² By A. P. Garrod of the Gloucester City Museum Excavation Unit (see p. 315). Information from M. J. Watkins of the City Museum. We are grateful to Dr J. P. Wild and Professor J. F. Gilliam for discussing the clothing depicted on the relief and the form of the deceased name with us.

³ Dr Wild compares the cape to that worn by the figure of a deceased soldier recovered from the Camomile Street bastion. RCHM London, Vol. III, pl. 7.

⁴ cf. the remarks by Saddington, D. B. ‘Towards the dating of early inscriptions recording Roman auxiliary regiments’, Akten des VI internationalen Kongresses für griechische und lateinische Epigraphik, München 1972 = Vestigia vol. 17, 538, and note also the omission of voting tribe and origo.Google Scholar

⁵ The name Aurelius is here technically a cognomen. The examples cited by Professor Gilliam, : CIL iii 2712 Ti. Claudius Aurelius, 1st century; ILS 2199 P. Aelius Aurelius, 2nd century (probably); ILS 5787 lulius Aurelius (a centurion), Diocletian, show that although as a nomen its use would normally indicate a third-century date, this may not apply when used as a cognomen, but the examples are too few to draw statistically valid conclusions.Google Scholar

⁶ Analysis of this and the other tablets by Dr A. M. Pollard of the Oxford University Research Laboratory for Archaeology and the History of Art shows that the alloy is 22·7% lead: 77·2% tin (0·1 % copper). The proportion of the metals in the others is: No. 3, 47·4% lead: 52·4% tin (0·1% copper); No. 4, 70·9% lead: 29·1% tin; No. 5, 38·3% lead: 61·2% tin (0·6% copper).

⁷ During excavation for the Bath Archaeological Trust (Britannia xi (1980), 387–8Google Scholar) directed by Professor B. W. Cunliffe, who made them available to RSOT. Other tablets from the deposit have been published in Britannia xii (1981), 370–9, xiii (1982), 396–407, and xiv (1983), 336–41. As before, we are grateful to Mrs S. Pollard for cleaning and conservation, Mr R. L. Wilkins for photography, and Mr I. Marriott for his word and letter-group Index of Curses.Google Scholar

⁸ This method of achieving a ‘secret’ text is found in the unpublished caballarem fragment and in Britannia xiv (1983), 336–9Google Scholar, No. 5 (one line only reversed); cf. also No. 4 below (mirror-image letters as well), RIB 154 (letter-sequence in each word reversed), Britannia xiii (1982), 404–6, No. 7 (entire text reversed).Google Scholar

⁹ (a) Failure to reverse pairs of letters: pure (for puer), a mistake also found in the unpublished caballarem fragment; this mistake was avoided on Side (b), probably because he had just begun the line (3) with a ligatured ER, realized his mistake before completing it, and wrote E on top of it; two lines later (5), he repeated his mistake by beginning to write a ligatured EM, then realized his mistake before completing it, and again wrote E on top of it. These mistakes imply that the primary text was written in a ligatured cursive akin to the writer's own handwriting, and probably was in fact written by him.

(b) Omission of letters: excon(f)ictus (see next note); ba(ro); in a 6–7 he first wrote ipsu(m) aen(u)m and then inserted the missing letters between the lines but slightly out of sequence, perhaps because his eye was caught by the familiar sequence VM which he forgot to reverse. Sangu(in)em may be a ‘Vulgar’ spelling from the primary text, since it is also found in an unpublished curse tablet from Hampshire.

(c) Possible confusion of similar letters: a 5, ut for et would be syntactically neater, but he has written £; b 5, the R of rem is a composite of P and R. The peculiar / of in (a 6) was caused by the writing-point being deflected by surface irregularities.

¹⁰ Commentary a 1. a[e]n[um me]um: the restoration of aenum is guaranteed by its reappearance in 6-7, but that of meum is cramped. levant: the verb levare in the sense of ‘steal’ is never found in Classical Latin and very seldom thereafter, but it occurs in the Lex Visigothorum and once in Gregory of Tours, HF vi. 45, hospiciola pauperum expoliabant, vineas devastabant... levantes pecora vel quicquid invenire potuissent. Like baro (‘man’), therefore, and indeed like the English colloquial usage of ‘lift’ = ‘steal’, it is a ‘Vulgar Latin’ usage which, until the discovery of the Bath tablets, only surfaced in the written language in the post-Classical period.

1–2. [e]xconic[tu]s: if the reading and restoration are correct, the word is unique. Perhaps excon(v)ictus, an intensive form of convictus, was intended; but since this is a curse tablet, excon(f)ictus for exconfixus (‘utterly accursed’) is better: cf. RIB 6, L. Egnatius Tyranus defictus est; Britannia xiv (1983), 338, No. 3, conftgatur.

