² With the next three items 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 them available to RSOT. Professor J.N. Adams discussed two difficult readings. 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 one is noted on p. 129, as No. 55. 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 This is not a nail-hole (the tablet could not have been folded as it was, had it been nailed), but is probably caused by a ‘cold shut’, a casting flaw. It looks as if line 2 deliberately avoids it.

⁴ Commentary 1–2. The ‘address’ is of usual form (see Tab. Sulis, pp. 68-9): to the god from a (named) petitioner. But the name is difficult. At the end of 1, a terminal I was first written (like those in 2 and 5), then L was written over it. Like eos vel in 3, this would seem to have been a correction. In 2, what looks like ligatured VI has been written above the hole, but since this must be a name-ending (and Rufus is common), it is acceptable instead as a carelessly ligatured VS. (There is no sign of the second stroke of S.) Mintla is for Classical mentula (‘phallus’); mentla is the usual Vulgar form, but mintla is also found in RIB 631, Priminus mintla. These are both instances of mentula being used ‘quite neutrally’, like those noted by J.N. Adams, The Latin Sexual Vocabulary (1982), 12. Exactly how mintla is being used is not clear: Rufus like Priminus may have used it as a second name or self-identification, or perhaps the petitioner's name was composite, Mintlarufus (‘Redprick“).

2. donavi is formulaic (see Tab. Sulis, pp. 63–4), for the ‘gift’ of either the stolen property or the thief. Here the thief/thieves is intended.

3. eos: the scribe seems to have first written ESO, which he corrected by inserting O after E, and writing the V of vel over O. For the usual ligature of EO see deo in 1.

3–4. vel mulier vel PARIVS. The scribe intended the ‘whether male or female’ formula (see Tab. Sulis, pp. 67–8), which isolates the thief by means of mutually exclusive alternatives, usually linked by si not vel. (Note also that, as in the item, in spite of the plural eos the singular person was retained in the formula.) The formulaic alternatives to mulier are vir, baro or mascel, one of which would be expected after the second vel, but the reading PARIVS is certain.

4. After PARIVS the reading is again hardly in doubt, but the meaning remains obscure. LII is not a numeral (the context is against it, and there is no suprascript bar); and FASPATEM (P not T must be read) is gibberish. The explanation seems to be another copying error (like ESO for eos, but worse). The scribe was copying the usual formula vel mulier vel vir, vel servus vel liber, but he wrote PAR for VIR (they would have looked almost identical; indeed the same is true of the BAR of baro, which is also a possibility); then the scribe's eye slipped from vir to servus. His RIVS sequence would be almost identical in appearance with the RVVS of servus, and the succeeding LIIF sequence looks like a garbled LIB of liber. Finally ASPATEM. There are two formulas of which this may be a corruption: either maiestatem (for which see Tab. Sulis, p. 65) or the donation of a ‘part’ or fraction to the god (tertiam partem in Woodward and Leach, The Uley Shrines, 121, No. 2).

5. The beginning of the line is lost in damage due to the fold, but the restoration [majteriam is inescapable. In a concrete sense, materia means the ‘materials’ of which something is made, specifically ‘timber'; there seems to be no instance of materia alone in the modern specific sense of ‘material’, i.e. woven fabric. However, note Tertullian, De Pallio 3, where materia is used of the materials (especially wool) from which the cloak (pallium) is made; and cf. Augustine, De Civitate Dei XXII.24, ‘in fruticibus et pecoribus indumentorum conficiendorum quanta materies!’ (What a wealth of vegetable and animal material is available for making clothes.) Where sagi is explicit, therefore, we may understand materiam sagi as ‘the material of a cloak’ [presumably the woollen cloth from which it was made]. (SAGI incidentally must be read; there is no sign of initial F for fagi, ‘of beech’.) The theft of cloaks is frequent in the Bath tablets (see Tab. Sulis, p. 80, and note 62.1-2, la[enam pajllewn sagum paxsam), which might be expected at a bathing- establishment, but the Uley tablets do mention linen (Woodward and Leach, The Uley Shrines, 121, No. 2), wool (ibid., 129, No. 58), and a pallium (ibid., No. 62). The syntax requires qui involaverunt (or similar) if the material is the object of theft by eos (3), as it seems to be, but this formula has evidently disappeared in the confusions of line 4.

