² During excavation directed by Dr Ann Ellison for the Committee of Rescue Archaeology in Avon, Gloucestershire and Somerset. Ian Longworth and Catherine Johns of the British Museum made it available to RSOT. For the site see A. Woodward and P. Leach, The Uley Shrines: Excavation of a Ritual Complex on West Hill, Uley, Gloucestershire: 1977–9 (1993), which contains an interim report on the inscribed lead tablets, 113–30. This tablet is noted on p. 130, as No. 80. Tab. Sulis cited below is R.S.O. Tomlin, Tabellae Sulis: Roman Inscribed Tablets of Tin and Lead from the Sacred Spring at Bath (1988), reprinted from B.W. Cunliffe (ed.). The Temple of Sulis Minerva at Bath, II: the Finds from the Sacred Spring (1988).

³ The date is deduced from the ‘late’ ORC form of E, which is made in two halves, first a downward hooked stroke, then above it, two strokes at 45 degrees to one another. The hand has two idiosyncrasies: (i) the scribe made the first two strokes of M separately, but ran the third and fourth together, usually with a preliminary downstroke (e.g. in Mercurium (1) or ultionem (3)); (ii) when he wrote LL in illos (3) and ili (4 and 6), though not in tollat (5), he added a descender only to the second L. The scribe three or four times repeated one or two letters in error, copying-mistakes presumably: see notes to 3, 4 and 6 below.

⁴ Commentary

1. carta. This is the first explicit instance of a curse tablet being called a ‘sheet (of lead)’, although it is implicit in Tab. Sulis 8, which concludes with carta picta perscripta, translated there as ‘the written page has been copied out’. That χάρτα > c(h)arta means a ‘roll’ [not page] of papyrus, and thus generically ‘paper’, is convincingly argued by N. Lewis, Papyrus in Classical Antiquity (1974), 70–7. Rolls of lead, μολιβδíνονσ χάρτασ in Latin chartas plumbeas, were the cement overcoat of antiquity: see Josephus, C. Apionem 1.307 (with CSEL xxxvii, p. 69, ‘leprosi vero in chartis plumbeis involuti’). But the other references collected by TLL s.v. charta III (999.55–60) make it clear that charta plumbea generally means ‘lead sheet’, as used for coffin-making or waterproofing a dome. The emperor Nero famously lay down with a plumbeam chartam on his chest to strengthen his voice (Suetonius, Nero 20.1), but this was a sheet of lead, not a roll: compare Pliny NH XXXIV.166, Suetonius' likely source, where the term used is lamna.

qu(a)e. The hooked downstroke of E allows no room for A, so the scribe must have written QVE (a Vulgarism for quae), although a long diagonal was then added, as if to do duty for A. But there is no sign of any first stroke of A.

2. manecilis. The correct spelling mani- is found in 5, but although E here is damaged by the fold, it cannot be read as I. The Vulgar confusion between unstressed e and i is frequent. The word is a diminutive of manica (‘sleeve’), neuter manicilium elsewhere, but in this text it is taken to be masculine, as if manicilius; note qui (2), illos (3) and ipsos (5). Whatever its form, the word is very rare: except for a gloss, the only other instance is in Tab. Sulis 5, which likewise curses whoever stole two gloves. St Columbanus also cursed a glove thief, successfully: for this, and a critique of the received opinion that the Romans did not wear gloves, see the note to Tab. Sulis 5,2–3.

per[i]erunt. British curse tablets often use perdere in the sense of ‘to lose (by theft)’: see Tab. Sulis, p. 65. But except for peril in the Eccles tablet (Britannia xvii (1986), 428Google Scholar, No. 2), where the context is obscure, this is the first instance of perire. It is used as if it were the passive of perdere; compare Catullus 8.2: ‘quod vides perisse, perditum ducas’. A ‘Lost or Stolen’ notice from Pompeii (CIL I².1680) begins: ‘urna aenia pereit de taberna’. A reward is offered for its return or the discovery of the thief.

3. ultionem requirat. The first instance of this phrase in a British curse tablet. For another plea for ‘vengeance’ from Uley see note to pareas (6) below; and compare Tab. Sulis, p. 68. The initial V is peculiar: it is elongated, with trace below it (in the fold) of what looks like the downstroke of T. Possibly T was repeated in error from the end of the previous line: compare IT (6).

illos. The last two letters are very cramped, but the reading is not in doubt; for the masculine accusative plural, compare ipsos (5).

