² During excavation by a Norfolk Archaeological Unit Manpower Services Commission excavation team. Full information from David Gurney who also supplied photographs, radiographs and a drawing. For inscribed plaster from the same site, see below, No. 67.

³ For the name Atticianus see RIB 1199. Kajanto cites 22 examples in CIL.

⁴ During excavation directed by Dr Ann Ellison for the Committee for Rescue Archaeology in Avon, Gloucestershire and Somerset. For an interim report see Rodwell, W. (ed.), Temples, Churches and Religion in Roman Britain (1980), 305–25.Google Scholar Three of the inscribed lead tablets found were published by us in Britannia x (1979), 340–5, from electrotype copies. The originals have now all been deposited on loan from the British Museum in the Ashmolean Museum, Oxford, where they are being examined by RSOT for publication, either in Britannia or in a separate corpus like the Bath tablets. This is the first to be examined by him. For making the tablets available in Oxford, and for providing enlarged photographs as an aid to drawing, our thanks are due to Dr Ian Longworth, Dr Tim Potter, Catherine Johns, and Dr Stephen Johnson; in Oxford our thanks are due to David Brown and Dr John Rea. We are also grateful to Dr Ellison (now Ann Woodward) for details of the excavation, and other assistance.

⁵ The hand is similar to several found among the Bath curse tablets (e.g. Tab. Sulis 8). It is characterised by distinct serifs, sometimes so elaborate as to make for confusion between I and E, and I and T; and also by a downstroke found in E and I, and occasionally in other letters, which in mid-course bends to the right. The redundant initial downstroke of the second N in 9 and of M in u suggests that the scribe also wrote a New Roman Cursive, and thus that the tablet should be dated to the fourth century; it is true that the cursive 5 in 10 is of Old Roman Cursive form, but this is found in some New Roman Cursive texts from Bath. The second Q and both Ds of quidquid (2) are also of cursive form. C in 7 and 8 includes a vestigial second stroke, but elsewhere is made with a single movement of the stilus. There is one ligature, presumably prompted by the cursive hand, GI in 7, although G could be read alone. There seems to be some word-division, marked by an occasional space between words.

⁶ For detailed analysis see next note. There are three copying mistakes, pe(r)d(id)it (3), MAIET (4), NESSA (9–10); and the use of quidquid (2) may be due to mechanical copying of a formulary. Ne taceat may have been omitted after ne loquatur (5); there is certainly an omission after de (13), and probably of habeat in 9 and of a subject (‘the thief’) for the string of verbs. The formula si vir si mascel (3–4) is confused, and like the malapropism co(n)scientia (12–13) raises the question of whether the scribe understood what he was writing. Yet his hand is practised and implies a good standard of literacy. His language also contains traces of the ‘Vulgar’, spoken Latin, in nessi (9–10, 14), co(n)scientia (12–13), and in the inept use of in (10) and de (13). By implication, therefore, he was familiar with Latin both written and spoken; but he was handling formulas he only half understood, or which perhaps were too familiar to retain much meaning.

⁷ Commentary

(The Bath curse tablets are cited by Tab. Sulis number as published by RSOT in B.W. Cunliffe (ed.), The Temple of Sulis Minerva at Bath, 11: the Finds from the Sacred Spring (1988); the published tablets from Uley are cited as ‘Uley’ and number as in Britannia x (1979), 340–5, Nos. 2–4.)

1. Biccus: ‘Celtic’ personal name, already attested in CIL xiii 5366a (Vesontio). Probably a variant of Beccus (xii 2514) or Becco (xii 5381, Tolosa), the childhood cognomen of M. Antonius Primus of Tolosa, meaning ‘beak’ (Suetonius, Vitellius 18). Beccus (‘beak’) is one of the few Celtic loan-words to be taken into Romance (Elcock, W.D., The Romance Languages (1975), 196).Google Scholar

dat: British curse tablets regularly ‘give’ the stolen property or the thief himself to the god, although donare is the usual verb.

1–2. Mercurio: see below, 10–11.

2. quidquid: the object ‘lost’ is often specified (e.g. Tab. Sulis 34, dono numini tuo pecuniam quam [a]misi, id est (denarios) (quinque)), although not always. It is possible that the scribe was mechanically transcribing a formulary (transcription errors are noted below), or that he meant to write quicumque: see below, 4–5.

3. pe(r)d(id)it: a transcription error. The verb is common in the Bath tablets, and is also garbled in Tab. Sulis 6, EERIDID (perdidi), and 103, PEDRE (perdere).

3–4. si vir si mascel: the seribe has confused two variants of the ‘whether man or woman’ formula common in British Curse tablets, si vir si femina and si mascel si femina.

4–5. ne meiat ne cacet: the anagram MAIET is another transcription error. This is the first occurrence of the formula, but a variant is found in Tab. Sulis 41, nec adsellare nec [meiere]. The subject of the verbs (and ne loquatur, etc.) is not explicit, but must be the thief; either this must-be understood, or fur, qui involavit (etc.) has been omitted in error; alternatively, quidquid (2) is an error for quicumque, perdere (‘cause loss to’) being used in the sense of involare (‘steal’) as in Tab. Sulis 99.

5–6. ne loquatur: the first occurrence of this formula in a British curse tablet, but cf. Audollent, Defixionum Tabellae No. 270 (Hadrumetum), non dormiat neque sedeat neque loquatur. RIB 7 (London) also silences its victim, but the context is unclear. Since the other formulas in this Uley tablet fall into pairs, it is possible that ne taceat has been lost, perhaps because it was confused with ne cacet.

6–7. ne dormiat n[e] vigilet: for this formula cf. Audollent No. 270 (previous note) and Britannia iii (1972), 363 (Wanborough), ne Uli permittas bibere nec [esse nec vigilare nec do|rmire nec ambulare. Denial of somnum (‘sleep’), often coupled with sanitatem (‘health’), is frequent in the Bath tablets.

