Tablet 581

description (a) 40 x 93 mm. (b) 40 x 93 mm. (c) 37 x 93 mm. (d) 38 x 93 mm. (e) 41 x 93 mm. Bowman and Thomas (1996), no.2 with Plates VIII-IX. Plates 2-3.
Archaeological data. Location: SG (bonfire site). Period: 3.
This group of tablets was found at the heart of the bonfire site on the intervallum road, by the side of the turf rampart, immediately outside the Period 3 praetorium, congealed along with many other fragments in a mixture of ash, soot, bone fragments and pottery (see Introduction, above, p.00). Some of the tablets in this group may well have fragments dispersed under various inventory numbers (and some of the pieces inventoried under this number do not belong with the text here published). The tablets contain an account consisting of five half-diptychs (a-e). Three of the five pieces (b, c and d) have writing on both sides, the others (a and e) on one side only. This is evidently part of a notebook containing domestic accounts of the praetorium in Period 3. It may be the work of the same hand which wrote 191, and therefore perhaps also 194, 197 and part of 196. This is most obvious in the case of lines 59-86. This section, however, seems to be in a finer and more elegant hand; it may be that more than one hand was responsible for the document but it perhaps more likely that the writer simply sharpened or replaced his pen. Also related is 582 which is an account mentioning the poultryman Chnisso whose name occurs in line 24 of the present text. Cf. also 302.2, 616.B.1, 679.3.
The leaves are all complete at top, bottom and left margin, but lack the right margin. It is uncertain how much is missing at the right. The break occurs at the point at which the second of the two tie-holes in each leaf has been punched. Comparison with the dimensions of 310, a complete diptych with tie-holes, suggests something in the region of at least 30-40%, which would accommodate perhaps 6-8 letters. Some of the restorations proposed in the notes, however, appear to require more than this to have been lost.
Examination of the physical characteristics of the leaves has enabled us to offer a reconstruction of the sequence of the leaves in what we believe to be the correct order. We have observed as criteria for matching and ordering the pieces: (i) the placement of the notches and tie-holes; (ii) the grain of the wood; (iii) patterns of abrasion and deposit of dirt on the surfaces, indicating clearly which surfaces were face-to-face; (iv) a few holes presumably made by a nail or by an insect boring through the wood. These observations enable us to show beyond all reasonable doubt that 581 uses the concertina format which we suggested for 190. The diptychs were tied together by thongs passed through the tie-holes bored close to the top and bottom edges. In discussing the format of 190, we noted that it "can in no way be described as a primitive codex" and that it "clearly lacks one of the features essential to the codex form, namely writing on both sides of the leaf" (Tab.Vindol.I, pp.40-4). It should therefore be emphasised that 581 differs in that it does have writing on both sides of the leaf; after reading one side the user would have to turn it over and rotate it through 180 degrees to read the other side (see Fig. 1). It should also be noted that in the order we have suggested there is a blank half-leaf at the end of the sequence on each of the faces.
Even with these features and differences, we would still not wish to claim 581 as any kind of a primitive codex; merely as another important piece of evidence for the format and use of folded wooden writing-material in a period and a context which is surely of significance for the development of the early codex. In this connection it is interesting to note that A.Wouters has shown very good reason for thinking that a papyrus codex in the Chester Beatty Library was made up in concertina fashion: see Wouters (1988), 19-23. Also relevant is Tomlin's discussion of Tab.Luguval. 1a. (Tomlin (1998), 39-40) which was written on a single strip of wood, scored three times, folded into four and secured by a single hole driven through one corner; this could then be used like a short roll of papyrus. The subject has also been discussed by Haran (1996), who thinks that an objection to the concertina hypothesis lies in the fact that other tablets which could not have been assembled in this fashion also have tie-holes. We see no problem with this and suppose that local and individual variations in the use of such writing material, such as that at Carlisle, are by no means surprising. Tie-holes in single diptychs can be used for sealing letters and, indeed, there are cases where it can be shown that the tie-holes were bored before the tablets were used (see e.g. below, lines 14-15 note, 645).
The foregoing paragraphs essentially restate what we wrote in the editio princeps. A.R.Birley (2002), 128 has suggested a different order for the leaves, and one which we had ourselves considered, placing e.Back (lines 59-72) at the beginning of the sequence. The consequence of this would be that the account would start in AD 101 (our lines 59-64), followed by entries for AD 102 (our lines 65-72, then lines 1-8), AD 103 and 104 (our lines 9-19 and 20-58), then AD 105 (our lines 73-111). It must be acknowledged that this arrangement has two advantages: (i) it offers a more logical sequence of dates and (ii) it places both the blank leaves (a.Back and e.Front) at the end of the sequence. Since this is obviously important, we have readdressed the question and re-examined the physical characteristics of the leaves to see whether the new order should be adopted. While we would not to rule out the new suggestion definitively, there is no doubt that the physical characteristics of the leaves, particularly the holes in the top and bottom edges which need to correspond, offer a much better fit in our original sequence. In particular, the small hole in the top margin of (e) matches one in the bottom of (d), whereas there is no such match between (e) and (a) when (e) is placed at the beginning of the sequence. The physical arrangement of the tablet is crucial and Birleys suggestion ((2002), 173, n.13) that the leaves may have been separated into individual sheets before use seems to us extremely unlikely. A second advantage in our sequence is that it presents two sections, one "consumed" and one "disbursed" and the language seems to support this: a(b) is used in our first section but not the second, and the second section has names in the dative not the ablative. We have therefore retained our original order in this re-edition. See further notes to lines 60 and 65.
Like 190, 191, 194, 196 and 197 the present text clearly belongs in the context of the presence in the praetorium of Flavius Cerialis, prefect of the Ninth Cohort of Batavians. This can be inferred not merely from the archaeological context, but also from the occurrence in the account of the name of Brocchus, certainly Aelius Brocchus, Cerialis' most frequent correspondent (line 62 etc., cf. 622 introd.).
There are difficulties in elucidating the connected questions of the nature of the account and the range and sequence of dates in the document. In the we suggested the probability that the text is not complete, and we hold to this view and think that if line 1 is to be understood as a heading (see note), it must be a sub-heading; since, on our arrangement, the text begins on the bottom half of a diptych it cannot be the start of the account. There is a substantial uacat after line 111, which could be the end of the account.
As regards the nature of the account, we think that the key to understanding the organisation of the text lies in supposing that there are in fact two sequences of entries, rather than a single continuous series, which do not divide neatly between the front and back faces. We suggest that the first sequence runs as far as line 42 or 44 and that the second begins at line 43 or 45, with the heading expensa[. It may be that we should further subdivide the second section into three parts: (i) lines 46-58 xv K(alendas) Iunias - perierunt, followed by a blank half-leaf; (ii) lines 59-86; (iii) lines 87-111.
The entries preserved all appear to relate to poultry with one probable exception (see lines 3-4 note). It is unfortunately not possible to be certain whether this is a cash account or merely a quantification; in no entry is a cash sum preserved but this may simply be due to the loss of the right-hand side of the leaves (the same may well be true of 190). Since what we take to be the heading of the first preserved section (line 1) is not fully preserved, we cannot be certain of its interpretation (see note). The first sequence of entries records the items with the names of various persons in the ablative plus ab. We now think that these should be understood either as "accepta ab NN", which would presuppose receipt of cash sums lost at the right; or as "absumpti [sc. pulli] or absumpta ab NN", that is "fowl [or items] consumed (or taken) by NN". The second sequence, with the heading expensa[ (line 45), appears to record poultry disbursed mainly for various events in the social calendar of the inhabitants of the praetorium, but in two cases supplied for use elsewhere (Coria, line 55, "foris", line 98). On this hypothesis, the term expensa[ would refer not to cash expenditure, but to outgoing items disbursed from the writer's "department". For convenience we designate the first section of the account as "consumed" and the second as "disbursed".
We are inclined to interpret the account not as a running daybook but as a collation of rough working notes compiled retrospectively by a domestic manager, probably one of the slaves, in Cerialis' household; the rough and retrospective nature of the record is strongly suggested by the errors, inconsistencies and gaps in the date entries especially at lines 61-3, 91 and 95. There are substantial lapses of time between the dates in lines 7 and 10, and between those in lines 18 and 21 (cf. lines 54 and 59, and 91 and 95).
For convenience we append a tabulation of the dates, followed by the line numbers in brackets, in the order in which they occur in the account and our proposed division into two sections.