2–3. templo Sulis dono: a unique instance of the criminal being ‘given’ to the temple, the writer having conflated two formulas, the ‘gift’ of the criminal to the god, and the ‘gift’ of the stolen property to the temple or the god.

3–4. si mulier si baro (etc.): this formulation of mutually exclusive alternatives is common in British curse tablets, but unique to them; for baro (‘man’) see note to levavit above, and cf. Britannia xiv (1983), 340, No. 6, with note ad loc. (ibid., 353).Google Scholar

5. et: the reading is certain, but V and E in this hand are similar and ut would be neater since it would subordinate fundat to dono (‘I give that he may…’), cf. Britannia xiv (1983), 338, No. 3Google Scholar, Basilia donat in templum Martis anilum… ut… configatur. However, a similar use of et is found in Britannia xiii (1982), 403, No. 6Google Scholar, dono numini tuo pecuniam… et… exsigatur, although donatjdonatur… ut exigat is more usual (JRS liii (1963), 123Google Scholar; Britannia x (1979), 343, No. 3). fecerit: the explicit involaverit is more usual; the perfect subjunctive in a relative clause within an indirect command (or similar clause) is a favourite usage in curse tablets from Britain.Google Scholar

5–7. sanguiin)em suum in ipsum aenum fundat: see note 9(6) above; the whole phrase is a variation unique to this occasion of the formula commonly found in British curses, but not elsewhere, of ‘paying’ (solvat) or ‘redeeming’ (redimat) with one's own blood. b 1–3. si mul[ie]r si ba(ro) (etc.): see note to a 3-4 and, for ba(ro), note 9(6) above. 3. The first letter is E written on top of an incomplete ER; see note 9(a) above. 4–5. latr[on]em: a forceful variant of the usual furem (Britannia xiv (1983), 340, No. 6, etc.). The first letter of 5 is £ written on top of an incomplete EM; see note 9(0) above.Google Scholar

5. rem: the first letter has been written as a composite of P and R; see note 9(c) above. 5–7. qui rem ipsam involavi[t] deus [i]nvenia[t]. This can be understood as a relative clause defining eum latronem followed by an indirect command dependent on dono with ellipse or accidental omission of ut (see note to et above), and has been so translated. But it could also be taken as a separate sentence (‘let the god discover the thief), cf. Britannia x (1979), 344–5, No. 4, where we would now restore lines 4–5 as fur quifraudem fecerit deus inveniat.Google Scholar

6. involavi[t]: one would expect involaverit (see note to fecerit), but there is not enough room; for the indicative cf. JRS liii (1963), 123, donatur… quisquis involavit denarios… ut (etc.).Google Scholar

¹¹ The difference between the two 5s raises the question of whether Britivenda and Venibelia were written by different hands. There are minor differences in the other letters common to both names. The two Is of Britivenda are left-sloping or vertical, the V was written with the left stroke first, the horizontal strokes of the reversed E were made from left to right, the N is rather square, the diagonals of the A are almost straight. By contrast, the two Is of Venibelia are vertical or right-sloping, the V was written with the right stroke first, the horizontal strokes of the second reversed E, but not the first, were made from right to left, the N is rather narrow, the diagonals of the A are sinuous. But the question must remain open, since both names are written with reversed Es and there is even a difference in the way the Es were made within Venibelia itself.

¹² For these two ‘Celtic’ names, cf. Brittuenda (Britannia xiii (1982), 397–8, No. 2Google Scholar); and Vendibedis, Sedebelia, Belia, three of the names in Britannia xiv (1983), 339, No. 4; 341, No. 7, two tablets by the same hand.Google Scholar

¹³ For ‘secret’ texts like this see note 8 above. Mirror-image cursive seems to be unparalleled, but Mr A. K. Gregory has sent us a drawing of a curse tablet found by metal-detector in Norfolk, and now lost, which was inscribed in mirror-image capitals.

¹⁴ Being written from right to left and in mirror-image, the letters are unevenly spaced and sometimes distorted: note the LL of puella and in femina how the horizontal strokes of F have been misplaced and the N ligatured to A is indistinguishable from E. The writer began line 6 correctly with INV, but his eye slipped to the second V of involaverit in the text in front of him and caused him to continue with -ERI, before he realized his mistake and wrote -VOLA on top of it.

¹⁵ Commentary 1. Lovernisca: the name is unattested, but cf. Lovernius (V. Nash-Williams, E., The Early Christian Monuments of Wales (1950), No. 89Google Scholar; Britannia xii (1981), 370, No. 5)Google Scholar, Lovernianus, (JRS lix (1969), 239, No. 23), and Lovernac (Nash-Williams, No. 170).Google Scholar

2. d[onat] will fit the space available; for the formula cf. No. 2 above, etc. 2-4. sive v[ir] sive femina sive puer sive puella: this type of formula, common in British curse tablets but unique to them, is always introduced by si; the only previous exception is Britannia xiii (1982), 404–5, No. 7. a seu pair followed by three utrum pairs.Google Scholar

5. qui: repeated in error.