4. invalavi<i>t. I was repeated in error; compare IT (6) and quantoci<ci>us (6). There is another instance of the form invalav-, in a small group of curse-tablet texts (unpublished) which have affinities with Uley but are said to have been found at a site near Cirencester, so invalavit should not be regarded as a scribal error; instead it is an instance of the regressive assimilation of the pre-tonic vowel to the following accented vowel (see J.N. Adams, The Vulgar Latin of the Letters of Claudius Terentianus (1977), 14). Nonetheless it is exceptional, since involare is the regular verb ‘to steal’ in British curse tablets and is very frequent in the correct involav- form, which is what survives into Romance (Väänänen cites Fr. ‘voler’, It. dial, ‘involare’, Prov. ‘envolar’).

4–5. sangu(in)em [e]t sanitatem tollat. Almost the same formula was transliterated into Greek letters in another Uley text (Inv. No. 2169 (d) 1, unpublished): ‘tollas illi medullas, sanguem, animam’. The contracted form sanguem also occurs in Tab. Sulis 44, 5–6, evidently as pronounced, since it survives in Romance (e.g. Fr. sang, Portug. sangue). ‘Blood’ and ‘health’ are linked in Tab. Sulis 41, 3: ‘exigas hoc per sanguinem et sa[nitatem sua]m et suorum’.

5. ipsos manicili[o]s. There is no trace of the second O, but there is space for it, and manicilia cannot be read; the writer thought the word was masculine (see above, note to 2).

tulit. Tulit in the sense of ‘has stolen’ is already found in the Caerleon curse tablet, RIB 323 + add., whose script suggests a date earlier than the mid-second century; see also Tab. Sulis 47, 3, tulerit. Adams notes (in Britannia xxiii (1992), 2–3)Google Scholar that it carries this sense in late Latin.

6. [u]t. The first letter must be read as I, not V, but ut is clearly required to introduce the third in the sequence of ut- clauses (whether they are indirect commands or the ‘free-standing ut introducing a wish’ noted by Adams in Britannia xxiii (1992), 6)Google Scholar. The scribe's IT is presumably a reminiscence of the TVLIT he had just written; compare the initial letter of 3 (see note above).

quantoci<ci>us. Compare Tab. Sulis 54, 9, with note: ‘quantocius consumas’. It is just possible the scribe coined a new form, quanto citius, but in the pattern of damage due to the fold it is more likely that the second (vertical) stroke of C has been lost, rather than the second (horizontal) stroke of T. I was simply repeated in error, like IT (6) and the I of invalavi<i>t (4).

6–8. These lines do not correspond to any known formula, which hinders restoration of the missing text. 7 is very cramped, but is just legible except for the damage at the end: after rogamus two letters have been lost, the second possibly E; then N or R, followed by C or S; finally VRA. Censura or better pensura could be read, but makes no sense in this context. Overleaf, 8 has largely disappeared in a patination of corrosion products where the tablet was more exposed to attack. The line may have begun with a relative pronoun, and evidently ended with a verb like tollat (third-person singular present subjunctive). The traces suggest: q[u]o s[a]nc[t]u[m]

6. pareat. This also occurs towards the end of the other ‘vengeance’ text from Uley (Britannia xxiii (1992), 310Google Scholar, No. 5), where the writer prays ‘ut petitio mea statim pareat me vindicatum esse a maiestate tua’. There pareat = pereat is out of the question, and in the commentary it was suggested that pareat is for Classical pariat (‘accomplish’), which would also seem to be true of the present text. What Mercury is being asked to accomplish (quod deum Mercurium r[o]gamus) would then be ‘vengeance’, the punishment of the thief.

⁵ During excavation for Tyne and Wear Museums directed by Dr N. Hodgson and Mr P. Bidwell, for which see Britannia xxv (1994), 264.Google Scholar Alexandra Croom sent a rubbing, drawing and photograph.

⁶ The personal name Vitalis is very common.

⁷ By Miriam Hornsby, a staff member of English Heritage at Birdoswald Roman Fort, where the stone now is. Alan Whitworth sent photographs and other details.

⁸ To the left of V is ‘the remnant of a deliberate cut and part of a larger letter worn away’. The centurion is otherwise unknown.