7–8. nee salutem nec sanitatem: understand habeat, unless this was omitted in error, or the scribe's mechanical grasp of Latin was so weak that he thought this pair of nouns was another pair of verbs. The two nouns are coupled in Britannia xvii (1986), 431, No. 2 (Eccles).

9–10. ness[i]: the scribe wrote NESSA, no doubt by confusion with the initial letters of nee salutem and nee sanitatem just before; the first S seems to be, not a corrected C, but a confusion between cursive and capital forms; the second S is cursive. The ‘Vulgar’ spelling nessi (for nisi) is also found in Tab. Sulis 65; nissi is found in Uley Nos. 2 and 3.

9–12. ness[i] … pertuierit: this formulation is common in British curse tablets: the curse is not to be lifted unless (or until) the stolen goods are returned, pertuierit or attulerit being the usual verbs and tense. The tense is probably future perfect.

10–11. in tempio Mercurii: the god's temple is often specified as the place to which stolen goods are to be returned, as in Uley No. 3, nissi quando res s(upra)dictas adfanum s(upra)dictum attulerit. This is the first explicit mention of the templum Mercurii at Uley, but it has already been deduced from the previously published tablets and other finds, which include figurines of Mercury and an inscribed altar base (Britannia xii (1981). 370, No. 5).

in templo: templo alone (dative) would be correct, but the uncertain use of prepositions, and a preference for simple prepositions like in and de instead of the Classical case-ending, are typical of ‘Vulgar’ Latin. Cf. Tab. Sulis 97, Basilia donat in templum Martis anilum argenteum.

12–13. ne co(n)scientiam: perhaps an error for the less abrupt nec co(n)scientiam. The assimilation of medial ns to s in co(n)scientiam is typical of spoken Latin, and occurred very early, e.g. in the epitaph (ILS 2) of L. Cornelius Scipio, COSOL (consul), CESOR (censor): see E. Pulgram, Italic, Latin, Italian, 600 B.C. to A.D. 1260: Texts and Commentaries (1978), 180. An instance from Britain is probably RIB 876, Co(n)st[ant…].

co(n)scientiam: the context demands the sense ‘forgiveness’, ‘pardon’, which is not found in Latin conscientia, which means variously ‘complicity’, ‘consciousness’, ‘conscience’. The writer may have been groping for the very rare and post-Classical ignoscentia (‘forgiveness’), a synonym of the Classical venia or indulgentia (see TLL s.v.).

13. de: hoc (‘this’), furto (‘(his) theft’) or similar has been omitted in error. The use of de here, instead of the Classical genitive, is typical of ‘Vulgar’ Latin.

14–16. ness[i] me intercedente: the context requires the sense ‘unless I intervene’, but the syntax is certainly not Classical, and does not seem to be ‘Vulgar’ either. It might be saved by supposing the loss of a final verb (e.g. deus placetur), on the lines of Cod. lust. x 34.2, nisi … ordinis intercedente decreto … causa probetur (‘unless permission is given by decision of the town council’); much more likely, however, the writer did not know how to construct a conditional clause of his own, and may indeed have confused nessi (nisi) with sine or even non.

intercedente: the sense of ‘intercede’, although familiar in English, seems to have been introduced by Christian writers (see TLL s.v. intercedere, 2156, 47ff.); here it is probably being used in a more strictly legal sense, to ‘intervene’ or ‘interpose’. This is the verb's first occurrence in a curse tablet; they often contain quasi-legal language.

⁸ The tablet was offered for sale by B.A. Seaby Ltd in their Coin and Medal Bulletin, August 1983, p. 221 (V245). In appearance it resembles tablets found at Uley. The similarity of its formulas to those found in British curse tablets, especially at Bath (cited as Tab. Sulis), guarantees a provenance in southern Britain. Mr P.A. Clayton made it available to RSOT.

⁹ Commentary

The restoration of m[aie]statem (2–3) determines the width of the missing portion, which is confirmed by the likely restoration of si [puella] (4). One line of text, in addition to virtually all of 1, is likely to have been lost at the top.

2. The first stroke of the initial A, like the second stroke of the initial S in 3, seems to have disappeared in a patch of damage.

amisi: cf. Tab. Sulis 34, deae Suli Minervae Docca. dono numini tuo pecuniam quam [… a]misi, which suggests the likely form of the two lost lines.

2–3. oro tuam m[aie]|statem: some letters are damaged and the third T must be seen as a rare form occasionally found in Tab. Sulis, in which the cross-bar is made in two strokes, each ligatured with an adjacent letter; but the reading and restoration are supported by Tab. Sulis 35, rogo sanctissimam maiestatem tuam ut vindices …. and Audollent, Defixionum Tabellae No. 122 (Emerita), per tuam maiestatem te rogo oro obsecro uti vindices quot mihi furti factum est.

3. ut <f>furem: uli could be read, but this form of ut has not yet been found in British curse tablets; the long descender, the possible trace of a diagonal stroke, and the likelihood that a second has been lost in the area of damage, all suggest that F was written twice in error. For furem cf. Tab. Sulis 65, donavi furem (etc.).

4. Pairs of mutually exclusive alternatives introduced by si … si are typical of British curse tablets, including si puer si puella (in that order). Si puella cannot be restored at the beginning of 4, whose second word must end in … ila. This, and what survives of the first two letters, suggests si a[nc]il(l)a; cf. Tab. Sulis 52, […] si ancilla, and an unpublished tablet from Brandon (Norfolk), si aricela (sic) si liberta. If the original width of the tablet has been correctly calculated (see above), si libera must have been omitted in error.

5. ext[i]nguas: the G is a form occasionally found in Tab. Sulis, made in one continuous stroke without a second cross-stroke; the verb does not occur in any other British curse tablet, but cf. Tab. Sulis 54, consumas.

Surviving traces suggest the next two letters are CR, but there is no word cruiillis. nor any easy emendation. -VT- inevitably suggests the conjunction ut (as in 3), perhaps preceded by (i)ta; its force would be to qualify extinguas (‘destroy him, so long as the property is returned’), ILLIS, since the antecedent furem istum is singular, should probably be understood as illi s[…].