What follows has been damaged by the nail-hole. There is an uninscribed space to the left of the hole, and since the text exhibits word-division, one should probably take sortium as a complete word. (Consortium seems to be the only Latin word whose accusative singular ends in sortium.) What follows qui cannot be the M it superficially resembles (compare the M in eurri), but seems to be a cruder form of the IN A in femina. There is a rare word ina (‘fibre’, esp. of paper), whose plural could be restored here, but the meaning of ina[s] sortium is a matter for conjecture; it must be the object of involaverit and thus the object stolen.

6. See note 14 above. The perfect subjunctive involaverit is so common in more complex curse tablet texts that it seems to have been written here, formulaically, where the indicative might have been expected syntactically.

¹⁶ Commentary 1. [permjittas: the restoration is guaranteed by the formula (see note to 7 ff. of No. 7 below, in note 19). There is a knife-cut here and this is the last word in the line.

2. nee sanita[tem/: see note to 7 ff. of No. 7 below, in note 19; the previous words might be [nee somni]um, cf. Britannia xii (1981), 376–7, No. 8.Google Scholar

3–5. This complex of correlative clauses (tandiu… quandiu, and perhaps also hoc… illud) seems to be unparalleled, but it may have qualified the ill-health wished upon the victim while he retained stolen property, like the unpublished Bath tablet (small find No. 691) which reads:‘… you are not to permit him eyes or health, except blindness and lethargy so long as he shall live, unless…’

6. si vir si femina is easily paralleled, but not the linking of such pairs by et (or etiam).

7. [si libera or libertd] si ancilla: cf. si ancella si liberta, Brandon curse tablet (unpublished); si libera si serva, Bath caballarem fragment (unpublished). The restoration, if complete, would suggest that the tablet was originally about 90 mm wide.

¹⁷ With a metal detector, no further details known. All three were bought by the antiquities dealers Fox & Co. of Yeovil, whose Mr D. Darwish informed RSOT and made them available. They were cleaned by Mrs S. Pollard and photographed by Mr S. C. Renow. For the temple complex at Pagans Hill excavated in 1949–52 by Mr (now Professor) P. Rahtz, see his reports in Proceedings of the Somersetshire Archaeological and Natural History Society xcvi (1951), 112 ff., and ibid., ci-cii (1956/7), 15 ff.Google Scholar

¹⁸ The hand has similarities with, as well as differences from, the hand of a writing-tablet found in a nearby Roman villa at Chew Stoke: see Turner, E. G., ‘A Roman writing tablet from Somerset’, JRS xlvi (1956), 115 ff. Both hands can be dated to the first half of the third century, but hardly later. A closer dating would be desirable, to know what ‘3,000 denarii’ was worth in real terms; and also because excavation dated the whole temple complex to after 258 and probably to the last quarter of the third century (see Proceedings cited in note 17 above, ci (1956/7), 31). However, Antonine coins were found on the site in c. 1830 (ibid., xcvi (1951), 114), and it may be conjectured that this tablet, and probably the next two items as well, belong to an earlier temple.Google Scholar

¹⁹ Commentary 1–2. Almost entirely lost. The ablative is Hi milibus of line 3, though preceded by in not de, suggests that lines 1–2 may have contained a ‘complaint’ or ‘memorandum’ about stolen property to the god from a named person, like two of the Uley curse tablets (Britannia x (1979), 342–3, Nos. 2 and 3). The possible RI of line 1 suggests the restoration [deo sancto Mercu]ri[o], but this can only be conjecture. The dedication of the Pagans Hill temple is not known.Google Scholar

3. in (denari)is (tribus) milibus: the partial abbreviation rios for (dena)rios is found in the Ratcliffe-on-Soar curse tablet (JRS liii (1963), 121 ff.Google Scholar), but seems to be otherwise unparalleled. The sum, presumably the sum stolen, is much larger than the other sums attested by curse tablets from Britain: argentiolos sex (Britannia xii (1981), 371, No. 6); (denarios) (quinque)Google Scholar (Britannia xiii (1982), 403, No. 6); argent[eos] sex (unpublished tablet from Hampshire). However, it is a suspiciously round figure from a time of accelerating inflation, whose real value is unknown.Google Scholar