⁹ During excavation by the Rochford Hundred Field Archaeology Group directed by Neil Faulkner. Pam Gilmour provided details and made the sherd available.

¹⁰ During excavation directed by Mark Atkinson for Essex County Council, Bovis Homes and English Heritage, for which see Britannia xxvi (1995), 360.Google Scholar Colin Wallace, Essex County Council Planning Department (archaeological section), sent details of this and the next item.

¹¹ A provisional reading by Mr Wallace (see previous note).

¹² During excavation directed by R. Havis for Essex County Council, British Airports Authority and English Heritage. For the site see Britannia xix (1988), 459.Google Scholar

¹³ Identified by Brenda Dickinson.

¹⁴ For this well-attested Celtic name see Holder, Alt-celtischer Sprachschatz, s.v., and compare Britannia viii (1977), 430, No. 18 (Viducius Placidus and his father(?) Viducus); but note also RIB II.8, 2503.453 and 454, VIDI and VIDIC[…].

¹⁵ By metal detector. See Williams, H.P.G., ‘A Roman Lead-sealing of Carausius from Gloucestershire’, Britannia xxvi (1995), 323–4CrossRefGoogle Scholar, with fig. 7. The impressions closely resemble official Carausian coinage, and in particular the reverse type is identical to RIC V.ii.906, ‘dated early in the reign, probably late 286–7’. Williams suggests that official coin dies were used.

¹⁶ Williams (see previous note) suggests ‘possibly of parchment’, but papyrus or a wooden leaf ink-tablet are also possibilities.

¹⁷ Probably the standard legend IMPCARAVSIV[SPFAVG], Imp(erator) Carausiu[s p(ius) f(elix) Aug(ustus)], but IMPCARAVSIV[SPAVG] and IMPCARAVSIV[SAVG] are also possibilities.

¹⁸ By metal-detector at a date unknown. D.F. Mackreth, who will publish the brooch in a report on the site being prepared by the Cotswold Archaeological Trust (see Britannia xxvi (1995), 365–6)Google Scholar, sent a drawing and other details to RSOT.

¹⁹ No other example of the name is known to Mr Mackreth (see previous note). It is not recorded in RIB 2421 or in G. Behrens, ‘RPömische Fibeln mit Inschrift’, Reinecke Festschrift (1950), 1–12.

²⁰ By Mrs P.H. Albone while field-walking. Mr J.E. Albone sent full details including the results of his own research; he would like to acknowledge the assistance of Professors S.S. Frere, M. Millett and M. Steinby, and (for the petrological analysis) Dr David Williams.

²¹ When complete, there would have been room for about twelve letters in (a) and about seventeen in (b), so that if both carried the same text (as is suggested by the surviving letters), there will have been a greater degree of abbreviation in the inner circle. In (a) there are traces of what may be the right-hand arm of V before the first C. In (b) there are traces of a letter preceding the V which are consistent with a damaged C.

It is tempting to interpret both (a) and (b) as alluding to Roman Colchester, which was known variously as Victricensis (JRS 1ii (1962), 190–1)Google Scholar, Colonia Victricensis (CIL XIV.3955) and, we may suppose, in full as C(olonia) C(laudia) Victricensis, perhaps with the addition of the initials of eponymous magistrates as on the colonial tile stamps from Gloucester (RIB II.5, 2487 and 2488). But these stamps have not been found in Colchester itself, and there is only one example from Britain (RIB II.5, 2482) of the circular stamps which are typical of Italy. So it is possible that this object is a modern import, in which case the letters may be part of a personal name. Petrological analysis was unable to point to a source for the tile, although it excluded the volcanic regions of Italy.

²² During excavation by Tempus Reparatum, Oxford, on behalf of Lincolnshire County Council, directed by Dr J.R. Hunn and Mr E. Guttman: see Britannia xxvi (1995), 350.Google Scholar Mr M. Wood and Dr D.P. Davison made it available with a second lead tablet from the same site, 25 by 31 mm, which when unrolled was found not to be inscribed. The tablet is published with full commentary in Zeitschrift für Papyrologie und Epigraphik, forthcoming. RSOT.

²³ Comparable hands are tabulated in Tabellae Sulis (see note 2 above), p. 94. The reading is not discussed in detail here, but note that the magical names in 4–7 retain their Greek omega.