Section 1 (consumed)
11 April (2)
16 May (5) AD 102
18 May (7)
Fifth consulship of Trajan (9):
26 April (10)
5 June (14) AD 103
10 June (16)
11 June (18)
Consulship of Attius Suburanus (20):
1 January (21)
2 January (25)
1 March (27)
30 March (29) AD 104
31 March (30)
23 April (32)
30 April (39) ... (?) (43)
Section 2 (disbursed) We do not know the years to which the dates in this section belong. We think it most probable that they relate to the same years as those attested in the preceding section, thus: lines 46-58, May-June 102; lines 59-86, June 103-May 104; lines 87-111, May-July 104.
18 May (46)
25 May (50)
13 June (52) AD 102
14 June (54)
10 June (59)
30 August (61) AD 103
25 December (63)
1 January (65)
17 January (67)
21 February (69)
1 March (71)
15 March (73)
21 March (75)
4 April (77)
29 April (79) AD 104
4 May (81)
9 May (87)
29 May (91)
1 May (95)
6 June (99)
16 July (101)
The sequences and range of dates have important implications for the chronology of the phases of occupation at Vindolanda. If we are correct in our belief that this account belongs to the household of Flavius Cerialis, and therefore to the period when the Ninth Cohort of Batavians was at Vindolanda, it is clear that its presence lasted until at least 30 April 104 (line 39), or even 16 July 104 (line 101). This is welcome as one of the few pieces of firm chronological evidence for the location of Batavian auxiliary units in this period. On the evidence for Batavian cohorts in Britain at this period, and for the difficulties in identifying the series of Batavian units see Jarrett (2003), 54-5. A cohors IX Batauorum milliaria is now attested in Raetia in AD 116 (ibid. 55 n.70). The evidence of 581 would seem, however, to demand some modification of the conclusion reached by Robin Birley on the basis of the archaeological evidence: " ... there was evidence ... that Period 4 timbers had been felled in the winter of AD 103/4. Taking into account the evidence for a period of abandonment before demolition took place, a closing date for Period 3 in either AD 102 or early AD 103 seems certain" (VRR I, 89). We can only suggest that if the dendrochronological evidence is precise and accurate, there must have been a considerable interval between the felling of the timbers and their use in the Period 4 constructions, and that the Ninth Cohort of Batavians was still in occupation in the first half of AD 104. We are grateful for David Taylor for his observation (pers.comm.) that the ideal time for working oak, used by the army for structural purposes, is 18 months after felling. A.R.Birley (2002), 69 and passim would extend this occupation to the summer of AD 105, based on his suggested sequence of dates in the present account (on which see above).
pulli and anseres which we identifed in the as chickens and geese. For evidence of the use of poultry in the Roman military diet see Davies (1989), 195-6; for its use at Vindolanda see Pearce (2002) and cf. VRR III, 113-4; and for eggs see 193.5, 302.4 and cf. 194.A.5. We are grateful to Dr. D.J.Breeze for the information that chicken bones are generally only found in very small numbers at Roman forts. He suggests that this may be because the bones of such small animals tend to disintegrate more quickly than the bones of domestic and wild animals (cf. Hodgson (1968), 127-62). For the rearing and care of chickens and geese see Columella 8.2-7, 13-14. It is worth noting that Columella uses the term pullus to refer to goslings as well as chickens (and it is applied to the young of other animals too), which leaves open the possibility that this account may be concerned solely with geese and goslings. We are, however, still inclined to think that pulli are chickens; they seem to be considerably commoner than geese and it is likely that in 302, where we have 20 pulli and 100 eggs, they are chickens rather than geese. If our suggested reading and interpretation of lines 37-8 and 109-10 is plausible, we must be dealing with chickens. The nice illustration of chickens on the gravestone of a pullularius, see Plate 12a in Maxfield (1981), also supports this. Despite the fact that poultry does not seem to be a luxury item, several of the entries in this account seem to relate to its use on special occasions, a possible discharge (line 60), the visit of the provincial governor (line 96), the presence of fellow-officers from other forts (lines 62, 64 etc.), perhaps a religious festival (line 71, cf. 190.c.32-33 and see Henig (1982) for the epigraphic evidence for religion in Roman Britain). The use of chickens for sacrifice is of course well known, and the chickens were eaten as well. The second section of the text (lines 45-111) can hardly be described as the social diary of Cerialis and his family, nor as the festal calendar of the household, but it does contain indications of activities in both these categories.
Our reading and interpretation of the text differs in several significant respects from the (notably lines 1, 33, 74, 98, 103, 109-10, on which see the notes). We have also introduced a number of minor changes to the text (e.g. underdots and individual letters) on the basis scans of the tablet made by Dr. John Pearce. Fragment (f) was included in the, but we do not now think it belongs with this account; it has therefore been removed and placed in Descripta (847). We have also simplified the presentation of the transcript, leaving only designation of the leaves (now in lower case) and the line numbers, and separating the text on each of the pieces by a space.