3–4. [de]mediam partem tibi ut ita ilium [e]xigas: cf. RIB 306, demediam partem donavit No Jenti (which contains the same ‘Vulgar’ spelling of dimidiam); Britannia x (1979), 343, No. 3, deo… tertiam partem donat ita in exsigat istas res. The formula seems to have been only half-understood, since donat (etc.) has been omitted, the order of ita and ut has been reversed, and ilium is apparently an ungrammatical reference to (feminine) partem.Google Scholar

4–5. a Vassicillo [.]pecomini filio: the name and patronymic are unattested, and the division is conjectural (Comini is a plausible patronymic, but entails a Celtic/Greek hybrid ‘Vassicillopsis’ and disregarding the space before P, so that Vassicillo should be preferred). Vassicillus is a plausible diminutive from vassos, cognate with Vassius, Vassillus, etc. The name Comina (C1L v 4258) and the derived gentilicium Cominius (xii 5371, etc.) are known.

5. quoniam: the writer seems to have conflated a causal clause introduced by quoniam and a relative clause introduced by quod, which is why the syntax has broken down.

6. [..]rtussum: an antecedent of quod is required, which excludes any restoration (even if it made sense) on the lines of [tojrtus sum; [pe]rtussum is a possibility, but no more, if one sees it as a variant spelling of percussum, conjecturally derived from the supine of percutio in its sense (Oxford Latin Dictionary s.v. 5) ‘to strike, stamp (coinage)’. Percussum, in the sense of ‘stamped (coinage)’, which is unattested, would at least make sense in this context. de hospitio Lon[…]: cf. Britannia xiv (1983), 339, No. 5, a theft de hospitio suo, the ‘Vulgar Latin’ word for ‘house’. lon[…] is presumably the owner's name (Longinus?).Google Scholar

7. [2-3/alaverint: clearly the perfect subjunctive of a first-declension verb, and presumably meaning ‘steal’, but the obvious candidate involaverint cannot be restored here.

7–11. For the formulas in these lines cf. RIB 306, nollis permittas sanitatem donee perferat usque templum Nodentis; Britannia iii (1972), 365, ne illi permittas bibere nee dormire nee ambulare; ibid., xii (1981), 376–7, N o. 8, nee ei somnum permit tat nee natos nee nascent es donee caracal/am meam ad templum sui numinis pertulerit; ibid., x (1979), 342, No. 2, ut nee ante sanitatem habeant nissi; and No. 5 above, with the unpublished Bath tablet there cited (note 16 above).Google Scholar

8. ma[n]d[u]care: rare in Classical Latin, where edere is the usual word, but a well-known example of the ‘Vulgar’ synonym which passes by speech into the Romance languages.

9. [nee nat]os, although a little short for the space available, is a reasonable restoration (cf. Britannia xii (1981), 376–7, No. 8Google Scholar) since the victims of the curse are a married couple. nessi: this ‘Vulgar’ spelling is also found in Britannia xiv (1983), 340, No. 6. 10Google Scholar. [meant] fits better than [ipsam]. [at]tulerint: cf. Britannia x (1979), 343, No. 3. [per]tulerint is possible but cramped.Google Scholar

The medial stop which follows is a rare if not unique example of punctuation in a curse tablet. (Those in Britannia xiii (1982), 400, No. 4, are of uncertain purpose.)Google Scholar

10–11. iteratis [pre]c[i]bus: the restoration fits the surviving traces; for the phrase cf. Cyprian, , de mortalitate 18 (CSEL iii, p. 308), quid precibus frequenter iteratis rogamus…?Google Scholar

11. [ab ip]sis nominibus: cf. Britannia xii (1981), 371, No. 6, a nominibus infrascriptis deae exactura est, with note ad loc. (on p. 372) that nomen is used ‘in a transferred sense of creditor or debtor’.Google Scholar

12. The antecedent of meorum is lost; perhaps [inimicorum], cf. A. Audollent, Defixionum Tabellae (1904), No. 96, a list of inimicorum nomina; ibid., No. 222, a curse to silence inimicorum meorum linguas.Google Scholar

12–13. The last two lines are badly damaged, and the writing became cramped as space ran out (note the verticality of the VM of meorum). The word that follows hoc (if this is the correct reading) might be pertussum again (see line 6 and note ad loc), if one supposes that the final – VM has become distorted like that of meorum. This can only be conjecture, but it does at least offer a restatement of the original plea (iteratis precibus) for the recovery of stolen coin (?).

²⁰ Commentary 10. octies novem: a magical number, since it is the product of 2 ³ and 3².

11–12. sit omnigen[ere ld]borum fatigatu[s]: cf. Livy 40.22, 15, Philippus omni genere laboris sine ullo effectu fatigatis militibus… in Macedoniam redüt. Is this the first evidence that anyone had read Livy, or indeed any prose literature, in Britain? cf. Barrett, A. A., ‘Knowledge of the Literary Classics in Roman Britain’, Britannia ix (1978), 307 ff.CrossRefGoogle Scholar

²¹ The break has occurred where the tablet was folded, and about a letter's breadth of text has been lost. The uninscribed space below the last surviving line suggests that it was the bottom line of the original text.