5–6. s[c.4] |cias: if ut is correctly understood in 5, this would be another second-person singular subjunctive like extinguas (5); s[uspic]cias seems to be the only verb that fits the space, but it yields no useful sense in this context.

6. perduci: the reading is not certain; the passive infinitive is hard to explain, but one would expect some reference to the recovery or return of stolen property, cf. Tab. Sulis 64, pallium [re]ducat.

If perduci is correctly read, the next (lost) letter must begin a new word, which can only be [r]em, since -EMRA- is an impossible sequence of letters. Res (plur.) is frequent in British curse tablets in the sense of ‘(stolen) property’, and rem is found in Tab. Sulis 44. Ra[ptam] is a reasonable restoration, in view of Britannia x (1979), 342, No. 2 (Uley), (iumentum) quod erapium est … quod rapuerunl.

7. Too damaged for restoration.

¹⁰ By a schoolboy, B. Daniel, and presented by him to the Roman Legionary Museum at Caerleon. Mr J.D. Zienkiewicz provided information and photographs of this and the following item. The complete tombstone would have resembled RIB 365, perhaps from the same cemetery.

¹¹ Cemenelum, modern Cimiez outside Nice, was capital of the small province Alpes Maritimae. It was enrolled in the Claudia voting-tribe. It is already attested as the origin of legionaries on the Rhine or in Britain, the latter being two who were seconded for service as frumentarii at Rome, where they died: L. Gratius L. f. Clau. Verinus Cemeneli of Leg. II Aug. (CIL vi 3339 = ILS 2364) and T. Aurelius Cl. Certus Cemeneli of Leg. XXV. V. (CIL vi 32872). The restoration offered here has been prompted by a third tombstone, that of a retired legionary of Leg. II Aug., which was found in the territory of Cemenelum: Maturius Fuscus emeritus ex legione II Aug(usta) missus honesta missione. See Octobon, F.-C., Lamboglia, N., ‘La “pierre écrite” d'Ascros (Alpes-Maritimes)’, Rivista di Studi Liguri xxv (1959), 221–31.Google Scholar (The version in AE 1961.169, which derives from Gallia xviii (1960), 321–2, is inaccurate.) Maturius is a manufactured nomen well attested in CIL xiii and borne by another legionary (of Leg. VIII Aug.) from Cemenelum (C. Laguerre, Fouilles de Cemenelum II, Inscriptions (1975), No. 46); it probably incorporates the Celtic name-element matu- (Ellis Evans, D., Gaulish Personal Names (1967). 228–30).Google Scholar The coincidence of what remains of the Caerleon legionary's nomen and cognomen with those of an emeritus from the same legion and origin makes the restoration of Maturius Fuscus attractive; it would assume that the Caerleon legionary followed his father, grandfather or close kinsman, into the same legion. But the stone is too fragmentary for the restoration to be certain; Valerius, a common nomen well attested at Cemenelum, and Tuscus could also be restored. The formulation also presents problems. The abbreviation of beneficiarius by B cut by a horizontal stroke has not been found in Britain, but occurs elsewhere (see Dessau, ILS iii, p. 756); there is another example of beneficiarius legati in RIB 505 (cf. RIB 293). This is presumably the post held by the deceased, preceded by a numeral (of the years of service), but it is unusual for it to follow the reference to the heir. It is also unusual for the years of service to be separated from the age at death. However, beneficiarius legati cannot be applied to the heir. both because it is not usual to specify his post, and because it is preceded by a numeral, RSOT.

¹² By Mr C. Williams. It is now in the Roman Legionary Museum at Caerleon.

¹³ While it is possible that the nomen of the deceased was Varius, a longer name such as Ianuarius seems on grounds of spacing more likely. If, as is also possible, there was no space between the R and T, then the name was almost certainly Quartus or cognates.

¹⁴ During excavation directed by Miss V. M. Metcalf for the Glamorgan-Gwent Archaeological Trust. Information from Mr S. H. Sell. For the site see Britannia xii (1981), 317, with plan p. 318, fig. 4. Fragments (a). (b), (c) and (e) were destroyed in a fire at the Trust's offices in 13th March 1983. Fragments (d) and (f) remain in the possession of the owner of the site, Mr I. Kegie.

¹⁵ By two visitors to the shrines, Mr and Mrs James, who informed Denis Coggins of the Bowes Museum. Mr Coggins removed it to the museum, and on a subsequent visit with Mr K. J. Fairless recovered a fragment of another buff sandstone altar. This was only the capital and extreme R. corner of the die, 0·30 by 0·35 by 0·24 m. with trace of an incised groove separating capital and die. No text survives. The two altars have been given to the Bowes Museum by the landowner, Mrs Colling.

¹⁶ For the dedication to Vinotonus Silvanus and the format cf. RIB 732 and 733. This is the first instance from Britain of the title Augustus applied to Silvanus, but it is common elsewhere (see Dessau, ILS iii, p. 551). The only other possible restoration of the prefect's nomen is Urbius which is, however, much less common than Orbius. The cognomen is almost certainly derived from primus or priscus; Primanus is too short and Priscianus too long, which leaves Primianus and Priscinus. There is a worn space before N in 5, and no trace of a letter. The spacing of 5 suits [MIA] better than [SCI], but neither can be excluded. The prefect seems to be otherwise unknown.

¹⁷ See the Carlisle Journal for 17 September 1814: ‘A Roman inscription, of which the following is a fac-simile [FIG. 3], was this week found amongst the rubbish on the line of the road now making from the Irish-Gate to the Court-Houses:- several other stones, bearing Roman inscriptions, were disposed of by the workmen before we could have an opportunity of examining them.’ The reference was noticed by Mr D. R. Perriam. and communicated to us by Ian Caruana.

¹⁸ It is not clear whether either margin was preserved, but it is likely that the right-hand margin was, the final letter of line i being an 5 of cursive form. Line 2 would have identified the deceased within the legion (miles, etc.); his age, years of service, and the name of his executor would have followed. Maritus is an uncommon cognomen. and it is possible that Martius should have been read.

