Vindolanda Tablets Online Tablets Exhibition Reference Help

Volume One (TVI) material for tablet 190 (TVI tablet number 4)


Publication: Bowman, Britannia v (1974), pp.360-73,

References: Bowman-Thomas, VWT, p.26, pl. XIII; Bowman-Thomas, Historia 24 (1975), pp.466, 476-7; Bowman-Thomas-Wright, Britannia v (1974), p.477; Thomas, Scriptorium 30 (1976), pp.41-2; Wright, Proc.XIVth Int. Congress of Papyrologists, p.355; Birley, Vindolanda, pp.154-5, pls. 79-80; Bowman, ZPE 18 (1975), pp. 237-52; J. N. Adams, BICS 22 (1975), p.23; Turner, RV, pp.51-3


This multi-leaved set of tablets is the longest single document in the collection. The series contained at least two full diptychs and a half of a diptych at the end; but there were probably more if we may judge by the number of minor fragments. We have conjecturally assigned two of these fragments to the beginning of the text, as we have it, but the others cannot be placed. If the half-diptych which contains lines 42-5 is the end of the document, as we think, then these other fragments must all be assigned to the earlier part of the text. We think that these diptychs were tied together in a kind of concertina format (see Vol. 1., Ch. 2). The diptychs were scored and folded with the writing on the inner faces and the text was written across the grain of the wood and parallel to the short edge of the leaf. This practice is also noticeable in other accounts (e. g. No. 5 (191), or CIL 32, p.953, No. xv, pag. posterior) and has obvious advantages for records of this kind in offering the use of long, narrow columns.

The sequence of the leaves in the document had to be established on the basis of (a) the way in which the various fragments were found in the ground and (b) the sequence of dates in the text itself. Beginning with the latter, we had no difficulty in establishing the order of the pieces which contain lines 10-45 and run from xiii Kal. Iulias through to vii Kal. lulias. The fragment containing lines 7-9 was found adhering face to face with that containing lines 13-19, whilst that containing lines 10-13 adhered to a blank piece. Since the join at line 13 could be made with certainty, the relative position of the other two fragments could also be established as forming the top half of a diptych (with the lower part of this half being blank, possibly because line 9 marks the end of one series of accounts).

We are confident that only one hand is at work in lines 10-45, but we are uncertain whether the same hand wrote lines 1-9. Both parts of the text are in similar hands - competent, business-like cursives with no pretensions to elegance - but the idiosyncratic form of r which occurs in lines 12, 16, 26, 32, 34, 37 and 41 is not found in lines 1-9 (nor elsewhere in the tablets except in No. 7.3 which may well be written by the hand which wrote No. 4.10-45 (190)). In fact, this section has its own rather odd form of r in sacrum (line 7). But since so little survives in lines 7-9 it is unsafe to suppose that a different hand is necessarily at work. Palaeographically noteworthy are the forms of d (sometimes almost like a modern c as in hordei, line 21); the three forms of u (contrast uini in line 25 with the same word in line 40, and the more usual form of u, for example in line 10); n, which sometimes has a very long cross-stroke beginning far to the left, e.g. uini, line 25 (contrast the form in axungiae, line 35); i, which often has a pronounced serif to the right at the top, or even only part way up the hasta (e.g. uini, line 25, domino, line 36); and in general the forms of h, b, g and m. There is no other hand in the collection which closely resembles this (apart from that of No. 7, mentioned above (199)). Its cursiveness is emphasised by the number of ligatures, especially between e and a following letter (ei in hordei, passim, er and es in ceruesae, line 12 etc.) and between r and a preceding letter (er, as noted above, pr in priuatum, line 32 etc., or in hordei, passim and, most interestingly, ur in muriae, line 33, on which see the note).