Commentary 1. There may be trace of N or A before C. The surviving letters can be divided in several ways.

2. Perhaps quicumque in one of its cases. Quomin[us] implies a verb of prohibition or prevention.

3. frau[d]e sua ul[la], ‘by any fraud of his': cf. Britannia xiii (1982), 404, No. 7, si mihi per fraudem aliquam, where some precaution seems to be implied.Google Scholar

4. donav[i or it], preceded by a personal name, or a dative plural such as quibus or nominibus.

5. The possibilities include [r]eus and [m]eus, but perhaps best is [d]eus, a concluding plea for divine intervention like No. 2 above, qui rem ipsam involavit deus inveniat.

²² Information per A. K. Gregory from Mrs Mansell Wildwood in whose possession it remains.

²³ Excavations as part of the Wasperton Archaeological Project, directed by G. M. Crawford, who supplied full details. For the site, see above p. 297.

²⁴ During excavations directed by Miss V. M. Metcalf Dickinson for the Glamorgan-Gwent Archaeological Trust. For the site see Britannia xii (1981), 317. Information from S. H. Sell who provided photographs and submitted the main fragment, as well as the following item, for inspection.Google Scholar

²⁵ During excavations by E. and A. C. Waddelove, supported by Cheshire County Council Planning Department. For the site see Britannia xii (1981), 333 and ibid, xiv (1983), 299. We are grateful to Mr E. Waddelove for supplying full details and photographs.Google Scholar

²⁶ Excavations supervised by S. Ward for the Grosvenor Museum and the Department of the Environment. Information on this and the following two items from Dr P. Carrington. For the site see Britannia ix (1980), 364–5, and its location ibid, xiv (1983), 298, fig. 8.Google Scholar

²⁷ For the site see above p. 286.

²⁸ We are grateful to Professor G. Rodriguez Almeida for reading the inscription (from a drawing by P. H. Alebon) and for his comments. He points out that the weight indicated, of the order of 200 pounds, is not found on amphorae from Monte Testaccio but on earlier examples. The sherd is perhaps Flavian.

²⁹ See Callender, M. H., Roman Amphorae (1965), 21.Google Scholar

³⁰ Excavations by the Grosvenor Museum Excavation section directed by T. Strickland for Chester City Council and the Department of the Environment. For the site see Britannia xiii (1982), 352–3 and xiv (1983), 298, ‘c’Google Scholar on plan-, fig. 8. The stamp (wrongly called a tile) is illustrated in Current Archaeology Ixxxiv (1982), 12, where the suggestion is made that the impression (sic) was made by a stamp originally intended for butter or bread. The correct reading was given in ibid, lxxxv (1982), 63, in a letter by Professor E. Birley.Google Scholar

³¹ Deposited by the farmer, Mr J. C. Myott, on loan to the Salt Museum, Weaver Hall, Northwich. Full information and casts of the inscriptions from D. F. Petch.

³² The name Cunitus, though spelt with double T, is matched on a defixio from Leintwardine, JRS xlix (1969), 241, No. 31. His associate probably had a Greek cognomen, such as Erastus or Eros and was presumably an imperial freedman. Cf. Ti Claudius divi I. Erastus (ILS 1671), 77 Claudius Aug. lib. Eros (ILS 2914) and 77. Claudius Eros Aug. lib. (ILS 7859).Google Scholar

³³ During archaeological investigations by the Peakland and Buxton archaeological societies, conducted by D. Bramwell and K. W. Dalton. For the site see Derbyshire Archaeological Journal ciii (1983), 47–74, and especially pp. 63–5 for an illustration and discussion of the text. We are grateful to Dr Bramwell for providing full information and submitting the object for inspection.Google Scholar

³¹ Excavations for the Trent Valley Archaeological Research Committee and the Department of the Environment directed by Hazel Wheeler. Information on this and the following item from Miss R. Birss.

³⁵ During the excavations noted in Britannia x (1979), 285. Miss L. Allason-Jones noticed the inscription and sent a photograph with a note by Dr F. Jenkins.Google Scholar

³⁶ Dr Jenkins (see previous note) reads , which is a more plausible name, but seems to fit what is there less well. The name Versius (but cf. Versinia, CIL xi 1620) is apparently attested only as that of a secretary of Sertorius in Spain (Sallust, cited by Servius, Comm. on Aeneid I 698).