6. donavi is separated from the main text by a space, and by being centred; there is nothing written on either side of it. It is repeated from 2 for emphasis; cf. Tab. Sulis 9. 14, a list of names in capitals followed by dono in cursive, likewise centred at the foot of the ‘page'. These are the only instances of this usage.

⁵ Letter by letter, without punctuation but retaining the original word-separation and layout. Subscript dots mark ‘doubtful’ letters, which are none the less complete but whose interpretation is uncertain; for details see the commentary above.

⁶ With the previous item. See Woodward and Leach, The Uley Shrines, 130, No. 76.

⁷ Commentary

ai. Both ends of the line have been lost by corrosion, but the restoration is certain; cf. Britannia xxiii (1992), 310, No. 5 (Uley), deo sancto Mercurio Honoratus. conqueror numini tuo….

2–3. mihi male cogitant: except for the use of the dative instead of de, this idiom is Classical; cf. Caelius in Cicero, ad Fam. VIII.12.1, male eum de me cogitare. This is its first occurrence in a curse-tablet text.

3. et male faciunt: there is no exact parallel in curse-tablet texts, but cf. Britannia xx (1989), 329Google Scholar, No. 3 (Uley), qu[i] pecori meo dolum malum intulerunt; Britannia xxv (1994), 296Google Scholar, No. 2 (Weeting with Broomhill), [male]fic(i)um; and cf. Tab. Sulis, p. 64, for fraudem facere.

4. This line is quite badly damaged, which is unfortunate, since it was non-formulaic and evidently gave details of the malice complained of in 2–3. The drawing represents all that is visible. Taking it letter by letter (with possible alternatives in brackets) we may read: S(or C)VR(or P)RA space ED[3]S(or C)IVMEN(or EX3-4]. Supra fits: not part of supradictus or similar (cf. Woodward and Leach, The Uley Shrines, 121, No. 2.7 with note), but possibly introducing some phrase meaning ‘beyond what is right’ or, more likely, supra as the Vulgar equivalent of super, used with ablative like de in the sense of ‘over’, ‘in the matter of’. What follows can then be separated in two ways, either ed[3]cium en[3–4] or ed[…]s iumen‘3–4]. For the first alternative there is no obvious restoration (hardly ed[ifi]cium), but for the second the sequence IVMEN[…] is attractive, since it suggests iumentum (a ‘draught animal’, especially a mule), the object of theft in Woodward and Leach, The Uley Shrines, 118, No. I = Britannia x (1979), 340Google Scholar, No. 2; cf. Britannia xxiii (1992), 310Google Scholar, No. 5 (the theft of four cows at Uley), and Britannia xx (1989), 329Google Scholar, No. 3 (malice towards a farm animal (pecus) at Uley). But ED[3]S remains insoluble; we have rejected supra <ra>ed[uli]s iumen[taris] and supra ed[iti]s iumen[ti].

5–6. For these frequent formulas see Tab. Sulis, p. 67. As in the previous item, they are in the nominative singular as usual, even though the syntax would require a plural (cf. Mis in a2 and a6, cogitant in a2, faciunt in a3). Note also possit below (b2).

6–7. ut non permittas: this formula is also frequent (Tab. Sulis, pp. 65–6). Ut might be taken as dependent upon queror (ai), a final clause or an indirect command, but Adams has suggested (Britannia xxiii (1992), 6) that it should be seen as free-standing, and introducing a wish. This is an attractive suggestion, which has been adopted here, although there is no curse-tablet text where the distinction is beyond doubt.Google Scholar

8–9. nec stare nec sedere: this formula also occurs in two unpublished curse-tablet texts from a site near Cirencester, and is a natural variant of nec iaeere nec sedere in Tab. Sulis 54.5 and Britannia xxiii (1992), 310, No. 5 (Uley).Google Scholar

a9–bi. nec bibere nec manducar[e] also occurs in the three texts just cited from [Cirencester] and Uley, as well as in Tab. Sulis 41. 4.