¹⁹ During excavation for the Tyne and Wear Museums Service directed by Messrs P. T. Bidwell and S. C. Speak. Paul Bidwell sent full details, with a squeeze and photograph.

²⁰ During ploughing by Mr A. N. Greenwell of Greenwell Ford, Lanchester, in whose possession the stone remains. John Casey informed us, and Mr Greenwell made the stone available.

²¹ The stone is published with full commentary by RSOT in Zeitschrift für Papyrologie und Epigraphik, forthcoming. The dedication to Jupiter is implied by the formula c(ui) p(raeest), which is found on altars dedicated officially by military units and their commanding officer, usually to Jupiter; this is confirmed by the wheel-symbol, which on inscribed altars from Britain (RIB 827, 1877, 1981, 1983) is confined to Jupiter: see Green, M. J., The Wheel as a Cult-Symbol in the Romano-Celtic World (1984), 345–7.Google Scholar The unit responsible is clearly the Coh. I Lingonum, already attested at Lanchester: see RIB 1091, 1092 (both of A.D. 238/44), 1075. It was equitata (RIB 1276). The spacing of the second line to survive suggests that there was a leaf stop before and after [E]Q. The name of the prefect is […]VLVIVS […]IX. The cognomen can only be Felix, thus proving that the previous line has lost two letters. Allowing one letter for the praenomen, the nomen can only be Fulvius of Mulvius; it is long odds on Fulvius, since there are c. 385 instances of Fulvius (and Fulvia) to c. 26 of Mulvius in Mócsy, Nomenclator, and CIL vi and viii. There are nine instances of ‘Fulvius Felix’ in CIL vi, mostly freedmen, and none of them an eques; otherwise the name occurs only twice in CIL, in a list of names at Fulginiae (xi 5228) and at Mustis in Africa Proconsularis (viii 1601). The latter was a native of Mustis, since he was enrolled in its tribe Cornelia, but his tombstone makes no mention of an equestrian military command. However, it proves that the name was current at Mustis, if nowhere else in Africa, and supports the identification of the prefect Fulvius Felix with the unknown Fulvius of CIL viii 1578 (Mustis). This is a dedication, probably to Elagabalus (A.D. 218/22) and his mother, by a lady and her two equestrian sons: the elder son is L(ucius) Fulvius Kastus Ful[vianus fisci advocatus patrimoni tr]actus Karl(haginensis) et Galliae Narbo[nensis); the name of the younger son is lost, but he held a post […] Britanniae inferior(is). The restoration of the elder son's post is due to H.-G. Pflaum (Carrières, No. 311), who notes that the description of the younger son's post implies an equestrian military command in Lower Britain (cf. ILS 2762 and CIL viii 5180 with 17266). This would seem to have been the prefecture at Lanchester, thus dating the altar to the period A.D. 218/22, and proving that RIB 1091 and 1092 should not be taken as evidence of a reoccupation of Lanchester in A.D. 238/44.

²² During excavation for the Joint Management Committee of the Department of the Environment and the Vindolanda Trust, directed by Mr P. T. Bidwell; see Britannia xii (1981), 323. Robin Birley made it available.

²³ The first letter, exceptionally, is a cursive Q; the three letters are presumably the initials of tria nomina, something unusual on a building stone, but for other possible examples see RIB 300, 301, and Britannia ix (1978), 474–5, Nos. 10 and 11. RIB 467 is a centurial stone which also carries someone's three initials; cf. also RIB 710 (tombstone), 952, 1421 (religious dedications). For the abbreviation of tria nomina in general, see Wiseman, T. P., ‘Tile-stamps and Roman nomenclature’, in McWhirr, A. (ed.), Roman Brick and Tile (1979), 221–30.Google Scholar The second letter(s) of line 2 is more difficult. Unless it is a symbol (and it does not much resemble a phallus), it may be best understood as IL ligatured, a second L (for legionis) being omitted by haplography. For building stones of the Twentieth Legion at Vindolanda, see RIB 1708 and Britannia xii (1981), 380, No. 14.

²⁴ By the Vindolanda Trust. Robin Birley made the stone available.

²⁵ It is most unusual for the legion's cognomen to be omitted. For another instance, see Britannia viii (1977), 432, No. 24.

²⁶ By the farmer, Mr Baxter. Information from Tony Wilmott, Project Director, Birdoswald Central Excavation Unit, who made the stone available.

²⁷ By Alan Whitworth, in the course of his measured drawing of the Wall. Mr Whitworth pointed them out to RSOT, and provided photographs and details of location. He has also detected three facing stones bearing recessed panels as if for a centurial inscription, but with no text now visible, (a) East of Birdoswald, south face, 78·4 m west of west wall of Milecastle 49, stone 0·32 by 0·20 m. (b) West of Birdoswald, north face, 52·5 m west of the fort (measured from west side of drive to the farmhouse), stone 0·42 by 0·13 m. (c) West of Birdoswald, south face, 32·2 m east of the east wall of Turret 49b, stone 0·30 by 0·18 m.

²⁸ See Britannia vii (1976), 381, Nos. 11 and 12, with notes. For other examples of incised X see RIB 1372, 1382, 1518, JRS I (1960), 237, No. 12(b).

²⁹ A stone bearing an incised X has already been recorded from this sector (JRS 1 (1960), 237, No. 12(b)), south face, 123 m west of the west wall of Milecastle 49.

³⁰ The frame thus resembles RIB 1518, save that only three sides are rectangular; the fourth side (of the rectangle) consists of two incised lines like half an X.

³¹ By Mr R. J. Chapman of Sandy, Beds., to whom it was returned after being the subject of a coroner's inquest. Information and drawing of the inscription from Mr C. Going.

³² The reading is certain. The writer may have been influenced, consciously or unconsciously, by the common Christian tag, vivas in deo.

³³ Information from Ian Caruana, who sent a copy of a report by Dr D. A. Welsby. Both items are now in Tullie House Museum, Carlisle.