d: traces [ n
traces [ n
45 expensa[ n
xv K(alendas) Iunias m[ n
nas pullus [ n
eodem die cena.[ n
absumptus p[ullus] n
50 ṿiii K(alendas) Iunias [
nio pullụṣ [ n
Idibus Iunis .[ n
legati [ n
xviii K(alendas) Iụḷịaṣ [
55 C̣oriṣ iussụ [ n
[[traces ṣ]]
eodem die ịṇ.[ n
perierunt [


11(?) April
the decurion(s)
of the 1st beer (?)
16 May, (by?)
the brewer
18 May, by (?)
In the fifth consulship of Trajan
26 April
by Crescens

on the same day, by
a goose

5 June, (by?)
Suetius (?)
10 June, (by?)
the brewer
11 June, (by?)
In the consulship of Sex. Attius Suburanus
1 January, by
veteran, chickens
on the same day, by Sautenus (?)
by Chnisso
2 January

1 March, by Ma
on the same day, by Candidus (?)
30 March, by Mar
31 March, by Exsomnius (?)
line deleted
23 April, by V
in charge of the draft-animals of Brocchus (?)
total, geese
likewise, geese
nursling chicks (?)
likewise, nurslings (?)
30 April
likewise, chickens
through Comm (?)
total, chickens

Date ?

18 May
a chicken
on the same day, for 's dinner (?) ...
chicken (?) consumed, 1 (?)
25 May
, a chicken ...
13 June
of (?) the legate
14 June
at Coria, on the instructions of

on the same day in
there have died

10 June
discharge of Flavinus (?)
30 August
for Niger and Brocchus
25 December
for Brocchus' dinner (?)
1 January, through (?)

