–350 | 144 |
2.1 | 9 | 230 | Debris layer | ca 275–350 | 50 |
2.1 | 9 | 230 | Debris layer | ca 275–350 | 227 |
2.1 | 9 | 232 | Fill in foundation trench for North wall of house | 330/1? | 231 |
2.1 | 10 | 119 | Debris layer above floor level | ca 350–370 | 75 |
2.1 | 10 | 119 | Debris layer above floor level | ca 350–370 | 268 |
2.1 | 10 | 142 | Dump layer under floor | 294/5 or 316/7, 322/3 or later? | 111 |
2.1 | 10 | 142 | Dump layer under floor | 290/1? | 202 |
2.1 | 10 | 191 | Dump layer under floor | ca 275–350 | 95 |
2.1 | 10 | 191 | Dump layer under floor | 311/2? 325/6? | 158 |
2.1 | 10 | 191 | Dump layer under floor | 290/1? | 241 |
2.1 | 10 | 191 | Dump layer under floor | ca 275–350 | 353 |
2.1 | 10 | 191 | Dump layer under floor | ca 275–350 | 429 |
2.1 | 10 | 202 | Possible dump directly on gebel | ca 275–350 | 352 |
2.1 | 10 | 220 | Debris after or contemporary with floor F106 | ca 350–370 | 226 |
2.1 | 10 | 220 | Debris after or contemporary with floor F106 | 327/8, 337/8, or 345/6 | 228 |
2.1 | 10 | 220 | Debris after or contemporary with floor F106 | ca 275–350 | 237 |
2.1 | 10 | 220 | Debris after or contemporary with floor F106 | ca 275–350 | 274 |
2.1 | 10 | 220 | Debris after or contemporary with floor F106 | ca 275–350 | 321 |
2.1 | 10 | 241 | Channel fill from construction | ca 250–325 | 62 |
2.1 | 10 | 241 | Channel fill from construction | ca 275–350 | 342 |
2.1 | 10 | 241 | Channel fill from construction | ca 275–350 | 446 |
2.1 | 10 | 255 | Channel fill from construction | ca 275–350 | 337 |
2.1 | 11 | 0 (backfill) | Backfill | Unknown | 399 |
2.1 | 11 | 79 | Wind-blown sand fill | 344/5? | 160 |
2.1 | 11 | 229 | Wind-blown sand fill | ca 350–370 | 232 |
2.1 | 11 | 235 | Collapse onto occupation debris | ca 350–370 | 145 |
2.1 | 11 | 235 | Collapse onto occupation debris | 319/20? | 233 |
2.1 | 11 | 235 | Collapse onto occupation debris | ca 350–370 | 234 |
2.1 | 11 | 235 | Collapse onto occupation debris | ca 350–370 | 239 |
2.1 | 11 | 235 | Collapse onto occupation debris | ca 350–370 | 273 |
2.1 | 12 | 226 | Wind-blown sand fill | ca 350–370 | 52 |
2.1 | 12 | 226 | Wind-blown sand fill | ca 350–370 | 54 |
2.1 | 12 | 226 | Wind-blown sand fill | ca 350–370 | 238 |
2.1 | 12 | 226 | Wind-blown sand fill | ca 350–370 | 236 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 53 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 56 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 57 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 58 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 59 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 61 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 89 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 90 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 91 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 92 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 93 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 94 |
2.1 | 12 | 245 | Collapse of north wall and vault | 357/8 | 235 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 240 |
2.1 | 12 | 245 | Collapse of north wall and vault | 356/7 | 242 |
2.1 | 12 | 245 | Collapse of north wall and vault | 362/3 | 243 |
2.1 | 12 | 245 | Collapse of north wall and vault | 357/8 | 244 |
2.1 | 12 | 245 | Collapse of north wall and vault | 356/7 | 245 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 275 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 276 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 277 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 325 |
2.1 | 12 | 245 | Collapse of north wall and vault | 354/5? | 326 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 327 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 328 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 447 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 448 |
2.1 | 12 | 245 | Collapse of north wall and vault | ca 350–370 | 449 |
2.1 | 13 | F150 | Top floor level | ca 350–370 | 55 |