²⁴ The invocation of the three magical protective deities in 4–7 identifies this lead tablet as a ‘phylactery’ or protective magical text, not a ‘curse tablet’. Their names ultimately derive from attributes of God in the Hebrew scriptures, but they are frequent in Greco-Roman magic as independent deities. That this text is an uterine phylactery can be deduced from the distinctive command ‘Stay in your place’, and it is in fact a Latin translation of a well-attested Greek spell: see K. Preisendanz, Papyri Graecae Magicae (2nd edn, 1973–4), VII 260–71, and R. Kotansky, Greek Magical Amulets: The Inscribed Gold, Silver, Copper, and Bronze Lamellae, Part I, Published Texts of Known Provenance (Papyrologica Coloniensia XXII/1, 1994), 265–9, No. 51. Dr Kotansky has drawn our attention to an Aramaic translation also: see P. Schäfer and S. Shaked (eds), Magische Texte aus der Kairoer Geniza I (1994), 108–19 (K 1.157, lines 12–23). The spell is prompted by the ancient belief (e.g. in Plato, Timaeus 91C) that the womb is a separate living creature liable to move about a woman's body and cause her various illnesses. In general see J.J. Aubert, ‘Threatened wombs: aspects of ancient uterine magic’, Greek, Roman and Byzantine Studies xxx (1989), 421–49. A non-specific Greco-Egyptian haematite uterine amulet has already been found in Britain (RIB II.3, 2423.1).

²⁵ Letters are transcribed as in the original, without spacing or punctuation. The dotted letters are now incomplete, and their reading depends on the context.

²⁶ Lost letters have been restored where possible, words separated, modern punctuation and hyphenation added, proper names given initial capitals. The capital letters in 3–4 mark a word which has not been interpreted. The personal name in 11 is not certain.

²⁷ With the next two items during excavation for the Museum of London Archaeological Service (MOLAS) directed by Bruce Watson and Trevor Brigham, who provided full details of this and the next fourteen items. Angela Wardle of MOLAS made them all available, and Robin Symonds, also of MOLAS, identified the pottery. For the Regis House site see above, p. 427.

²⁸ The drawing (FIG. 5) omits the first triangular stop; the stops are all very faint. Ten lead pigs of Vespasian's reign have already been recorded from the Mendips field (see RIB II.1, 2404.4–13), so the total is now thirteen. Inscription (a) of this item seems to have been produced by the same die as the top-face inscription of RIB 2404.11.

²⁹ This strip is not recorded for RIB 2404.11 (see previous note), and may therefore be distinct from the die which produced inscription (a).

³⁰ Both stops, the first (circular) after BRIT and the second (triangular or leaf-stop) after ARG, are again very faint. The same text occurs on the front face of the next two items, and on the front face of almost all the other Vespasianic lead pigs from the Mendips field (RIB 2404.4–10). (The only exceptions are RIB 2404.11 and 12, both incompletely recorded; and RIB 2404.13, which is later than all the others, being dated to A.D. 79, when EX AR(G) has been incorporated into the top-face inscription.) Several dies were used. The die here is not identical with the die(s) used for the next two items, but it measures the same length as the die used for RIB 2404.7 and 8, and would seem to be identical. VEB is probably the abbreviated name of the lead-mining district in the Mendips, Somerset: see A.L.F. Rivet and C. Smith, The Place-names of Roman Britain (1979), 487.

³¹ The first stop is very faint and may be circular rather than triangular. Trevor Brigham has drawn our attention to four rectangular raised lumps of lead, between the arms of the first V, between the top of the E and the following S, between the I and A of VESPASIAN, and between the two strokes of the A of AVG. These projections correspond to depressions in the mould, which conceivably were caused by the projecting heads of nails somehow used in the original die. They occur in the same positions on the surface of the next item, and can also be recognized in R.P. Wright's drawing of RIB 2404.4, thus proving that the identical top-face die was used in the moulds of all three ingots.

³² No stops are actually visible. For the text see note 30 above. The die is not identical with that used for the front-face inscription (b) of the previous item, but it measures the same as that used for the front-face inscriptions (b) of the next item and of RIB 2404.4, which also both share the same top-face inscription (a) (see previous note). Whether the die used for inscription (b) was identical is not quite certain: the letters in inscription (b) of the next item appear to be slightly more spaced out than here.