17 January
for Brocchus
21 February
from the pen
1 March, for the lord(s) (?)
of the Matronalia (?)

15 March
for Niger and Lae
21 March,

4 April
for Brocchus
29 April,
for September
4 May
with Sautenus
total disbursed

and geese, number
from these (?)

9 May
for (?) Onesimus with the standards (?)
on the same day, for Sautenus
in the pen
29 May as lunch for
and Flavinus, consumed (?)
on the same day,
with Sautenus
1 May, for the singulares (?)
on the visit of the governor
consumed (?) at lunch
likewise outside for Myr (?)
6 June
chickens, number 4(?)

16 July
through Surenus the centurion(?)
, number 12
on the same day
for nus(?) chickens
in the hands of Sautenus
, chickens
total, chickens, number 20+
from these, Tanagrian(?)
remainder, sterile (?)
total (?), chickens, number 7+


a.1. traces [ ] ab .[ :; on the scan there are no traces of writing before a and what follows it looks less like than it does on the photograph. If the writing is centred, as it appears to be on the scan, then we might have a sub-heading and we could read aḅṣ[ (the mark at the edge is consistent with s), suggesting absumpti (sc. pulli/anseres), or perhaps absumpta (neuter plural, cf. line 45 and note). An alternative is to disregard the mark at the edge and read ac̣[, i.e. presumably accepta; this would provide a better counterpart for expensa[ (line 45) but would presuppose that there were cash sums in the missing portion at the right, since it is hardly to be supposed that the birds themselves were received from the named individuals. On balance, we favour the first solution. Cf. also line 36 note. Since this piece is the bottom half of a diptych, on our arrangement of the leaves, it cannot be the beginning of the whole account.

a.2. The traces of the first digit look more like a long initial than part of but we cannot exclude ṿi.

a.3-4. Following the pattern of the other entries in this section, we would expect to have decurioṇ[e or decurioṇ[ibus with a/ab in the preceding line followed by coh(ortis) or alae, with the number and name of the unit in line 4. The medial dot on either side of i at the start of line 4 must indicate that it is to be understood as a numeral (cf. 295.4). Following that, we still think c̣eru[ most probable, though the first letter could be f or s. We have found no ala or cohort whose name begins with these letters (there are several units with the number i; see now Jarrett (2003), 33-77). If we supposed that the name was simply omitted in an unofficial account, we might have decurion[e/-ibus alae or coh(ortis) / ·i· .The restoration of coh(ortis) would be justified if we were dealing with a cohors equitata such as the cohors i Batauorum (Jarrett (2003), 55-6). As an alternative, we have considered the possibility that we might restore turmae / â‹…iâ‹…, referring to the decurion in charge of a cavalry troop of an ala or cohors equitata; auxiliary turmae are, however, like centuriae normally identified by the name of the officer in command (see e.g. RMR, Index 4, s.v., O.Flor.3-5; RIB 557 might conceivably conceal turma followed by a number but this text is garbled); legionary cavalry were not organised in turmae. The best hypothesis is perhaps that we have a reference to the decurion(s) of an ala; the ala i Tungrorum, for instance, is known to have been in Britain in this period and might well have been in the region of Vindolanda, see Jarrett (2003), 44 (equally, there might have been a detachment of the unit at Vindolanda temporarily, cf Tab.Vindol.II, p.93). The traces of the last letter can hardly be read as anything but u, and this suggests c̣ẹru[esae followed by a quantity (cf. 190.c.6 etc.); note the occurrence of a ceruesarius two lines below. It is tempting to think that the entry might refer to the allocation of beer requested from Cerialis by the decurion Masc(u)lus in 628, but the terms in which he writes to Cerialis make it clear that he is a decurion of the Ninth Cohort of Batavians (which was equitate). Another possibility which we have also considered is a reference here to venison (ceruina) or deer (cf. 191.10). If either of these suggestions is correct, it follows that the account does not record only poultry, although this entry would be the only clear exception. A.R.Birley (2002), 130 n.17, following P.Holder, suggests that this should be read as a distribution of equal quantities of beer (ceru[esae), presumably 1 modius or 1 sextarius, to a number of decurions. All our accounts which specify quantities, however, put the name of the commodity before the quantity, not after it.

a.6. ceruesaṛị[o: for a ceruesarius named Atrectus at Vindolanda see 182.14.

a.8. pulli: pullus as a term for the young of animals is by no means confined to poultry, see OLD s.v.1; it is used of goslings but the context here suggests very strongly that chickens are meant (see introd., above). The syntax is of some interest and importance. Here and elsewhere in this account we have the nominative case, which is by no means normal in the accounts from Vindolanda (cf. Adams (2003)). This account gives us no evidence of a price cited with the commodity; for an inferred unit price for chickens see 582.b.4 note.

a.9. Traiano V[: it is noteworthy that there is a considerable gap between the name and the digit. There is no reason to doubt that this is a consular date (AD 103), especially in view of line 20, where the name of the consul of AD 104 certainly occurs (see note ad loc.). It is notable that there is a gap of over 11 months between the dates in lines 7 and 10.

a.10. It is difficult to determine what is and what is not ink after the date. The trace at the right is perhaps compatible with ạ[.