2.1 | 13 | 83 | Wind-blown sand | ca 350–370 | 193 |
2.1 | 13 | 185 | Surface layer, Wind-blown sand | ca 350–370 | 261 |
2.1 | 13 | 186 | Vault collapse in W. part of room | ca 350–370 | 262 |
2.1 | 13 | 186 | Vault collapse in W. part of room | ca 350–370 | 263 |
2.1 | 13 | 186 | Vault collapse in W. part of room | ca 350–370 | 265 |
2.1 | 13 | 186 | Vault collapse in W. part of room | ca 350–370 | 266 |
2.1 | 13 | 187 | Clean wind-blown sand fill | ca 350–370 | 225 |
2.1 | 13 | 199 | Vault collapse outside room | ca 350–370 | 34 |
2.1 | 13 | 199 | Vault collapse outside room | ca 350–370 (354/5?) | 35 |
2.1 | 13 | 199 | Vault collapse outside room | ca 350–370 | 48 |
2.1 | 13 | 199 | Vault collapse outside room | ca 350–370 | 79 |
2.1 | 13 | 199 | Vault collapse outside room | ca 350–370 | 138 |
2.1 | 13 | 199 | Vault collapse outside room | ca 350–370 | 272 |
2.1 | 13 | 207 | Vault collapse from above doorway | ca 350–370 | 80 |
2.1 | 13 | 207 | Vault collapse from above doorway | ca 350–370 | 81 |
2.1 | 13 | 207 | Vault collapse from above doorway | 353/4? | 214 |
2.1 | 13 | 207 | Vault collapse from above doorway | ca 350–370 | 289 |
2.1 | 13 | 208 | Fill of northwest bin | ca 350–370 | 38 |
2.1 | 13 | 208 | Fill of northwest bin | ca 350–370 | 39 |
2.1 | 13 | 208 | Fill of northwest bin | ca 350–370 | 82 |
2.1 | 13 | 208 | Fill of northwest bin | ca 350–370 | 84 |
2.1 | 13 | 208 | Fill of northwest bin | ca 350–370 | 288 |
2.1 | 13 | 208 | Fill of northwest bin | ca 350–370 | 432 |
2.1 | 13 | 209 | Fill of northeast bin | ca 350–370 | 64 |
2.1 | 13 | 216 | Occupation debris on floor | 358/9 | 41 |
2.1 | 13 | 216 | Occupation debris on floor | ca 350–370 | 42 |
2.1 | 13 | 216 | Occupation debris on floor | ca 350–370 | 44 |
2.1 | 13 | 216 | Occupation debris on floor | ca 350–370 | 216 |
2.1 | 13 | 216 | Occupation debris on floor | ca 350–370 | 269 |
2.1 | 13 | 216 | Occupation debris on floor | ca 350–370 | 290 |
2.1 | 14 | 184 | Wall and vault collapse | ca 350–370 | 209 |
2.1 | 14 | 184 | Wall and vault collapse | ca 350–370 | 210 |
2.1 | 14 | 194 | Wall and vault collapse | ca 350–370 | 37 |
2.1 | 14 | 194 | Wall and vault collapse | ca 350–370 | 154 |
2.1 | 14 | 194 | Wall and vault collapse | ca 350–370 | 229 |
2.1 | 14 | 194 | Wall and vault collapse | ca 350–370 | 264 |
2.1 | 14 | 194 | Wall and vault collapse | ca 350–370 | 267 |
2.1 | 14 | 211 | Wall collapse and floor debris | 361/2 or 367/8 | 270 |
2.1 | 14 | 221 | Occupation debris with brick & sand | ca 350–370 | 66 |
2.1 | 14 | 221 | Occupation debris with brick & sand | ca 350–370 | 220 |
2.1 | 14 | 221 | Occupation debris with brick & sand | ca 350–370 | 221 |
2.1 | 14 | 221 | Occupation debris with brick & sand | ca 350–370 | 439 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 65 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 67 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 68 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 69 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 70 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 88 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 247 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 292 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 293 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 322 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 323 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 324 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 330 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 440 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 441 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 442 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 443 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 444 |
2.1 | 14 | 233 | Habitation layer just above floor | ca 350–370 | 453 |
2.1 | 15 | F109 | Stack of baked brick | ca 275–350 | 451 |
2.1 | 15 | F78 | North-South wall | ca 275–350 | 452 |
2.1 | 15 | 0 (backfill) | Backfill | ca 350–370 | 291 |