³⁴ During excavation by the Carlisle Archaeological Unit directed by Ian Caruana, who provided a drawing and full details.

³⁵ The L. and upper edges of the impression are lost, but there was probably a one-line text within a rectangular frame. The restoration is uncertain; possibly the abbreviated title of one of the cohorts of Lingones.

³⁶ By Ian Caruana, while observing the renewal of a sewer. Mr Caruana made it available with a photograph and other details.

³⁷ The obverse die is identical with that published from Street, Castle, Britannia xvii (1986), 439Google Scholar , No. 24, which can now be restored as Obv. P[S] | COS. Another version is found in the next item. For the abbreviation of the unit's title, cf. CIL iii 12633 (Apulum), tile stamps of the pedites singulares consularis of Upper Dacia, which include (f) P(edites) Si(n)g(ulares) Co(n)s(ularis) and (g) P(edites) S(ingulares) C(onsularis). The unit is the infantry guard of the governor of Britain, based at the Cripplegate Fort in London. In general see M.P. Speidel, Guards of the Roman Armies (1978). This is the first unequivocal epigraphic evidence of the singulares in Britain; for the question of whether SC on five British inscriptions refers to s(ingularis) c(onsularis) or s(ummus) c(urator), see Speidel, 126. None of the sealings is stratified, but it is unlikely that the unit was the guards of the third-century legate of Lower Britain at York: he was not strictly a consularis, and the singulares of the governors of praetorian provinces do not describe themselves as consularis. (There are only two possible exceptions: see Speidel, No. 13 (= ILS 2525) and 29 (= CIL iii 14693).)

³⁸ By Ian Caruana, while observing the renewal of a sewer. Mr Caruana made it available, and provided details.

³⁹ See note to previous item.

⁴⁰ During excavation by Carlisle Archaeological Unit on behalf of the Department of the Environment and Carlisle City Council directed by Mr. J. A. Dacre, Mr. T. G. Padley, who noted the discovery in Rescue News 31 (Autumn 1983), 8, made them available and discussed them with us.

⁴¹ They will be published more fully by RSOT in the final report edited by Mr T. G. Padley.

⁴² This is the earliest instance of both place-names (see A.L.F. Rivet and Colin Smith, PNRB, s.v.), written within ten years of the Roman army's arrival at Trimontium (Newstead). Two stilus writing-tablets of identical type, and ten stili, were found there (J. Curie, Newstead, 308). Luguvalio (Carlisle) is found, correctly spelt, in Tab. Vind. 22, 9, a few years later. The ‘Vulgar’ loss of intervocalic v may be due to the writer's unfamiliarity with the name. Iulius Martialis is quite a common name, and the recipient cannot be identified; he is hardly likely to have been the centurion at Mainz (A.D. 43/69) or the Praetorian tribune (A.D. 69), see PIR ² s.v. 1410. It is not clear why he is addressed ‘at Newstead or Carlisle’: perhaps his legion had detailed men for service at both, and his correspondent did not know which, or he may have travelled regularly between them, possibly in the course of the enlargement and remodelling of Newstead in the late 80s.

⁴³ Terlius is a common Latin cognomen; dominus is used in the Vindolanda Tablets between equals (21, 30), as well as to a superior (37). The reading and meaning of bra2–3ario is discussed in the final report. Clearly it is the rank or function of Tertius. Two possible restorations are *bractiario and *braceario, the former better suited to the surviving traces, but not entirely satisfactory; it would be a variant of brattiarius (‘a worker in gold leaf’). *Braceario might conjecturally be derived from bracis, a Celtic loan-word found in Tab. Vind. 5, 16 (‘a kind of cereal which could be used in malting’), and would mean ‘maltster’ or ‘brewer’. For Lugu(v)alio see previous note.

⁴⁴ The epistolary formula dabis (‘you are to give’) may be found on Tab. Vind. 107, and is common on the Vindonissa stilus writing-tablets (see Speidel, M. A. in Jahresbericht der Gesellschaft pro Vindonissa 1986, 52).Google Scholar Michael A. Speidel has kindly discussed them with us. The name lulius Macrinus could be read on Tab. Vind. 4, B3 (see note ad loc), but the identification is unlikely; the name is common, and none of the instances in CIL is likely to have been a soldier in Britain.

⁴⁵ Sedatus is a common Latin cognomen, but a nomen Kimius seems to be unattested (but cf. the cognomen Cimio, CIL iii 7975); it might be a variant of Gimius (cf. CIL xii 4980) or, more likely, cognate with the Etruscan place-name Ciminius (see W. Schulze, Eigennamen, 566, n. 3).

⁴⁶ The mis-spelling Britania is very rare in the Roman period, Britannia or Brittania being the usual forms. It is found after the Roman period in the Ravenna Cosmography and as a MS variant (see Rivet and Smith, PNRB, 49–102), but only two contemporary, epigraphic, instances are noted by TLL. As might be expected, they are not of British provenance: CIL iii 5300 (Noricum) and viii 4800 (Numidia), the veteran who served in Britain Brauniaco (i.e. at Bravoniacum). This tablet was surely posted from another province.

⁴⁷ During excavation by the Carlisle Archaeological Unit directed by Ian Caruana, who provided squeezes and other details.

⁴⁸ The numerals are cut across the grain of the finished surface, and by their position were added after the timbers had been felled and cut to shape. However, since some of them occur on off-cuts which formed no part of the final structure, and have even been sawn through, they cannot have been intended as an aid to assembly. For a similar incised numeral on an oak plank from Chesterholm, see below. No. 72.

⁴⁹ With the next six items by Ian Caruana, while observing the excavation of foundations for a block of flats on the site of the demolished Miles Mclnnes Hall. Mr Caruana provided drawings and other details.

⁵⁰ This is the first epigraphic evidence of the ala Petriana at Stanwix, whose presence there (cf. Not. Dig. Occ. xi.45) has never been doubted. Sealings are sometimes found at the fort where the unit named is otherwise attested: e.g. Britannia xviii (1987), 372, No. 28.