Lines 1-9 of the text record deposits of money or payments ad sacrum. The actual amounts, recorded in denarii, are fairly small if our interpretation of the numerals is correct. We can think of only two plausible explanations of ad sacrum in a military context. It might refer to the unit's strong-box which was kept in the sacellum (see Webster, RIA2, pp.189-90) and to which individuals might have made small deposits (for deposita in pay records see RMR 68.iii.10 and for contributions ad signa, interpreted as payments towards a burial fund or for the cult of the signa, where the word sacrum might find a context, RMR 68.ii.19). But we cannot parallel this use of sacrum, and lines 38-9 where one modius of wine is credited ad sacrum would be particularly hard to explain (cf. also line 5). Therefore we think it on the whole more likely that we are dealing with items set aside for sacrificial or religious purposes. We know, for instance, that extraordinary deductions were made from pay for special meals at the saturnalia (see RMR 68.ii.8 and iii.7 and, for ritual aspects of such banquets to which the word sacrum might be appropriate, Josephus, BJ 7.16); and there are other features of this document which suggest a religious connection (see note to lines 38-9). We might also note here the observation of Seneca, NQ 4a.2.7 on Egypt, although the context is not specifically military: in haec ora stipem sacerdotes et aurea dona praefecti, cum sollemne uenit sacrum, iaciunt. The nature of the rest of the account leads us to assume that this section was preceded by a date-heading and it is noteworthy that payments ad sacrum are recorded in at least three consecutive lines (7-9).

Lines 10-45 contain accounts of foodstuffs. The entries are arranged by date, beginning with xiii Kal. Iulias and include various kinds of food and drink reckoned in modii and sextarii. It seems obvious that we have here a kind of day-book recording either receipts or disbursements of foodstuffs over a period of several days (for a conjecture at the existence of such records see P.Mich.VII, p.32 and note the statement of Vegetius quoted Vol. 1., Ch. 3). A good parallel for the form of the account exists in RMR 82 (=P.Ryl.223); it contains entries under consecutive date-headings of small quantities of materiel relevant to shipbuilding or maintenance. The small quantities involved suggest that the supplies were being expended rather than acquired, and this is supported by the fact that the document specifies the use to which the materials are to be put.

There are several reasons for regarding most of the items in the Vindolanda account as disbursements rather than receipts. First, the amounts recorded are not very large and it seems unlikely that supplies would be taken in on consecutive days in such small quantities. Second, the time of year is inappropriate for the intake of commodities such as barley from harvest, though this cannot be regarded as conclusive since it might have been purchased or brought in from depots outside the camp. Third, and more forceful, is the point that the phrases attached to some items (per priuatum, ad sacrum, ad stipes) are much easier to explain if the supplies are regarded as outgoing rather than incoming (see below). Finally, we may note the occurrence of the word allatus in line 22, a line which has caused us considerable difficulty; the reading of this word is unimpeachable and it must, we think, indicate something brought in (probably wine); and if this is so it is natural to suppose that the other items are outgoing.

We consider the account, then, as a record of (a) contributions in cash and kind for some religious purpose and (b) the disbursement (with one item excepted) of food-supplies to military personnel over a period of a few days at the end of June of an unspecifiable year. We must now consider briefly the significance of the qualifying phrases mentioned above:

per priuatum (lines 18?, 32, 34, 37, 41, see also No. 7.3 (199)). The explanation offered in the editio princeps, that these were items purchased privately by soldiers in addition to their regular rations, now seems to us less certain. It is true that this practice is attested (R. W. Davies, Britannia ii (1971), p.124 n. 15, perhaps also in the Pselkis ostraka, SB 6955-76, which need not necessarily be receipts for regular rations); but we should now prefer an explanation which allows per to mean something more like ‘through the agency of’. One possibility, suggested to us by Professor Eric Birley, is to take Priuatus as a personal name (Kajanto, LC, p.315); thus the items would be issued through a person, perhaps a slave, named Priuatus. An alternative, along the same line, is to take priuatus as meaning a civilian as opposed to a soldier (see OLD s.v. 2c). We can find no parallel for our original interpretation which implied a meaning like ‘through a private transaction’. Whatever the correct explanation, we still think it likely that the phrase would denote a disbursement outside regular rations. In each case the phrase refers to the item preceding it (as lines 40-41 prove).

ad stipes (line 36). The full context is domino ad stipes per priuatum, referring to axungia. We understand this as an issue of a commodity to be used for charitable purposes, though we are now less certain that domino indicates that it was made to the commanding officer (see note to line 36). For stipes see the quotation from Seneca (above).