a.11. The name Crescens is common at Vindolanda, see Tab.Vindol.II, Index II, s.v. and 574.9 note.

a.13. anseṛẹṣ,; it is impossible to be sure whether there are traces of ink after anseṛ and it is more likely that we have the singular.

b.14-15. The pattern of entries suggests that we are likely to have ab at the end of line 14. The name Suetius is very uncommon; NPEL offers only one example of its use as a cognomen in Belgica/Germany, T.Iulius Suetius, CIL 13.7911; it does not appear in the index of LC. There is also an inscription from Italy recording a military tribune (and praefectus cohortis?) L.Suetius Marcellus (?) (PME S82 bis), but there is no reason to identify him with a Suetius at Vindolanda, other than the rarity of the name. An alternative possibility would be to restore ab Ma(n)-/suetio (the name is well attested, cf. 188.12 and 580.6) and the fact that suetio in line 15 is indented would not rule it out. As for what follows, adiu[ looks good, with the tie-hole coming between a and d. This could be interpreted in various ways: as the beginning of adiutore (cf. 199.2), if ab was in the missing part of line 14; or it might be descriptive of his function: compare 180.33, patri [a]d ị[uu]ẹncoṣ ; here, iu[menta might be another possibility, cf. line 33.

b.17. ceruesario: cf. line 6 and note.

b.19. Vatto is not a common cognomen but it does occur in Belgica/Germany (see NPEL s.v.). The suggests that after that we might have p̣ụḷ[li, but the scan does not support this.

b.20-21. The writer seems to have begun K(alendis) I(anuaris) and then to have written the beginning of the consular name over the i without erasing what he had written. There is no indication of abbreviation after Sex. This entry is important because it offers the one clear and unimpeachable reading of a consular date in this text (see line 9 note). Sex. Attius Suburanus held only one ordinary consulship, in AD 104 (his first tenure, in AD 101, was as suffect consul). In line 21 the obvious reading is Ianuaras or Ianuaria and we think it likely that the writer put one of these before correcting it to Ianuarịs. The trace at the right-hand edge of line 21 is too exiguous to suggest any reading; we suppose that it must be the beginning of a personal name, going with ueterạno in the following line.

b.22. Although the traces at the right are abraded, it seems impossible to read ueterinario here. and our reading looks more or less certain on the scan. For a ueteranus named Ingenuus see 187.i.11.

ib.23. Sau.[ : probably the personal name Sautenus, which also occurs in line 82 etc.; see also 188.back 7 and note and 709.1.

b.24. Chnisson[e: the name, which is presumably Germanic, also occurs in 582.a.i.2, ii.3 in connection with poultry.

c.27. Mạ..[: although the upper part of the letter after may be obscured by dirt, mạṛị[ (cf. line 29 and note) is not an easy reading; also difficult is mạṇ[ (e.g. Mạṇ[suet(i)o), unless we assume that n is in a form different from that usually used by this writer; the form of n for n(umero) in line 85 is perhaps close enough to support it.

c.28. C̣an[: this suggests the name Candidus, which often occurs in the tablets (see Tab.Vindol.II, Index II s.v.). However, the first letter here could well be ; LC and NPEL list a number of cognomina beginning Pan-, of which Panno, Pancuius and Panturo are attested (albeit rarely) in Belgica/Germany.

c.29. Mar[: we are confident of the reading. Note Marinus in 343.i.3 and 641, Maritimus, the addressee of 645, Marcellinus (609.7) and possibly Martialis (609.back 3).

c.30. ab Exṣ..[: perhaps Exṣp̣ẹ[dito (the name Expeditus is found in 161.5 and 171.a.1), or better, Exsọṃ[nio (for a centurion of this name see 575.5n., 182.ii.13).

c.31. The erased letters are very abraded, but are probably a date.

c.33. iumenṭaṛio : this seems to us the most probable reading, although is far from easy. The term does not occur elsewhere in the Vindolanda tablets, but the occurrence of someone in charge of, or dealing with, iumenta is not unexpected (cf. 343.20). In the previous line, after ab, there was presumably a personal name. The traces after br are very abraded and ambiguous; they are compatible with Brọc̣[chi, which looks quite plausible on the photo; that is, the iumentarius in charge of Brocchus' animals. brịg̣[, which puts us in mind of the place-name Briga (190.c.38, 292.c.2), is another possibility; i.e. the iumentarius at Briga, or perhaps iumenṭaṛio Br ịg̣[ensi (cf. 180.25 note).

c.34. anṣ[er-: we regard the restoration as certain in view of the next line. What is uncertain is whether we should restore the genitive plural or the nominative plural (for the latter cf. line 108).

c.35ff.. Some of these lines may have been inserted after the writing of the main part of the text, cf. line 39 note.