³³ The letters at first sight look as if they have been cold-struck with a chisel, but they were in fact produced by the same die as the next item, inscription (c), whose whole appearance is consistent with the impress of an iron die. There is no sign of the horizontal stroke of L here or in the next item, but it can be read by analogy with the die-struck inscription LRAD on RIB 2404.6 (probably) and on RIB 2404.8. The letter A will have had a pronounced leftward serif: see note 36 below. The corresponding die-struck impressions on the other Vespasianic lead pigs from the Mendips field consist of abbreviated Roman names: Ti(berius) Cl(audius) Tri(ferna) on RIB 2404.7 (from the same mould as RIB 2404.8, which carries LRAD), and on 2404.9 and 10; and later, in A.D. 79, G(aius) P(…) C(…) on RIB 2404.13. LRAD should be understood in the same way.

³⁴ The inscription carries the same rectangular projections as were noted on the previous item (see above, note 31) in the same positions, proving that the die used in the moulds for both these ingots and RIB 2404.4 was identical.

³⁵ As with inscription (b) of the previous item, no stops are actually visible and the letters are c. 20 mm high, but in this item the letters appear to be slightly more spaced out, as if the die was not quite identical. For the text see note 30 above.

³⁶ The die used was identical with that for inscription (c) of the previous item, but the bottom part of the letters is more deeply impressed, and shows that the foot of the left stroke of A was provided with a pronounced serif. The whole appearance is consistent with the impress of an iron die. For the reading of L and the interpretation of LRAD, see note 33 above.

³⁷ Damian Goodburn of MOLAS discussed this item with us, and provided a tracing of the branded inscriptions.

³⁸ For variations in the spelling of ‘Thracum’, see the Epigraphic Indexes to RIB I and II, 6.5, ‘Auxiliaries’; and note in particular RIB 109, alae Trhaec(um). This would then be a military unit otherwise unattested in first-century Britain. One ala (I) and four cohorts (I, II, VI and VII) of Thracians are known here, but none of them bore the title Augusta: see the Epigraphic Indexes to RIB, loc. cit. The only Thracian units so honoured were three other cohorts, I, II and III Thracum, of which the first two, and probably the third, were equitata; and three alae, I, II and III Thracum. See M.G. Jarrett, ‘Thracian Units in the Roman Army’, Israel Exploration Journal 19, No. 4 (1969), 215–24, and, most recently for the alae J. Spaul, Ala² (1994), 228–35. In the nomenclature of these units the honorific Augusta normally follows the word cohors or ala, but sometimes it follows the ethnic: e.g. CIL III.110, Coh(ors I) T(hracum) Aug(usta); AE 1930.89, Ala (I) Thracum Aug(usta); CIL XVI.56 (a diploma), Ala II Thracum Augusta; Roxan, RMD I No. 4 (another diploma), Ala III Thracum (sagittariorum) Augusta. But none of these units, so far as is known, ever served in Britain or indeed in any of the north-western provinces, except just possibly the Ala II Thracum Augusta p(ia) f(idelis). Its service in Lower Germany must be deduced from the titles p(ia) f(idelis), which might have been awarded for loyalty at the time of the revolt of the governor of Upper Germany, Antonius Saturninus, in A.D. 89. But if so, it had already been transferred (titles and all) to Mauretania by 107: see CIL XVI.56. The argument in any case is not conclusive, and G. Alföldy, Die Hilfstruppen in der römischen Provinz Germania inferior = Epigraphische Studien 6, pp. 8–9 n. 39, cites other units which bore the title p(ia) f(idelis) for which there is no independent evidence that they ever served in Lower Germany.

³⁹ The fifth letter could be R rather than B. There is a triangular cut-out after the first three letters, and a lunate cut-out after the S. The reading is fairly clear, but the combination of letters is unusual. It seems impossible to read OCTOBRIS, but a personal name is a possibility. If the letter C once occupied the space now taken by the triangular cut-out, such a name might have incorporated the Celtic name-element cobro-, for which see Holder, Alt-celtischer Sprachschatz, s.v., and compare Corrigendum (h) below.

⁴⁰ Less probably, if it is read inverted: IV, presumably ‘4’.