These accounts have an important bearing on our knowledge of the food-supply of the Roman army and the military diet. The information they offer is to be put in the context of the very full discussion by R.W. Davies, Britannia ii (1971), pp.122-42. As far as the actual commodities are concerned, the list does not provide any great surprises, apart from the fact that there is no mention of frumentum (it does occur in No. 5.9 (191)) which was certainly the staple item of the Roman military diet. The commodity most frequently mentioned is in fact hordeum (barley), which was rarely eaten by the soldiers except in peculiar circumstances or as a punishment, though it could be used as an ingredient in various dishes (see line 11 note). It is possible that the soldiers were reduced to eating it at the end of the season as a substitute for frumentum (for a similar phenomenon see IG xi.2.158.46-7, a reference which we owe to Dr. D. M. Lewis). Of course, its normal use was as animal fodder; and if we are right in supposing that Vindolanda was occupied, for at least part of the period covered by the deposit, by a cohors equitata (see No. 1 introductory comments (155)) there is no difficulty in explaining the repeated occurrence of hordeum (which, it should be noted, is never qualified by the phrase per priuatum). The other commodities mentioned are muria (fish-sauce), axungia (pork-fat), uinum, acetum (sour wine) and ceruesa (Celtic beer). The amounts of vintage wine (uinum) are in fact quite large compared with the small quantity of acetum. For a fuller picture of the variety of foods in use we should add that the account in No. 5 (191) also contains (in addition to frumentum) condimenta, sal, bracis (a cereal) and various kinds of meat, caprea, porcellum, perna and ceruina. The latter may not all be part of the standard diet of the regular soldiers in peace time (see No. 5 introductory comments (191)) and we might suspect that the items in No. 4(151), apart from hordeum, were not standard rations either.

As we have said, there is some indication in the account of a religious connection. Another curious feature, perhaps linked to this, is the fact that with one exception, no single date-heading has more than three items. The exception is viii Kal. Iulias, June 24, for which there are eight separate entries, which include the less common commodities and all the certain occurrences of the phrase per priuatum. This day was the festival of Fors Fortuna (see CIL 12, p.320, I.Ital. 13.2, p.473, RE VIL1, 16 ff.) and might therefore be a day on which extra food-supplies were needed and additional items purchased (note that the Dacian account discussed in No. 5 introductory comments (191)is dated to 1 May, which led the editor to suggest a possible connection with the Larales). It is also appropriate for the items designated ad stipes and ad sacrum d<i>uae (?). Whether these items were intended for ordinary soldiers or for officers remains uncertain. At any rate, the whole operation which these accounts represent was presumably part of the regular administration by the cibariatores (see Vol. 1., Ch. 3) and the horreorum librarii (R. W. Davies, Britannia ii (1971), p.136, n.96).

Changes from the editio princeps are discussed in the line notes. New photographs have enabled us to improve the text at several points, but it should be noted that we have adhered to some readings based on the original photographs which are not evident on the plates (e. g. because a fragment has become detached in the interim).



1. We have omitted to note the two tiny fragments which are placed before this line in the editio princeps. We think this placement is correct but the traces are so exiguous that we could not even hazard a guess at what they are supposed to represent. As the bent fibres at the edges indicate, they belong to the bottom edge of one half of a diptych and the top edge of the other. Lines 1-3 are on a fragment which is from the top right-hand corner of a leaf. At the top left is a cut which was probably part of a V-shaped notch.

2. In line 2 perhaps X (editio princeps). We still think this is the easiest reading; but as no trace of a numeral survives after it we should perhaps read s(emis) instead, assuming that the apparent cross-stroke is not in fact ink. The trace surviving in line 3 may be the same.

5. 'm(odii) ii (denarii) s(emis)' (editio princeps). There is a break between the measure and the numeral which could accommodate one digit. m is certain and apparently has over it a flat dash like that which surmounts it elsewhere in this text (line 15 etc. ), where it certainly stands for m(odius). If it stands for m(odius) in this line, it is noteworthy that here and here only (but see note to line 43) it is followed by an amount in denarii. It was suggested in the editio princeps that the cash sum might represent the cost of the item preceding (on prices of commodities see in general R.P. Duncan Jones, Chiron 6 (1976), pp.241-62), but it is perhaps more likely that the account is simply recording a credit of a commodity and a cash sum. Throughout the text we have resolved the measures in the nominative case, following the name of the commodity which is in the genitive (see our remarks in No. 5 introductory comments (191)and in the notes to No. 4.16-18 and 22 (190)). In the case of fractions, as here, we suppose that denarii is a genitive followed by s(emis) in the nominative. The symbol following s(emis) might be the sign for asses (cf. R. Cagnat, Cours d'epigraphie latine 4 (1914), p.472).

7. On the meaning of this notation see introductory comments.

8. Perhaps ].i a<d> sacrum' (editio princeps). The apparent long downstroke on the first letter is almost certainly not ink; a is difficult but we are now reasonably confident that d was written. The word or phrase preceding ad might be the same as in line 9 but we cannot guess what it might have been.