c.36. p̣ulli a demp̣ṭ[i : the phrase occurs only here and in the following line (but is probably to be restored in line 40, see note). We have not been able to find a precise parallel, but several possible interpretations occur to us. We could understand ademp̣ṭi to mean "dead", see TLL I 683.10ff; against this, (a) this usage seems to be almost entirely poetical, (b) the writer uses the verb perierunt to refer to dead birds at line 58 (cf. also line 93 and note) and (c) dead birds are unlikely to have been consumed or taken. A second possibility is to understand it to mean "stolen"; but this does not seem to us very likely. The verb adimere can also mean "to buy" (see OLD s.v.10) and we cannot exclude this (but it is not possible to read aḍẹ[mpti in line 1). The interpretation which we are inclined to prefer, is based on Columella 8.5.9, uerum suburbanis locis, ubi a matre pulli non exiguis pretiis ueneunt, probanda est aestiua educatio, suggesting that young chicks taken from the mother hen are an expensive delicacy. For the use of adimo with animate objects see TLL I 683.34ff.

c.37. At the right, if is correct perhaps we should restore ạ[nseres; if our explanation of adempti is correct (see previous note) there seems no reason why it should not also be applied to goslings and we might have digits after both ạ[nseres and pulli in the next line.

c.39. The beginning of this line is written almost on a level with the writing in the previous line and seems to have been an interlinear addition between lines 38 and 40.

c.40. The use of item suggests that a[dempti should be restored here.

c.41. We perhaps have per followed by a name; c in this hand is usually much more curved, but cf. the note to line 28 above. The scan suggests c̣ommị[, perhaps Commius or Commidus (the latter occurs in Britain, see RIB 1514, near Chesters). Alternatively, perhaps per c̣ommẹ[rcium (and note that OLD s.v. recognises the spelling commircium).

d.43-44. The traces are too abraded to permit a reading. There is perhaps a date in 43 but it is impossible to be sure whether the lines belong with what precedes or with what follows, as part of the heading.

d.45. expensa[: both the neuter and the feminine are attested; TLL V 1647ff. suggests that expensae can be used for items disbursed. A neuter plural would perhaps be easier to explain if this were a sub-heading (see line 1 note).

d.46-47. We can imagine a name at the end of line 46, followed by ad plus a noun, a phrase of the type indicating an occupation or function, see lines 14-15 note.

d.47. There might be a trace of ink after pullus.

d.48. Of the last letter in the line, only a sloping vertical survives, strongly suggesting the left-hand stroke of n; cenaṇ[ti, followed by a short name in the dative case, seems likely, cf. line 64.

d.49. Cf.582.a.ii.2 and note, where consumpti is used. TLL IV 618.19 cites absumo and consumo as synonyms, and we see no reason why they should not both mean "consumed". The latter does not occur in 581.

d.51. nio is likely to be the end of a personal name or a word indicating an occupation or function.

d.52-53. legati: in this text (line 96) and elsewhere in the Vindolanda tablets the term consularis is used of the governor (see Tab.Vindol.II, Index IV s.v.). This suggests that the reference here may be to a legionary legate; see 154.5 and note, where we suggested that it might denote a legionary legate (and cf. M.A.Speidel (1995)). The word also occurs in 660. Various restorations could be envisaged, e.g. c̣[enanti b(ene)f(iciario)]/ legati, or ạ[duentu / legati (cf. line 96).

d.55. The final letter of C̣oriṣ is damaged and the reading might be compromised by the descender of i of legati in 53; but it would be an unusually long one. We are reasonably confident of the place-name, despite our revision of line 98 (see note), but we could equally well read C̣oriạ; perhaps preceded by ad in the previous line (cf. 611.5).

d.57. At the broken edge there is a trace of a sloping vertical which is compatible with ; this suggests ṣ[tabulo, see line 90 and note.

e.back.60. missiọ : there is something odd about the final letter and we have considered the possibility that it is just a medial point; but this would not conform to the pattern elsewhere in this text. missiọ might indicate the celebration of a discharge, but it can have the more general meaning of leaving a place or a post (see TLL VIII 1140.13ff.); or we might conceivably have the end of manumissio (for freedmen in Cerialis' household see 616.A.1 note). We suggest that the nominative case conforms to the pattern found in religious calendars, e.g. the Feriale Cumanum (ILS 108), where some entries consist of a date followed by natalis diui Iuli (vel sim.). The name Flavinus occurs in line 92 and this suggests that it should be restored here, although there are other possibilities, e.g. Flavianus (cf. 172, 261), Flavus (594.b.2); if missio refers to a discharge, it would suggest someone below equestrian officer rank and there is no difficulty in supposing that Flavinus will still have been at or near Vindolanda as a veteran almost a year later (cf. line 22 and note). The name Flavinus does not occur in any of the other tablets. A.R.Birley (2002), 128 suggests that this might be "a ceremony to mark the departure of Flavius Genialis, Cerialis' predecessor as prefect." This depends on the correctness of the placing of e.Back (lines 59-72) at the beginning of the sequence, which we do not think is demonstrable (see introd., above). We also note that there is no actual proof in the tablets that Flavius Genialis was the predecessor of Cerialis (see Introduction, p.00, above). Finally, there is no example of a gentilicium used elsewhere in this account except in the consular date in line 20 (cf. e.g. line 62 where both Niger and Brocchus are mentioned).

e.back.61. The trace at the broken right-hand edge of the tablet is compatible with ; the most obviously attractive restoration is c̣[enantibus (cf. line 64), unless it is too long.