⁴¹ This is plausible as an abbreviated personal name, but L(?) must be interpreted as being written retrograde, like an inverted K but without the lower diagonal stroke.

⁴² Germanus is common; and for Germanicus and its derived personal names, see Kajanto Cognomina, 201. If the reading is correct, the only other possibility suggested by Mócsy Nomenclator is the unique Parmanicus.

⁴³ Before the first surviving letter (I) there are two horizontal strokes, of which the upper could be the top bar of the letter F, and the other the beginning of a line running below the whole exhortation. There does not appear to be another vertical stroke before the first surviving letter, to make II (for E). Filix for felix is already found at Pompeii (C1L IV.4511). If this interpretation is correct, feliciter might have been preceded by the name(s) of the person(s) to whom the exhortation was addressed.

⁴⁴ During excavations in the crypt directed by Mike Hammerson for the Southwark and Lambeth Archaeological Excavation Committee. For the site see Britannia ix (1978), 453–4; and M. Hammerson, ‘Excavations under Southwark Cathedral’, London Archaeologist 3 No. 8 (1978), 206–12. Angela Wardle of the Museum of London Archaeological Service sent details and made the sherd available.

⁴⁵ This is by far the commonest of the various possibilities suggested by Mócsy Nomenclator. Naevius as a nomen is fairly common, but in Britain it also occurs as the personal name of an imperial slave (RIB 179).

⁴⁶ With the next item, which is apparently part of the same amphora but does not conjoin, during excavation by the Central Marches Archaeological Research Group directed by J.H. Allen; see Britannia xxii (1991), 246–50.Google Scholar We have not seen the late-Roman silver spoon (ibid., 250) inscribed: VTERE FELIX VIVAS. Lynne Bevan of the Field Archaeology Unit, University of Birmingham, made this sherd and the next item available to RSOT.

⁴⁷ This seems to be the first instance of the personal name Gaio, but in the alternative spelling Caio it is twice found, at two places fifty miles apart just to the north-east of Italy: in CIL III.11592 (Virunum, Noricum) and in CIL III. 10795 (Municipium Latobicorum, Pannonia Superior). It is certain from both inscriptions that the form is Caio (nominative) with genitive Caionis, and almost certain that it is a Celtic personal name. Its resemblance to the Latin praenomen Gaius, quite often used as a cognomen (whence the names Gaianus and Gaiolus), is a coincidence. But note the unique names Geuo(?) at Hardknott (RIB II.7, 2501.699) and Gauo at Vindolanda (Tab. Vindol. II, No. 192, where the editors compare Gauolus at Aquileia (CIL V.337)). Gaio would seem to have been a Roman auxiliary soldier of Celtic origin, quite possibly recruited in Noricum or its neighbourhood.

⁴⁸ Between T and F there is what looks like a small O. But the readings this suggests, ToFT or ToFI (the I barred, and thus a numeral), make no sense, whereas ‘Titus Flavius’ is plausible at a fort of Flavian date. ‘O’ must then be regarded as an earlier graffito, or an insertion of unknown purpose (possibly a medial point). But in view of this uncertainty, and also the inconclusive forms of T and F (which do not exclude a reading ‘III’, ‘three’), the reading TFT and its interpretation are not beyond doubt.

⁴⁹ With a metal-detector by Mr Paul Whitehead, in whose possession it remains. Judith Plouviez of the Department of Environment and Transport (archaeological section), Suffolk County Council, sent full details and a drawing.

⁵⁰ For another silver ring inscribed DEO | MER, see RIB II.3, 2422.20 (Corbridge), and for others simply MER, ibid. 29 (Chesterholm), 30 (Billingford, Norfolk). A (stolen) gold ring is ‘given’ to Mars/Mercury in a text from Uley: see A. Woodward and P. Leach, The Uley Shrines (1993), 123 = Britannia x (1979), 344Google Scholar, No. 4 (revised).

⁵¹ During excavation for Tyne and Wear Museums directed by Dr N. Hodgson and Mr P. Bidwell. Details of this and the next seventeen items were sent by Alexandra Croom, who also provided rubbings, drawings and photographs.

⁵² They come from different moulds, but are identical in style and execution with each other and with the (lost) belt-plate from Brampton, Cumbria (RIB II.3, 2429.11).

⁵³ IX and XL are possible readings, but look less likely.