³⁷ Excavation for the Wimborne Archaeological Group directed by R. Tanner and subsequently A. G. Giles. Information and photograph from Lt.-Col. G. E. Gray. For the site, with plan, see Britannia xiii (1982), 386–7Google Scholar, fig. 23, and for other epigraphic material, Britannia xii (1981), 369 No. 4, 383 No. 31 and xiii (1982), 411 No. 38.Google Scholar

³⁸ See Davey, N. and Ling, R., Wall Painting in Roman Britain, Britannia Monograph No. 3 (1981), 165–8. (No. 38) with PLS. LXXVIII, LXXIX an d cxi.Google Scholar

³⁹ This slogan possibly occurs as a graffito on wall plaster from Exeter, Britannia ix (1978), 476 No. 23(a).Google Scholar

⁴⁰ Excavations directed by Philip Crummy for the Colchester Archaeological Trust. Information and drawings of this and the following item from Nina Crummy.

⁴¹ There are no examples in TLL where flamma carries the double sense that brand can have in English, so it may be better to treat Flamma as a proper name, cf. the names of horses attested on North African curse tablets, e.g. Igneus, , ‘Fiery’Google Scholar; Fulgidus, , ‘Lightning’Google Scholar; Nitidus, , ‘Flashing’; or Scintilla, ‘Spark’Google Scholar, see Audollent, A., Defixionum Tabellae, 456–60Google Scholar, Index III 2. Flamma is in fact attested as a personal name, see Kajanto, I., The Latin Cognomina, 341Google Scholar. The Celtic name Senovarus/a is attested twice in Roman Britain, see Britannia xiii (1982), 398, No. 4 ( + 401, n. 8), BathGoogle Scholar, and Britannia xiv (1983), 350, Addendac, 3, Caves Inn, Warwicks.Google Scholar

⁴² By Denis Gamble of the Engineers' Department of Colchester Borough Council. Information from P. Crummy. For the site see Crummy, P., Colchester Archaeological Report 3 (1984).Google Scholar

⁴³ Excavation by the W. Essex Arch. Group directed by Dr N. E. France and Miss B. M. Gobel. Information and photographs from R. Bartlett. For the site see JRS liii (1963), 138.Google Scholar

⁴⁴ Excavations by Hereford and Worcester Archaeology Department for Hereford and Worcester County Council, the Department of the Environment and the Manpower Services Commission, directed by John Sawle, who provided details of this and the following item and submitted the sherds for inspection.

⁴⁵ Vascella from vascellum, an alternative for the more common vasculum, the diminutive of vas, the form being attested epigraphically o n a funerary inscription from Rome, CIL vi 3428 of a.d. 214, where it refers to an ash chest. The word is probably neuter plural (rather than first declension feminine singular), but with a singular meaning, cf. vasa, the neuter plural of vas which can be so used (OxfordLatin Dictionary s.v. vas 2. Both vas and vas(,s)a (the latter used in the sense of cinerary urn) are attested from Roma n Britain (JRS xliv (1954), 107, No. 25Google Scholar and EE ix 1351(a)). Note, however, that in medieval Latin vascella can be both neuter plural and feminine singular, see Niermeyer, J. F., Mediae Latinitatis Lexicon MinusGoogle Scholar s.v. Mr Sawle draws our attention to Hooke, D., ‘The Droitwich Salt Industry: an examination of the West Midlands charter evidence’ in Brown, D. et al. , Anglo-Saxon Studies in Archaeology and History, BAR British Series 92 (1981), 123–69, where it is shown, pp. 136–7, that vasculum was used in the medieval period to describe vessels holding brine and it is interesting to not e the anticipation of this usage here.Google Scholar

⁴⁶ Excavations for the Canterbury Archaeological Trust directed by T. Tatton-Brown. See Britannia viii (1977), 423.Google Scholar

⁴⁷ Excavations by the Springhead Excavation Grou p for the Department of the Environment, directed by S. R. Harker. Information on this an d the following item from Debora h French, who submitted the objects for inspection.

⁴⁸ For Roman hackamores, see Taylor, A. K., ‘Römische Hackamoren und Kappzaume aus Metall’, Jahrbuch des romisch-germanischen Zentralmuseum 22 (1975) = Festschrift H. J. Hundt pt. 2, 106–33. Pl. 59.2 shows a complete unprovenanced example of similar form in the British Museum carrying a (different) stamp on the side piece above the loop (cf. ibid., 128).Google Scholar

⁴⁹ Information, drawing and photograph from Tim Ambrose. The ring remains in the possession of the finder.