2.1 | 15 | 109 | Ash deposit in test pit | ca 350–370 | 329 |
2.1 | 15 | 145 | Flat roof collapse on occupation debris | ca 350–370 | 77 |
2.1 | 15 | 145 | Flat roof collapse on occupation debris | ca 350–370 | 204 |
2.1 | 15 | 145 | Flat roof collapse on occupation debris | ca 350–370 | 281 |
2.1 | 15 | 148 | Sand and brick debris under floor | ca 350–370 | 412 |
2.1 | 15 | 152 | Flat roof collapse on occupation debris | ca 350–370 | 29 |
2.1 | 15 | 152 | Flat roof collapse on occupation debris | ca 350–370 | 405 |
2.1 | 15 | 152 | Flat roof collapse on occupation debris | ca 350–370 | 409 |
2.1 | 15 | 152 | Flat roof collapse on occupation debris | ca 350–370 | 410 |
2.1 | 15 | 157 | Occupational and mud brick debris | ca 350–370 | 30 |
2.1 | 15 | 157 | Occupational and mud brick debris | ca 350–370 | 31 |
2.1 | 15 | 157 | Occupational and mud brick debris | 360/1? | 32 |
2.1 | 15 | 157 | Occupational and mud brick debris | ca 350–370 | 60 |
2.1 | 15 | 157 | Occupational and mud brick debris | 363/4 or 364/5? | 146 |
2.1 | 15 | 157 | Occupational and mud brick debris | ca 350–370 | 260 |
2.1 | 15 | 157 | Occupational and mud brick debris | ca 350–370 | 284 |
2.1 | 15 | 157 | Occupational and mud brick debris | ca 350–370 | 285 |
2.1 | 15 | 157 | Occupational and mud brick debris | ca 350–370 | 350 |
2.1 | 15 | 157 | Occupational and mud brick debris | ca 350–370 | 413 |
2.1 | 15 | 157 | Occupational and mud brick debris | ca 350–370 | 423 |
2.1 | 15 | 157 | Occupational and mud brick debris | ca 350–370 | 425 |
2.1 | 15 | 157 | Occupational and mud brick debris | ca 350–370 | 426 |
2.1 | 15 | 261 | Rubble layer below floor | ca 275–350 | 246 |
2.1 | 15 | 261 | Rubble layer below floor | ca 275–350 | 343 |
2.1 | 15 | 152 | Flat roof collapse on occupation debris | ca 350–370 | 283 |
2.1 | 16 | 158 | Wind-blown sand below mud-brick collapse | ca 350–370 | 205 |
2.1 | 16 | 158 | Wind-blown sand below mud-brick collapse | ca 350–370 | 312 |
2.1 | 16 | 158 | Wind-blown sand below mud-brick collapse | ca 350–370 | 408 |
2.1 | 16 | 161 | Sand and dust above floor | ca 350–370 | 282 |
2.1 | 16 | 161 | Sand and dust above floor | ca 350–370 | 313 |
2.1 | 17 | 153 | Wind-blown sand | ca 350–370 | 27 |
2.1 | 17 | 153 | Wind-blown sand | ca 350–370 | 309 |
2.1 | 17 | 153 | Wind-blown sand | ca 350–370 | 310 |
2.1 | 17 | 155 | Sand and dust above floor | ca 350–370 | 311 |
2.1 | 17 | 155 | Sand and dust above floor | ca 350–370 | 403 |
2.1 | 17 | 161 | Sand and dust above floor | ca 350–370 | 411 |
4.1 | 0 | 1 | Ptolemaic-Early Rom. | 280 | |
4.1 | 0 | 1 | Ptolemaic-Early Rom. | 385 | |
4.1 | 0 | 2 | Ptolemaic-Early Rom. | 305 | |
4.1 | 0 | 2 | Dyn. 25 or earlier | 306 | |
4.1 | 0 | 8 | 3–4 c. | 134 | |
4.1 | 0 | 11 | 3–4 c. | 190 | |
4.1 | 0 | 11 | 293/4? or 321/2 or later | 195 | |
4.1 | 0 | 11 | Ptolemaic-Early Rom. | 378 | |
4.1 | 0 | 11 | Unknown | 382 | |
4.1 | 0 | 11 | ca 315–370 | 383 | |
4.1 | 0 | 11 | Ptolemaic-Early Rom. | 384 | |
4.1 | 0 | 14 | 3 c.? | 307 | |
4.1 | 0 | 18 | 356/7 | 127 | |
4.1 | 0 | 18 | 4 c. | 170 | |
4.1 | 0 | 18 | Unknown | 386 | |
4.1 | 0 | 43 | 291/2? | 107 | |
4.1 | 0 | 43 | 293/4? or 321/2 or later | 199 | |
4.1 | 0 | 60 | 3–4 c. | 74 | |
4.1 | 0 | 60 | 3–4 c. | 171 | |
4.1 | 0 | 61 | 285/6? or 328/9 or later | 97 | |
4.1 | 0 | 61 | 288/9? or 316/7 or later | 105 | |
4.1 | 0 | 64 | 284/5? or 327/8 or later | 203 | |
4.1 | 0 | 77 | 285/6? or 328/9 or later | 98 | |
4.1 | 0 | 79 | Unknown | 33 | |
4.1 | 0 | 81 | 3–4 c. | 28 | |
4.1 | 0 | 86 | Unknown | 418 | |
4.1 | 0 | 86 | Unknown | 424 | |
4.1 | 0 | 90 | ca 350–370 | 286 | |
4.1 | 0 | 99 | 4 c. | 78 | |
4.1 | 0 | 99 | 293/4? or 321/2 or later | 109 | |
4.1 | 0 | 99 | Ptolemaic-Early Rom. | 427 | |
4.2 | 0 | 1 | Unknown | 339 | |
4.2 | 0 | 6 | Ptolemaic-Early Rom. | 422 | |
4.2 | 0 | 7 | Unknown | 406 | |
4.2 | 0 | 8 | Unknown | 333 | |
4.2 | 0 | 10 | Unknown | 351 | |
4.2 | 0 | 13 | Unknown | 153 | |
4.2 | 0 | 13 | Ptolemaic-Early Rom. | 428 | |
4.2 | 0 | 17 | Later Ptolemaic | 315 |