9. ]atam: only the long tail of the first letter survives. a is preferable to r, since r is not elsewhere written with a long descender in these lines (nor in the account as a whole except in the rather special ligature in muriae, line 33, on which see the note).

10-13. About half the width of the complete column is missing at the right.

11. hordeum is the commodity most frequently mentioned in the account, occurring seven times. For possible explanations of this phenomenon see introductory comments, p. 86. For its consumption by soldiers see Vegetius, 1.13, Suetonius, Aug.24, Polybius 6.38.3 and cf. Galen (Kühn) vi.507, R. W. Davies, Britannia ii (1971), p.123. It could be used in cooking polenta for example (see André, ACR, pp.52, 57, 63-4). For barley and its collection cf. P.Dura introduction, p.41 and J.F. Gilliam, YCS 11 (1950), pp.243-4.

12. ceruesa[e: the more common forms are ceruesia and ceruisia but ceruesa is found elsewhere (see TLL, OLD s.v. and the remarks by J.N. Adams, BICS 22 (1975), p.23). ceruesa is Celtic beer which is known to have been in use in the army, see R.J. Forbes, Studies in Ancient Technology iii2 (1965), p.130 and Davies, op.cit. (note to line 11), p.133 who cites evidence for grain-malting in its manufacture. André, ACR, p.180 says that the cereal basis is not known, but the references given by Lauffer, DP, ad 2.11 show that it could be made from frumentum or bracis (for the latter see No. 5.16 and note (191)).

14. modius is throughout abbreviated as m; normally this is surmounted by a horizontal dash but here and in line 21 it is omitted, no doubt through inadvertence, cf. ChLA 436 and P.Mich.449 where it appears as m with superscript o (cf. No. 6.2 note )205)). It was stated in the editio princeps that the modius in use here is likely to be the castrensis modius. This was in standard use in the later Empire, but in the first century A.D. it is more likely to be the basic Italic modius of approximately 8.6 litres. The castrensis modius is 1.5 times this capacity according to R. P. Duncan Jones, ZPE 21 (1976), pp.43-52 and 53-62. The loss at the end of this line might have removed a quantity of sextarii.

15. The modius is normally used for dry goods, and it is rather odd that it should be used in this account for wine and beer as well as the other foodstuffs. We have not been able to find another document in which we can parallel this use. We can adduce only Pliny, NH 23.88 on walnut oil, si quid in nucleo putridi fuerit totus modius deperit, Sidonius, Ep.2.2.8 on the capacity of a baptisterium (cf. TLL s.v. 1242.57 ff.) and the late reference from Ionae uita Columbani quoted in No. 5.16 note (191).

16-18. There are difficulties in restoring these lines. There is no doubt about the date in view of lines 13 and 20. Normally this stands alone in the line, but here the writer has gone on to fill the right-hand side of the leaf. We might be tempted to suppose that this entry was of a different type from most of the others and we could restore: hordei [m(odium) i / allat]um ad hor[reum / per pri]uatum. The two difficulties are first, that the trace before m in line 17 does not suit u very well and second, that we would expect a nominative rather than an accusative case here, especially if our interpretation of line 22 is anywhere near the mark. But we have no other suggestion to offer. Note also that the r at the end of line 17 is far from certain.

20. There is a large gap between u and l of Iulias where there is a crack in the leaf which the writer had to avoid. The a in Iulias has been remade from what looks like s.

22. 'allatus ua..rasse.[' (editio princeps). The improved photograph confirms allatus, and we are fairly confident that it was followed by uin and that the letters immediately before the break are ssec. Very hesitantly, we suggest that we should understand m(odius) to have been lost at the right (to agree with allatus) and that uini Massec[i preceded. This is not easy palaeographically, since we probably have to assume that the writer first put uini mssec[ and then wrote in the a by means of a long descender from the final part of m. There are two further difficulties: (1) the spelling should be Massici but it hardly seems possible to read the letter after ss as anything other than e; (2) we are unsure whether such a high-class product would be likely to be found at Vindolanda (on Massic wine cf. Pliny, NH 14.64). This must be an incoming item (see note to lines 16-18) and as a parallel we note CIL 4.1239, cf. Add., p.205, faenu(m) al(l)atu(m) viii Idus Octo(bres); cf. perhaps also ChLA 475.