⁵⁰ Guildhall Museum Cat. (ed. 2, 1908), 43, No. 307. For a pottery ink-pot from Canon Street, London, and now in the British Museum, see JRS xlii (1952), 108, No. 34Google Scholar. B.M. Guide to Roman Britain (1951), 31, fig. 15. RPW.Google Scholar

⁵¹ Such amphorae are known as ‘pseudo-Koan’. For Koan and pseudo-Koan amphorae see, e.g. Grace, V. R., Excavations of the Athenian Agora, Picture Book No. 6, Princeton 1961, figs. 56 and 61 and accompanying text. Pseudo-Koan amphorae have been found at Pompeii thus giving a terminus ante quern of a.d. 79 for their manufacture, and were of local Italian make, as well as being produced elsewhere.Google Scholar

⁵² Watching brief by G. Milne. Information and drawing from Mrs B. J. Davies and Beth Richardson.

⁵³ For this term see note 51 above.

⁵⁴ Excavation for the Department of Greater London Archaeology (Southwark and Lambeth) of the Museum of London, directed by Derek Seeley with support from Eagle Star Properties Ltd. and the Greater London Council. For the site see above p. 308. W e are grateful to Professor Rodriguez Almeida for assistance with the reading.

⁵⁵ Antipolis, Antibes is known t o have produced muria, Pliny, N.H. xxxi, 43, cf. Martial, Epig. xiii, 103. which shows that muria made of tunny from Antipolis was cheap compared to sauces (also from Antipolis?) mad e from mackerel.

⁵⁶ Information from D. W. Blandford of Trinity School Park, Crbydon, where the sherd now is.

⁵⁷ Information on this and the following two items from A. K. Gregory.

⁵⁸ By Michael de Bootman in whose possession it remains.

⁵⁹ During excavation for the York Excavation Group directed by Mrs E. M. King, who provided a drawing.

⁶⁰ The c could be G, and the Man N or even a cursive R. Personal names beginning Ott… are most uncommon, but cf. CIL iii 3817 (Pannonia), Tertius Ottonis f(lius).

⁶¹ The graffito was noticed by Dr D. Williams. The two sherds were made available by Dr B. Heywood for the Yor k Minster Archaeology Office.

⁶² The stroke above the v seems to relate to the L. There is no trace of any letter after v, but the surface is damaged. For Bitucus cf. RIB 108, but this is not the only possibility. For Candi cf. RIB 1653, centuria Candiana, implying a personal name Cand(i)us.

⁶³ R. E. Birley in 1973 submitted this an d the next counter. For other finds in 1973 see Britannia v (1974), 467Google Scholar. For a list of comparable counters see JRS xxxii (1942), 117, No. 11. RPW.Google Scholar

⁶⁴ Gibson, , Arch. Ael. xxiv (1902), 40, fig. 2, with ‘1898’ inked on the sherd. Now in the Newcastle Museum of Antiquities and drawn by RPW in 1961.Google Scholar

⁶⁵ ‘Right’ and ‘left’ refer to the figure as seen in impression and not as cut in reverse on the stone itself.

⁶⁶ See Ant. J. lxii pt. 2 (1982), ‘Exhibits at Ballots’, 380–1, 386–7.Google Scholar

⁶⁷ During excavation directed for the Department of the Environment by Dr G. Webster, who made it available to RSOT.

⁶⁸ Found by Mr Alan Harrison of Winterton in whose possession it remains. Information and photographs for this and the following item from Kevin Leahy.

⁶⁹ Found by Mr Alan Harrison of Winterton, in whose possession it remains.

⁷⁰ Excavations by the Norfolk Archaeological Unit. Information of this and the following two items from A. K. Gregory.

⁷¹ Excavations by the Staines Archaeological Unit for the London and Middlesex Archaeological Society with the support of the Department of the Environment, directed by K. Crouch, who supplied details and submitted the sherd for inspection.

⁷² The bottoms of the letters are missing and it is just possible that the first letter is a B. Similarly, there may be stops between the letters, but since these are on the break they may simply be where small flakes of the surface have been detached.

⁷³ It came to the Museum of Antiquities, University of Newcastle, from the local collection formed by Bell, T. J.. Allason-Jones, L. and Miket, R., Catalogue of small finds from South Shields Roman Fort (1984), 3.331Google Scholar. The inscription was revealed by cleaning, and was brought to our notice by Miss Allason-Jones, who sent a drawing and details. It is possible that this spoon was one of a pair of which the corresponding member was inscribed VTERE, as suggested for inscribed spoons from Canterbury and Colchester, to make the slogan VTERE FELIX : see Britannia xiv (1983), 350, (e).Google Scholar

⁷¹ During excavation for the Department of the Environment and the Society of Antiquaries of Newcastle upon Tyne directed by Mr Daniels, C. M. (see Britannia xi (1980), 355 f.Google Scholar). Miss L. Allason-Jones made this and the next three items available in the Museum of Antiquities, University of Newcastle. For another stamped tile of the Sixth Legion from Wallsend, see Britannia vii (1976), 389, No. 50.Google Scholar

⁷⁵ The Fourth Cohort of Lingones garrisoned Wallsend during the third and fourth centuries: RIB 1299, 1300, 1301; Not. Dig. Occ. 40.33.