e.back.62. Clearly the writer intended a reference to two persons, Niger and Brocchus; the omission of et is not surprising in a memorandum of this kind (though contrast line 74). Niger and Brocchus are the joint authors of a letter to Flavius Cerialis (248); for Aelius Brocchus see 622 introd., and for the identity of Niger see 295 introd. Brocchus is mentioned three times in this part of the text and also in line 78, over a period of over 7 months. Presumably he was a frequent visitor (as is suggested by several texts in Tab.Vindol.II and see also 622 introd.), and therefore stationed nearby (perhaps Coria (Corbridge) rather than Briga, see 190, 291, 292 (Appendix, ad loc.), 611), rather than a resident. The name Niger occurs again in line 74; if it is the same Niger in both places, his visits are separated by more than 6 months. These entries clearly imply, as one would expect, that Cerialis' household was in the habit of entertaining visiting fellow-officers (cf. 291, 292, 616, 617).

e.back.63. The erasure of Maias and the substitution of Ian[uarias constitutes perhaps the clearest evidence that this text is not a running "day-book" compiled as and when the commodities were received, used or paid for, but an account compiled retrospectively.

e.back.64. ceṇantẹ:; the final letter is better read as with a pronounced serif, not unusual in this hand.

e.back.65. p̣er [: palaeographically this is the easiest reading but we would certainly not exclude c̣er; ṣer, however, is much harder to read . Line 102 suggests that an entry in the form per + name may be acceptable. The fact that the date is 1 January suggests, however, that we might have the name of a consul, as in lines 20-21. The only possibility revealed by the fasti for this period would be Iulius Servianus (cos. AD 102); A.R.Birley (2002), 172-3, n.13 suggests that we should read Ser as the beginning of Ser[uiano, moving this section (lines 59-72) to the beginning of the sequence. For arguments against this see introd., above, and line 60, note. It might be added that in line 20 the name of the consul precedes the date, as we would expect.

e.back.66. There are very abraded traces towards the right.

e.back.70. For stabulum as a pen for poultry see Columella 8.11.3, 17, cf. 8.13 (but cf. line 93 note).

e.back.71-72. The Kalends of March is the date of the Roman festival of the Matronalia and it is tempting to see a reference to it here. The final surviving letter in 72, of which there is a substantial part of a descender curling to the left, can hardly be ; it strongly suggests and we therefore suggest restoring matronaṛ[um. A possible restoration of the whole entry, suggested by the analogy of 190.c.30ff., would be: ḍom[ino (or –ae or -is) ad sacrum (or festum) matrona ṛ[um. For the Matronalia, see Schol.Hor.Carm. 3.8.1, Kalendis Martis matronalia dicebantur, eo quod mariti pro conseruatione coniugi supplicabant, et erat dies proprie festus matronis; Schol.Iuu. 9.53, Kalendis Martis in quibus Iunonis sacra celebrantur a matronis; tunc nam matronalia sunt.

d.back.74. For Niger see line 62 note. The traces after et (if any) are very unclear on the photograph but lae looks clearer on the scan; the most plausible reading is Laet[ or Laet.[. NPEL cites several names beginning with these letters, of which Laetus is well attested in Belgica/Germany. In view of the references to Niger and Brocchus in line 62, this man is also likely to be an officer.

d.back.75. ab[, we remarked that this is one of three or four entries in this second section which might have the formula ab plus a name in the ablative, the others being lines 80, 88, 92 and cf. 97. It is now clear, however, that in none of these cases is this the only possible interpretation of the incomplete entry. Here we now doubt the reading ab, preferring ad; see the next note.

d.back.76. uae: a possible restoration is Miner]/uae, in view of the fact that the calendar of Philocalus records this day as the birthday of Minerva (CIL I2, p. 260); if the last traces in 75 are read as aḍ, we might have something like aḍ [sacrum n(atalem) Miner-/uae (for the abbreviation of natalem, see CIL and cf. 190.32-3, ad sacrum duae.

d.back.79. The reading of what follows the date is very uncertain, but might be read prạṇ[ for which cf. line 91.

d.back.80. In view of the appearance of other fellow-officers Niger and Brocchus (line 62 etc.), this may well be Caecilius September (234 etc.). The traces at the right are too exiguous to hazard a reading.

d.back.82-83. There may be a space of one line between these two lines but there are some marks at the right-hand edge which may be ink. For Sautenus see 182.back 5, and cf. lines 23 (and note), and 89; the phrase apud Sautenum is repeated in line 94. expeṇ[̣sorum: see line 45 and note.

d.back.84. Perhaps reliqụị p̣[ulli n(umero), followed by numerals, with a similar entry for geese in line 85.

d.back.86. This is centred and could be a repeat of the heading expensa (see line 45) but ex ẹ[i]ṣ is preferable, cf. line 109.