⁵¹ This is a Latin praenomen, but is also a common ‘Celtic’ personal name.

⁵² Tertius is a common Latin cognomen, but it is occasionally found (in CIL xii and xiii) as a nomen. Vi[…] would then be a cognomen (Vitalis?).

⁵³ What may be extraneous scratches confuse the issue. Rufi (etc.) cannot be read.

⁵⁴ By Mr W. Caygill, using a metal detector. Information from Ian Caruana, who provided a drawing, and Colin Richardson.

⁵⁵ Other sealings of Leg. VI are known (e.g. Cumb. Westm. n.s. xxxvi (1936), 120), but this sealing is interesting because the reverse is almost identical with that of the only other sealing recorded from Kirkby Thore (ibid., the provenance need no longer be doubted). This has since been lost, but reads: Obv. CIB. Rev. OPTA |TID✹ The D has been taken to refer to a decurion, and thus to imply that the abbreviated cohort on the obverse was equitata (i.e. Coh. I Batavorum) (see Birley, E. in Arch. Ael. 4th ser. xiv (1937), 239Google Scholar , n. 30), but the centurial sign now associated with Optati makes this argument untenable. It is unusual for the centurial sign to follow the name, but this does occur on sealings (see Cumb. Westm. n.s. xxxvi (1936), 119, Nos. 17, 19–21; 121, No. 8). It would seem, therefore, that D is the centurion's abbreviated nomen: D(omitius) Optatus, or similar, cf. RIB 1534, T(itus) D(omitius?) Cosconianus. It is not clear why Optatus should be associated both with Leg. VI and with Coh. IB(…). The unusual and almost identical layout of both reverses implies that they are contemporary, and thus do not indicate his transfer from one unit to the other. The likeliest explanation is that he was a legionary centurion acting-commander of an auxiliary cohort: instances are collected by Birley, E. in Latomus xlii (1983), 79–82.Google Scholar

⁵⁶ During excavation for the University of Lancaster and the Department of the Environment, directed by Dr T. W. Potter. See Potter, T. W., Romans in North-West England (1979), 1–138.Google Scholar

⁵⁷ Published with discussion by Dr D. C. A. Shotter, ibid, (see previous note), 73–4.

⁵⁸ See note to previous item.

⁵⁹ Published in detail by Potter, T. W., Romans in North-West England (1979), 75–87.Google Scholar The game(s) involved is discussed by R. C. Turner, ibid., 76–9.

⁶⁰ See note to No. 52. Published with ten other sherds bearing brief and fragmentary graffiti by Dr D. C. A. Shotter, ibid, (see previous note), 101.

⁶¹ During excavation for the University of Lancaster and the Department of the Environment, directed by Dr T. W. Potter. This and the next two items, with twelve other brief and fragmentary graffiti, are published by DrShotter, D. C. A. in Cumb. Westm. n.s. lxxvi (1976), 42.Google Scholar

⁶² During excavation for the Chelmsford Excavation Committee directed by Paul Drury. Details on this and the following two items from Mr C. Going, who made the sherds available for inspection. For the excavations and locations of Site ‘M’ ‘K’ and ‘D’, all excavated by Mr Drury, see the forthcoming C.B. A. research report edited by N. Wickenden on the temple and other sites in the northeast sector of Caesaromagus.

⁶³ By Mr D. Gamble, Engineering Department, Colchester Borough Council, and handed over by him to the Colchester Archaeological Trust. Made available for inspection by the Trust's director, Mr P. Crummy, who also provided details.

⁶⁴ Cf. stamps of Legion V Alaudae which regularly begin with the letters L V for l(egio) V, (the title Alaudae is not included), followed by the initials of the junior officers responsible. (CIL xiii part vi, pp. 14–15). For arguments against identifying the series of tiles from Lincoln stamped LVL followed by the letters A. D. E and F, with the products of Legion V, see Britannia viii (1977), 275–8.

⁶⁵ During rescue excavation conducted for the Department of the Environment by Dr I. M. Stead. Full information on this and the following two items, all now in the British Museum, from Valerie Rigby of the Museum.

⁶⁶ For the deceased's name compare the contemporary coins with the legend ANDOCO. ‘These have generally been found on the western borders of the Catuvellauni, and suggest that this unknown king may have ruled over some lesser tribe in this neighbourhood, which was later swallowed up in the Catuvellauni, probably during the time of Cunobeline’, Mack, R. P., The Coinage of Ancient Britain, 2nd ed. (1964), 63.Google Scholar An identification between the two seems not altogether impossible, given not only the appropriate dating and provenance of the burial, but also the extreme rarity of inscribed material in Britain before the Roman period. Burial 322 is not a particularly rich one, but then we do not know the circumstances in which Andoco lived after the suggested incorporation of his kingdom into that of his powerful eastern neighbours.

⁶⁷ Statistically, by far the most likely restoration of the name would be Secundus although there are other possibilities such as Verecundus and lucundus.

⁶⁸ This is not the well known villa at Plaxtol which lies about half a kilometre to the south.

⁶⁹ During excavation by the Kent Archaeological Society directed by Roger Crockett with the support of the owners of the site, Fairlawne Estate. Information and drawing from Mr G. de la Bédoyère. The spoon is to go to Maidstone Museum.

⁷⁰ During excavation for the Lincoln Archaeological Trust and the Department of the Environment directed by Miss C. Colyer; information and drawing from Maggi Darling.

⁷¹ We are most grateful once more to Professor R. Rodriguez-Almeida, for reading the inscription (from a drawing) and providing a possible interpretation. In the second line PIC, he very tentatively suggests, may indicate an origin for the produce, as, e.g., from Picenum. The letters also call to mind the word picatus pitch-flavoured, or picatum, specifically, pitch-flavoured wine, which would, of course, be inappropriate if the restoration given above is correct.

⁷² During excavation directed by the late Mr Charles Green for the Ministry of Works. Information from Mr D. Gurney of the Norfolk Archaeological Unit.

⁷³ During excavation for the Vindolanda Trust directed by Robin Birley, who made this and the next sixteen items available.