⁵⁴ For an instance of IVN from Wallsend, see RIB II.7, 2501.274

⁵⁵ F resembles C, P is made with a short diagonal second stroke, not a loop, and the tail of Q is vertical. For other alphabets incised on tiles see RIB II.5, 2491.135–45, and in particular a somewhat similar alphabet from Wallsend (142) and ABC[…] from South Shields (139).

⁵⁶ The first L is written with a long diagonal descender, the second with a horizontal stroke that begins with a flourish well to the left; it thus resembles the elaborate serif that was added to the R. The combination of Lucius and Aurelius occurs in the name (after accession) of the emperor Lucius Verus (A.D. 161–9) and of persons who owed their citizenship to him, but otherwise it is rare.

⁵⁷ Two more inscribed fragments of tile from South Shields are omitted here, each of which contains only one or two letters of uncertain reading.

⁵⁸ A is perhaps the final letter, in which case [Ma]ura (cf. RIB II.7, 2501. 364) is possible.

⁵⁹ The third letter might otherwise be R, but for VIA as a graffito (on samian) see RIB II.7, 2501.604 and 605.

⁶⁰ The last letter might otherwise be G, but this seems less likely. The dish must have been turned upside-down, to hold it more firmly before inscribing. VIS might be an abbreviated name, or possibly a numeral: VI s(emis), ‘six and a half’.

⁶¹ Presumably to identify the dish, whether as IV, VL, or even S.

⁶² During excavation at 36 Fishbourne Road East before redevelopment, directed by John Wildman of Southern Archaeology (Chichester): see The Archaeology of Chichester and District 1995, 13–15. The ring remains the property of the site developer. It was found with first-century domestic refuse, including fragments of high-status pottery and part of a decorative cavalry harness, in the backfill of a ditch feeding a piped water-supply making for the palace. John Magilton sent details and a drawing. The ring was examined by Catherine Johns at the British Museum, who informs us that there is ‘no real doubt that it is first century and indeed pre-Flavian’. Martin Henig concurs, noting that the form is found at Pompeii; for Britain he compares M. Henig, A Corpus of Roman Engraved Gemstones from British Sites (1978), No. 784 with pl. LII (Wroxeter), and for Fishbourne in particular, B. Cunliffe, Excavations at Fishbourne 1961–1969, II, The Finds (1971), 88 No. 2 (a gold ring with intaglio) and 91 No. 5 (a silver ring with engraved bezel). RSOT.

⁶³ The inscription was read from a photograph provided by the British Museum. The first T has lost its cross-bar, but is guaranteed by the context since Claudius is regularly preceded by the praenomen Ti(berius), and invariably if it is the name of a first-generation Roman citizen (see next note). The second T has lost its downstroke, but no other letter is possible. Only the final letter is difficult; but since a genitive inflexion of the cognomen is required by the gentilicium CLAVDI, an I is probable, and the surviving traces are consistent with a slightly canted I having a bold upper serif, now abraded.

⁶⁴ The name Catuarus combines two elements found in many other Celtic personal names, initial Catu- (‘battle’) and the suffix -arus, but was not known to Holder. It is now attested at Milev in Numidia (CIL VIII.19989) as the name of the non-Roman father of an enfranchised veteran of Publius Sittius (the condottiere who died in 44 B.C.), called P(ublius) Sittius Catuari f(ilius) Senex. Sittius recruited his mercenaries from Italy and Spain, and the veteran's own name Senex is a typical ‘Roman’ name incorporating a Celtic name-element, so his Celtic origin need not be doubted. The Fishbourne ring is historically important because in the mid-first century, its likely date, only members of the equestrian aristocracy would have worn a gold ring: compare the other Fishbourne gold ring cited above, in note 62, and see Henig's commentary thereto (in particular Pliny, NH XXXIII.32). This exalted status would have been far above a veteran auxiliary soldier, which is what the combination of the reigning emperor's gentilicium with a Celtic personal name would normally suggest. (A Gallic aristocrat, say an equestrian officer in the invading army, would almost certainly have been a lulius.) Instead, Catuarus must be seen as a Briton of wealth and social position who owed his citizenship to the favour, direct or indirect, of Claudius or Nero. The classic example is of course the client king Tiberius Claudius [To]gidubnus (‘Cogidumnus’), already attested at Chichester (RIB 91 + add.). Catuarus is likely to have been his kinsman or ally, rewarded like him for collaborating with the Roman invaders.