c.back.88. Onesimus, is a common name among slaves and freedmen, and occurs in 610.10. The initial o is very large. After the name it would be palaeographically possible to read either aḅ or aḍ (the form of d would be somewhat similar to the form in line 13, diẹ and much like the second example in line 89; the sloping upright is more characteristic of d than b in this hand). The traces just before the break do not admit absu[mpti. If aḍ is to be read, the last trace is compatible with g. There is a word sigma meaning a couch, but if we are to read sig we think ad signa a more probable expression. The whole entry might then take the form vii Idus Maias Onesimo aḍ sig̣[na pullus (or pulli n(umero) .. if there is room). A perfect participle of adsigno is another possibility.

c.back.90. in stabuló: see line 70 note. It is not clear whether the mark at the right is ink or dirt.

c.back.91-92. Presumably we have the participle (or gerundive) of prandeo, referring to the midday meal rather than dinner, see OLD s.v., followed by a name in the dative (if there is room). For Flavinus see line 60 note. Following Flauino, it would be difficult to make sense of ab plus an ablative name, cf. line 75 note, line 97 note, where we suggest abṣụ[mpti.

c.back.93. One possibility is eodem die moṛ[tui sunt, although the writer uses perierunt in line 58; another is moṛ[bidi; a third possibility is to take moṛ as the beginning of a name (e.g. Morinus, a cognomen derived from the tribe in Gallia Belgica, see LC 201), although apud Sauteṇ[um in line 94 makes it difficult to see how to construe this. We have also considered a part of moror, e.g. moṛ[ante/-i (name) ] apud Sauteṇ[um. This in its turn might suggest that stabulum (line 70 etc.) is to be taken in the meaning of "hostelry" (as in e.g. Pliny, Ep. 6.19.4; cf. LS B2); but we think this less likely.

c.back.94. For the phrase apud Sauteṇ[um see line 82. The s at the beginning of the name is noteworthy, an exaggerated example of the form in which the vertical is made as a narrow loop.

c.back.95. For Kalendis followed by the genitive of the month-name, see TLL VII 757.2ff. There is no other example of this in the Vindolanda tablets. This entry is problematical, because the date is out of sequence (unless it is just an error for 1 June). It supports the notion that this account is not a day-book, cf. line 63 and note. At the end of the line the read sịṃ, suggesting that we might have a reference to Similis (cf. Tab.Vindol.II, Index II s.v.) or the word sịṃ[iliter. The scan rather suggests sịṇ, in which case we might consider sịṇ[gularibus, something provided for the singulares who accompanied the governor, cf. 154.5 and note.

c.back.96. For another possible reference to a visit by the governor to Vindolanda see 248.ii.10-11.

c.back.97. abṣụ[mpti is defensible as a reading and we suppose that it can be interpreted as an expression meaning "consumed at lunch", cf. line 49 note and line 92, rather than ab followed by a proper name.

c.back.98. C̣orịs, we are no longer confident that this is the place-name which seems to occur in line 55. f̣orịs looks better on the scan and can be interpreted as meaning "outside the camp". The letters at the end of the line are very abraded and may be obscured by dirt. A personal name would fit the context, and we think that Mỵṛịṇ[o is defensible (see NPEL, s.v.), though r is difficult and there is nothing clear thereafter.

b.back.102. We are not certain of the reading of the centurial sign which is at the edge and has a crack running through it. We assume that the remainder of this line contained the beginning of a word completed in line 103. Surenus is attested in CIL XI 4749 (cf. LE 235).

b.back.103. neṣ,, suggesting the restoration capo-/]nes. We no longer think that s was written; we might simply have in …-/ne.

b.back.104. At the end of the line ṃạ[ is perhaps possible, presumably the beginning of a name or description (like ceruesarius), of which no in line 105 is the end.

b.back.105. There may be some ink below this line at the right, perhaps ẹ[, in which case ̣we might envisage ẹ[x eis. But we think it more likely to be dirt.

b.back106. p̣enes Sautenụṃ: it is not clear how this differs in meaning from apud Sautenum, for which see lines 82 and 94. penes is also used in 159.5, 256.i.3, 582.a.ii.3, 593.ii.1-2, 642.ii.4, 712.b.1.

b.back.107. ribus must be the end of a word which began in the previous line. legionaribus occurs in 180.22 (see the note for the form), but there would not be room for militibus here as well. Another possibility is singula]/ribus, cf. 154.5 and line 95 and note.

b.back.108. There is no mark of abbreviation after s, but s(umma) seems very likely here; s(unt) would be the alternative.

b.back.109-110. ta.ạrụ [, We noted that Columella 8.2.13 refers to Tanagrici, a breed of hen, but thought it far-fetched to suppose we have a bungled attempt at this word here; the traces on the scan are not inconsistent with tana and we now think that Taṇạg̣ṛị[ci could be defended as a reading. After ṛ[el]iqui, of which we remain confident, we now think sṭer..[ likely. If these readings are correct, it would be natural to connect this entry with the statement of Varro, RR 3.9.6, Tanagricos … qui sine dubio sunt pulchri et ad proeliandum inter se maximi idonei, sed ad partus sunt steriliores.

b.back.111. ṣ(umma) is very uncertain; we might have ẹṭ or the marks might not be ink.

Download EpiDoc version using the CC license Creative Commons License and EpiDoc Schema v.5