⁷⁴ The Roman pound was the equivalent of 327 gm. The weight, which is uncorroded, is thus only 55 gm (almost 3·4%) underweight. For similar weights see JRS xliii (1953), 130, Nos. 10(a), 11(a) and (b).

⁷⁵ For pairs of bronze spoons found in Britain, one inscribed utere, the other felix, see Sherlock, D., ‘An inscribed Roman spoon from Canterbury’, Arch. Cant. c (1984), 81–6.Google Scholar

⁷⁶ For a spear head with a similar punched inscription, see J. Curie, Newstead, 188. Coh. I Tungrorum garrisoned Vindolanda in the reign of Trajan: Bowman, A. K. and Thomas, J. D., Vindolanda: the Latin Writing-Tablets (1983), 119.Google Scholar

⁷⁷ See JRS xli (1951), 143, No. 12, for a brass helmet with XII incised upon it.

⁷⁸ For similar incised numerals on oak timbers from Carlisle, see above, Nos. 36–43.

⁷⁹ Possible restorations include the rare man's name Tarsa (CIL xii 4883) and the Pannonian place-name Mursa. A similar bung was found in a second-century context at Carlisle (Britannia xvii (1986), 440, No. 26), bearing the branded inscription LEG XI[ ]. This bung too might have come from the Danubian frontier-zone, where Legions XI, XIII and XIV were stationed in the second century.

⁸⁰ The B resembles A with a high crossbar, but this must have been because it was difficult to cut a curved stroke with a knife. Coh. IX Batavorum garrisoned Vindolanda in c. A.D. 100–105 (see Britannia xviii (1987), 128).

⁸¹ This ‘Celtic’ personal name is not directly attested (but cf. CIL ii 4404, Sulo), but must be the basis of the personal name *Sullonius implied by the place-name Sulloniacis (see A. L. F. Rivet and Colin Smith, PNRB, s.v.).

⁸² C is probably the first letter, but a small piece is missing just left of it, sufficient for 1–2 letters. There is a space after S. The graffito is clearly a note of contents, something granular but large and valuable enough to be counted rather than weighed. Some spice or precious herb is indicated, coriander (coriandrum) perhaps, which has been found in Roman Britain (Britannia viii (1977), 330), but the S defies explanation. Corsica exported wax and honey, but nothing worth counting into a jar.

⁸³ This ‘Celtic’ personal name, well attested in CIL xii and xiii, contains the name-element matu- (Ellis Evans, D., Gaulish Personal Names (1967), 228–30)Google Scholar and probably meant ‘bear-son’. In Britain it is found as the name of a mortarium maker: Frere, S.S., Verulamium Excavations I (1972), 372.Google Scholar

⁸⁴ Valens (‘strong’, ‘healthy’) and its cognates were popular cognomina, especially with soldiers.

⁸⁵ N is miniature and enclosed by the lower loop of 5. Probably a personal name, e.g. Censorinus.

⁸⁶ Probably the genitive ending of a personal name, perhaps a ‘Celtic’ one like Vindomorucus (RIB 2053).

⁸⁷ During the series of excavations directed by Mr C. M. Daniels for HBMC and the Metropolitan Borough of North Tyneside. This and the next eleven items were made available by Mr P. Moffat. They will be published by RSOT in the final report being prepared by Mr Daniels and Mr Moffat, together with other inscribed material from the fort (see Britannia vii (1976), 388–90; xv (1984), 347–8; xviii (1987), 372), duplicate tile stamps, and graffiti of less than three letters or of doubtful reading.

⁸⁸ Probably Mat[ernus] or, less likely, Mat[urus]; but many other less common names cannot be excluded.

⁸⁹ The graffito is apparently complete, and thus presumably an abbreviated personal name; but the choice is wide (Primus, Priscus, Probus, Proculus, etc.).

⁹⁰ The R is cursive, but the P is capital-letter. This P, which is standard in fourth-century New Roman Cursive, is sometimes found in earlier cursive texts, as here. Many names could be restored (see previous item).

⁹¹ This seems to be the sequence of strokes, and would imply that the graffito was made, and meant to be read, from the outside of the bowl. The reading XII, ‘12’, should thus be excluded.

⁹² A number of personal names end in … nalis, but the plural is puzzling. Brothers can be excluded, since the names are all cognomina; they tend to be theophoric (Neptunalis, Saturnalis, etc.), so a group of worshippers or other religious reference is possible. In this military context, however, a more likely restoration is [conluber]nales, ‘messmates’.

⁹³ By Mr J. G. Hunt, who intends to deposit it in the site museum, Aldborough, where there are other objects found by him, including another sealing (Britannia xviii (1987), 373, No. 30). Mr A. D. Hooley made it available.

⁹⁴ The first letter has been almost obliterated by the cord or wire, the third letter is only rudimentary, but distinct from the animal's tail. The text is restored from the two PBI sealings already known, both of which depict a right-facing bull: Roach Smith, C., Coll. Ant. III, 197, pl. xxxiiGoogle Scholar (Felixstowe), where the I has been misinterpreted as the bull's horns (autopsy by RPW, confirmed by a plaster cast); and RCHM, City of York I, Roman York (1962), 133.Google Scholar No. 144 (York).

⁹⁵ During excavation for Malton Museum and the Department of the Environment directed by Mr L. P. Wenham (see Britannia ii (1971), 252). The sherd was sent by Dr Brenda Heywood, who is preparing the final report.

⁹⁶ An abbreviated personal name, or a numeral (‘four’).

⁹⁷ Like the next two items, during excavation for the West Yorkshire Archaeology Service directed by Mr P. Abramson. Mr J. C. Clarke made all three items available. They will be published, with other inscribed items (see Britannia xviii (1987), 376–7), in the final report. For a similar tag, see JRS xvii (1927), 216, No. 22.