⁷⁶ During excavation for the Manpower Services Commission and the University of Newcastle directed by Mr N. Neil. Miss L. Allason-Jones made this and the next item available in the Museum of Antiquities, University of Newcastle.

⁷⁷ For other examples of alphabets and other ‘writing practice’ on bricks and tiles before firing, see McWhirr, A. (ed.), Roman Brick and Tile (1979), 239.Google Scholar

⁷⁸ Directed by S. E. Winbolt, whose report on the site is in Sussex Arch. Collections lxxiii (1932), 13–32. Information and rubbings from D. R. Rudling.Google Scholar

⁷⁹ Excavations by the Wiltshire Archaeological Unit at Trowbridge directed by Kate Carless who provided details and submitted the object for inspection.

⁸⁰ Curie, J., A Roman Frontier Post and its People (1911), 168–70Google Scholar, an d PL. xxi = Garbsch, J., Romische Paraderustungen (1978), 57, H 1. Curie does not note a n inscription, although he notes the one o n the bronze ‘parade helmet’ (ibid., 167, Uffi t(urma) Ges(sH)). The inscriptions were noted by M r I. Scott an d brought to our notice by Dr J. Close-Brooks; Mr T. Cowie made it available at the National Museum of Antiquities of Scotland.Google Scholar

⁸¹ There is a space between the two xs; the numerals(?) seem to have been cut this way up. Th e trooper's nam e is badly worn; the name of his troop seems to be an adjective (the usual form would be -iana, an d on e might read Atinia(na)) derived from the name of the last decurion (Atinius, At(t)ius)), the comman d being temporarily vacant. Garbsch (see previous note) publishes a number of ownership inscriptions including F 1 (Theilenhofen), a helmet one of whose three successive owners is tiurmd) Patercliana Attonis.

⁸² Anderson, , PSAS xxxv (1899–1900), 393 with neither reading nor figure. Drawn in 1948 by RPW in the National Museum of Antiquities, Edinburgh.Google Scholar

⁸³ Macdonald, and Curie, , PSAS lxiii (1928–1929), 548, fig. 114, item 4. Now in the National Museum of Antiquities, Edinburgh, where in 1982 their illustrator drew this on request and provided detailed notes. IICA E Macdonald and Curie; TILCA E RPW. The ala Tungrorum was stationed at Mumrills, see RIB 2140. Ilca seems to be unmatched. It may have been derived by syncope from the place-name Ilici (Hispania Tarraconensis), or later there was the personal name Ilica (ILCV4129, Rome), RPW.Google Scholar

⁸⁴ During excavation for the Scottish Field School of Archaeology directed by Professor S. S. Frere. A drawing by T. Borthwick and a full description by Dr D. J. Breeze were provided by Professor S. S. Frere, who notes that this is only the third lead ingot to be found in Scotland: see Stuart, R., Caledonia Romana (2nd ed., 1852), 206 (Bertha) and 323 (Kirkintilloch), both now lost.Google Scholar

⁸⁵ During the excavations directed by Professor J. J. Wilkes. This and the next item were made available by Professor S. S. Frere.

⁸⁶ The exaggerated height of the I suggests it was the initial letter; there may have been a space after the last letter.

⁸⁷ T. A. Acton, who excavated the legionary depot at Holt from 1908 to 1915, found in 1912 a tegula reading:…]PREF LEG XX[…He sent it to Haverfield, who included it in EEix 1275a just before that fascicule went to press. This object, labelled in ink ‘Hilly Field’ was still retained among the epigraphic material which R. G. Collingwood transferred to RPW in March 1941. In 1947 RPW gave it to Sir Cyril Fox for the Holt Collection, since 1925 in the National Museum of Wales, Cardiff. A second broken tegula, with no provenance recorded on it, was included in this Haverfield material. The graffito described above has needed some elucidation; presumably Haverfield put it on one side while putting his material for EE ix into final order, and subsequently overlooked it. It has now been sent to Cardiff for the Holt Collection.

⁸⁸ Teg(u)la occurs on two other tile graffiti, EE vii 1144b (South Shields) and JRS xxxvii (1947), 182, No. 15 (Willowford Bridge, Cumbria, east of Milecastle 49, Harrow's Scar), RPW.Google Scholar

⁸⁹ Excavations by the Glamorgan-Gwent Archaeological Trust, directed by Metcalf, V. M.. For the site see Britannia xii (1981), 317. Information from S. H. Sell, who submitted the tile for inspection.Google Scholar

⁹⁰ Information from the Curator, Mrs E. Tyson.

⁹¹ Information from Miss L. Allason-Jones.