25. uini: the final i has an emphatic serif to the right and the middle i also has a tick half-way up the vertical. The sextarius sign is paralleled on the Carvoran corn-measure of the same period (see F.J. Haverfield, Arch. Aeliana (third series) 13 (1916), p.84). It also occurs in P.Mich.449.10 (not recognised by the editor) and in an account from Dacia (see No. 5 introductory comments (191)). R. Cagnat, Cours d'epigraphie latine4 (1914) gives this (p.461) and another symbol (p.473).

31. acetum is a low-quality wine which could be mixed with water to make a drink called posca, see R.W. Davies, Britannia ii (1971), p.124. It also appears in the Dacian account (see No. 5 introductory comments (191)) where the price is half a denarius per sextarius.

32. per priuatum has been written in afterwards and squashed in between two lines. On the interpretation of the phrase see introductory comments.

33. muriae: there is a remarkable ligature in ur (perhaps better described as a conjoint letter) in which the descender of r is prolonged to some distance below the line, a feature of r not found elsewhere in this hand. The writing of ur in this way reminds us of the manner in which these letters are occasionally joined in mediaeval manuscripts (a phenomenon which is attested much earlier as well). For muria (fish-sauce) see Davies, op.cit. (note to line 31), p.131 and cf. condimen[torum, No. 5.3 (191)).

34. ’per <p>riuatum' (editio princeps). The second p runs into the tail of r in the previous line, but the new photograph shows that it was not omitted as was assumed.

35. axungia is pork-fat, for which there were many uses. See Pliny, NH 28.136, usus axungiae ad emollienda, excalfacienda, discutienda purgandaque (cf. Galen (Kühn) xii.346) and the references given by André, ACR, p.186, Lauffer DP, ad 4.11. It appears on the Palmyrene Tariff, CIS II.iii.3913.49 (Greek) and in CIL 4.2070 (Pompeii), iix Id. Iulias axungia also the following note.

36. domino: this was taken in the editio princeps to refer to the commander of the unit at Vindolanda, but we are now very unsure of this since we cannot parallel the usage at all. The word is, of course, frequently used in military circles in addresses both to equals and superiors (for the third-person usage see No. 37.6 (225)). We have no alternative suggestion to offer and its meaning in line 44 is even more obscure (see note).

ad stipes: this means alms or charitable contributions and it can have a religious significance as is made clear by Varro, LL 5.182, id apparet, quod ut tum institutum etiam nunc diis cum thesauris asses dant stipem dicunt and by Seneca, NQ 4x.2.7 (quoted above). The meaning of the whole entry in lines 35-7 might be that the pork-fat was credited to the dominus, through a Priuatus(?), to be given away as a charitable distribution or dole (to local civilians?). The festival of Fors Fortuna might be a suitable occasion for the giving of charity. Or perhaps, as was suggested in the editio princeps, the commodity had some role in the performance of a sacrifice. For special needs on festival days cf. CPL 305.5, quia in die festo mi opus est.

39. '' (editio princeps). The correction to d<i>uae was suggested to us by M. Robert Marichal (by letter). It would presumably refer to a goddess and would make good sense with sacrum preceding it; it might then, however, mean ‘shrine’ rather than ‘festival’ (see the dedication of a shrine in CIL 12 801, sei deo sei deiuae sac(rum) and cf. ILS 2986). However, diua = goddess is scarcely ever found except in poetry and would be somewhat unexpected in a document of this kind. We have also considered the possibility that duae is an error for deae (for the spelling cf. CIL 8.11700, Africa, duis - but that might rather represent diis), but this would be strange.

42-5. The new readings result from a much improved photograph. For our conjecture that this was the last leaf in the series see Vol. 1., Ch. 2.

43. Apparent traces between hordei and f may not be ink. If they are, we should no doubt read them as m(odius). There is a mark at the edge of the leaf which looks like a denarius sign, but we think it is unlikely to be ink. The leaf is certainly complete at this edge and there would not be room for a numeral to follow.

44-5. We can make no sense of this though the reading of most of it seems clear. If runt is a verbal ending, as it surely must be, it would follow that domini is masculine plural (but we do not know who these ‘lords’ are, see note to line 36). man[se/runt is a possible reading but we cannot explain it satisfactorily. Is brigae a place-name? If so, surely not the place in Hampshire (see Rivet-Smith, PNRB, p.227 f.). Finally, we are puzzled by the fact that all this seems to have no evident connection with the rest of the text.

46. Not Iu]lias.

47. Possibly ]s.