⁹⁸ Cilo(nis) could be understood as ‘(property of) Cilo’, the individual soldier's name, but is more naturally taken as the centurion's cognomen. The centurion seems to be otherwise unknown. His nomen Attius may conceal the Celtic or German personal name Atto (see Alföldy, G. in Epigraphische Studien 4 (1967), 10–16)Google Scholar , but is also a Latin nomen widely distributed in Italy and the western provinces. His origin is therefore uncertain.

⁹⁹ Inscribed quernstones found on military sites usually carry the name of a century, and this would seem to be so here.

¹⁰⁰ R is cursive, but the E is capital-letter, assuming that a faint horizontal stroke belongs to it; the top half of the letter is lost. The graffito seems to be complete, but since no personal name Uverus is known, V was probably repeated in error, for the common name Verus, i.e. <u> Veri, ‘(property of) Verus’.

¹⁰¹ During excavation for the West Yorkshire Archaeological Service directed by Mr A. B. Sumpter and Mr A. Tindall. Mr J. C. Clarke made it available.

¹⁰² The fourth letter was made with an initial diagonal stroke, suggesting M or N; it was not E or T. There is no knowing whether they belong to one word or two, but the space before and after 5 is greater than that between the succeeding letters, suggesting perhaps a personal name: e.g. [… u]s Ca[ndidus].

¹⁰³ During excavation for the Welsh Office (Ancient Monuments Branch) directed by J. D. Zienkiewicz, who provided full details and photographs. See further J. D. Zienkiewicz, The Legionary Fortress Baths at Caerleon, vol. II: The Finds (1986), 129, No. 7 with pl v., 7a and b.

¹⁰⁴ During excavation directed by Miss V. M. Metcalf for the Glamorgan-Gwent Archaeological Trust. Information on this and the following item from Mr S. H. Sell who submitted both objects for inspection. For the site see Britannia xii (1981), 317 with plan, 318, fig. 4.

In (a) line 2, V replaces II, while some of the other letters given may belong to an earlier text.

¹⁰⁵ During excavation directed by A. G. Marshall for the Glamorgan-Gwent Archaeological Trust. Information and copy of a drawing from D. R. Evans of the Trust. For another graffito from the same site see Britannia xvii (1986), 453, No. 86, and for the excavations in 1984, Britannia xvi (1985), 256.

¹⁰⁶ Information from George Boon.

¹⁰⁷ Information from Mr L. P. Wenham.

¹⁰⁸ Where seen by RSOT. Information from the Curator, Mr W. K. Milne.

¹⁰⁹ With the help of the Victoria and Albert Museum and the Haverfield Trust. Information from Colin Richardson. The stones were despatched for sale at Sothebys (sale of antiquities, 23 May 1988, lots nos. 371–3, where catalogued as ‘circa ist Century A.D.’) by Mr James Graham of Netherby Hall, who did not inform Tullie House, which is where the other surviving inscribed stones from Netherby are housed. When approached by Tullie House, Mr Graham after a week's reflection refused a pre-sale offer almost identical with what he subsequently received at auction.

¹¹⁰ Information from the Honorary Curator, Mrs E. Healey. The inscriptions are RIB 1022, 1025, 1029, 1058, 1072, 1073, 1075, 1076, 1078, 1079, 1081, 1084, 1087, 1091, 1092, 1093, 1094, 1101, 1103, 1115, 1161, 1179, 1264, 1265, 1271, 1272, 1286, 1291, 1359, 1416, 1482, 1536, 1539, 1637, 1684, 1726, 1779, 1833, 2310. For a brief history of the collection see RPW in Trans. Durham and Northumberland Arch. and Arch. Soc., n.s. 6 (1982), 93.

¹¹¹ When it was donated by Sir Arthur Evans, from the collection of Sir John Evans. Information from Martin Henig and Arthur MacGregor.

¹¹² See Britannia xviii (1987), 377. It is now built into the south face of the garden wall in front of ‘Duffenfoot’, where the owner, Mr Keith Blood, allowed access.

¹¹³ Compare its present state (FIG. 12) with the drawing in RIB, which implies it originally measured 045 by 021 m. This is close to the transmitted dimensions (in Lapidarium Septentrionale) of ‘about 1 foot 8 inches by 10 inches’ [c. 0·51 by 0·25 m].

¹¹⁴ There are unexplained marks after the P and final I, which are probably casual damage. Despite the drawing in RIB, there are no letters after the final I, nor need they be expected. The centurion's name almost certainly ran Ponti(…), and it is possible that a ligatured T and perhaps I have been lost, like a suprascript bar to II, in damage to the top of the stone. The most likely nomen is Pontius, but some rare cognates are also known (see W. Schulze, Eigennamen, p. 212), among them Pontienus. An attractive possibility is that the centurion should be identified with Pontienus Magnus, primus pilus of Leg. X Fretensis in Palestine in January 150 (PSI ix, 1026 = Cavenaile, CPL 117). The primus pilus would be at least 50 (cf. ILS 2641) if not 60 years old (cf. Juvenal, Sat. xiv 197, even after direct entry into the centurionate); another second-century centurion of Leg. XX took 57 years to rise from recruit (tiro) to primus pilus (AE 1985.735). It is thus most likely that Pontienus Magnus was already a centurion in c. A.D. 125, when RIB 1845 was inscribed. A further possible identification is with the (centuria) Pont[…] attestedat Black Carts (MC 29–T 29 a) (Britannia viii (1977), 431, No. 19), since two other centurions are attested in both sectors, Lousius Suavis (RIB 1499, 1506, and 1859, 1861, Britannia xii (1981), 380, No. 18) and the primus pilus Serenus (RIB 1510, and JRS lii (1962), 194, No. 20). Cohors II, no doubt of Leg. XX, is already attested (RIB 1848, 1851) near the find-spot of RIB 1845.

¹¹⁵ See Ramm, H. G., ‘The Duel Cross Milestone and Roman Roads west of York’, in Addyman, P. V. and Black, V. E. (eds.), Archaeological Papers from York Presented to M. W. Barley (1984), 43–5.Google Scholar

¹¹⁶ Jones, G.D.B., Shotter, D. C A., Roman Lancaster (1988), 186Google Scholar , with fig. 65.