⁶⁵ By the owners, Messrs Simtor, in return for a replica. Accession No. 1995.197. Information from Ian Caruana.

⁶⁶ Where they were seen by Haverfield in 1903, but sought in vain by R.P. Wright in 1951; in 1985 the Museum told Kay Hartley (letter to R.P.W., 6 June 1985) that they had been rediscovered.

⁶⁷ By R. Kotansky, Greek Magical Amulets (see note 24 above), 3–12, No. 2, with pl. II.

⁶⁸ ibid., 1–2, No. 1, with pl. I.

⁶⁹ In the facsimile reprint (1995) of RIB with Addenda and Corrigenda, at p. 780.

⁷⁰ The square brackets distinguish letters already lost to Camden, but others were probably then visible although he missed them, i.e. T ligatured to R, A ligatured to M (twice), and T mis-read as I.

⁷¹ By R. Kotansky, Greek Magical Amulets (see note 24 above), 13–15, No. 3, with Pl. I.

⁷² Kotansky (ibid.) reads after autopsy: [----magic] signs vac. | [------]DONA | [----] ADONAE M | [---] (magic sign) Y (sign) barred SSSS TTT | [-------] (magic sign) EEEE | [--------].[---]. The original has also been re-examined by R.S.O.T., who has discussed the reading with Dr Kotansky. In 2, despite tentatively reading DONA, Kotansky draws ONDE but, as he remarks, ‘the reading is quite uncertain’; R.S.O.T. sees only the first two strokes of ‘N’, considers that ‘D’ resembles A with cross-bar, and sees no sign of E at all. In 4 Kotansky reads as Y what might otherwise be cursive E or a magic sign; then a magic sign resembling an acute-angled L (the horizontal stroke is omitted in R.P.W.'s drawing); and then a ‘four-barred S’ which he equates with the three-barred SSS of ‘Chnoubis’ amulets. The difficulty here is that there are not the usual three SSS with a single cross-stroke, but four, in a context of other repeated letters (TTT and EEEE); and that, while the first S and the fourth S are cut by short diagonal strokes (not drawn by R.P.W.), the second (so at least it seems to R.S.O.T.) is not, and the apparent stroke which cuts the third S, lengthened in Kotansky's drawing to include the fourth S, is faint and looks casual.

⁷³ The text thus apparently consists of magic signs, repeated Latin letters, and (in 2) the invocation Adonae for Greek ’αδωναí, ‘Lord’.

⁷⁴ Compare the previous item and footnote. This was pointed out by L. Schwinden in Hémecht 44 (1992), 97, note 61.

⁷⁵ By Fraser Hunter of the National Museums of Scotland, Edinburgh, who sent photographs and discussed the reading with us.

⁷⁶ (ii) is in smaller capitals. The second vertical stroke of N is rather faint, and S is of cursive form. Nestor is already attested in a northern military context: see RIB II.7, 2501.4 and 409, and compare RIB II.6, 2494.155. The cognomen Taurus is frequent in Gaul and Britain.

⁷⁷ By Frances Mawer in Britannia xxv (1994), 232–6, with drawing and photograph (ibid., pl. XVI B).

⁷⁸ Salmon's ‘Mark A’. If this is not pure decoration, it might be an unbarred alpha to which corresponded an omega on a companion tank; note that a lug from another tank was found with RIB 2416.10

⁷⁹ See E. Birley, R. Birley and A. Birley, Vindolanda Research Reports II: The Early Wooden Forts. Reports on the Auxiliaries, the Writing Tablets, Inscriptions, Brands and Graffiti (1993), 82–3, with fig. 7. Although we are still bound by RIB convention, we note with approval the ‘mild protest’ (p. 19) ‘against colleagues who continue to refer to this site as ‘Chesterholm’, the name of the house in the valley below. If ever a Roman site deserved to be referred to by its Roman name, this one surely does’.

⁸⁰ Possible nomina in CIL XIII, both derived from Celtic personal names, are Cobrovius (cf. RIB II.8, 2503.141, Cobrova) and Cobrunius. R. Birley has suggested (see previous note) C(aii) Ob(ult)r(onii) Sab(ini) c(larissimi) v(iri), but this expansion ‘naturally remains very uncertain’.