bI–2. ha[s] iras: this reading is not certain, but it is consistent with what remains. H is certain, and so is S; otherwise only downstrokes survive, except that R is suggested by what looks like the beginning of its diagonal stroke, T being excluded by the absence of a rightward termination to the downstroke. The formula of ‘redemption’ or buying-back only at the cost of one's life is well attested (see Tab. Sulis, pp. 65 s.v. nisi and 66 s.v. redemat), but the object ‘redeemed’ is problematic: it is the donum ‘given’ to the god in Tab. Sulis 65, apparently the thief himself; the stolen property in RIB 323 (Caerleon); and ‘what they have administered to me’ in Britannia xx (1989), 329Google Scholar, No. 3 (Uley). In effect it is the curse itself, which in curse-tablet language is activated by ‘giving’ to the god. There is no curse-tablet parallel for this idea of ‘redeeming’ iras, ‘(causes for) anger’, ‘provocations’, but it is understandable: note OLD s.v. redimo 8, ‘To get rid of by monetary payments or other concessions, etc., to buy off, citing Publilius Syrus, Sent. A19, ab amante lacrimis redimas iracundiam (buy off your lover's anger with tears).

b2. redemere: for Classical redimere, a back-formation from emere, like redemat in Tab. Sulis 65. 10 and 99. 6. possit: like the formulas in a5–6, the person is singular, even though the syntax of illis (etc.) would require the plural. The author presumably copied it without thinking from a standard formulary of some kind.

b3. nessi: this Vulgar form of nisi is also found in Tab. Sulis 65. 10. The scribe seems to have first written NS, before converting the S into E by a horizontal ligature to the following S.

suo: O is apparently written over another letter, perhaps E.

AENEI: the text could perfectly well have ended with nessi sanguine suo, as the same formula does in Tab. Sulis 65, so there is no knowing what followed. There are possible traces of one more line (b4), but thereafter the tablet is uninscribed. A vertical line has been scored on the extreme right, probably when the tablet was being cut.

⁸ Subscript dots mark ‘doubtful’ letters which are incomplete, but where the reading is reasonable even without knowledge of the context. Letters more doubtful still are represented only by a stop, and are restored within square brackets in the Reconstructed text. For details see the commentary above. The transcription observes the scribe's intermittent word-separation, which he marked by leaving a space before et male (a3), si liber (a5) and redemere (b2), or by lengthening the final E in sedere (a8), redemere (b2) and sanguine (b3) [contrast male faciunt in a3]. Note also that he avoided word-division at the end of a3, a7, a8, b2, preferring to extend the horizontal stroke (of T, C or E) to mark the end of the line. The only certain instance of word-division is per|mittas (a6–7), although in b3 the question remains open.

⁹ With the two previous items. See Woodward and Leach, The Uley Shrines, 129, No. 49.

¹⁰ The letters cut the layout lines, which therefore preceded them, but the scribe may have scored them line by line as he wrote, since the last two layout lines are closer-set, as if to anticipate the smaller letters that were to come.

¹¹ Commentary Curse-tablet lists of names are all in the nominative case, or occasionally in the accusative (e.g. Tab. Sulis 53). The mixture of cases in lines 1-7 must therefore be significant: the sequence is nominative masculine, nominative feminine, nominative masculine, a verb, accusative masculine, nominative feminine, accusative masculine. It looks as if the nominative names are all cursing (‘giving’) the two accusative names.

1. Aunillus: the first letter looks like the A in Minura (6), the second letter has the first diagonal of V but no sign (except the appropriate space) of the second. For the ‘Celtic’ name Aunillus cf. CIL XIII.1210 and 3282 (Aunilla) and V. 5374 (Aunillius).

2. V[ica]riana: the traces suit VIC to begin with, and exclude Valeriana; but there is not enough space for Victoriana. So read V[ica]riana, which is not found in Kajanto Cognomina, but is an acceptable development of the Latin cognomen Vicarius.

3. Covitius: there is no sign of the leftward extension of the crossbar of T, but C can be excluded since in this hand it is made in a single curving stroke. The name Covitius is not attested, but Covirius cannot be read.

4. The second letter is damaged, but lying between M and N it must be a vowel; E is the only other possibility, but in view of the frequency of Minius / Minnius, MINI can easily be understood as the patronymic of Covitius (3). Only a vertical stroke survives of the last letter in the line; the sequence DONA requires some part of the verb donare, but there is no trace either of the crossbar of T (for donat) or the diagonal of N (for donant). Donare is frequent in curse- tablet texts in the sense of ‘give’, i.e. devote to divine vengeance, and the formulaic dona[t] is the likeliest reading, the three persons in 1–3, like Minura in 6, being each regarded as the subject.

5. Varicillum: like Varianus at Uley (Britannia xx (1989), 329Google Scholar, No. 3), this looks like an elaboration of the Latin cognomen Varus, but it really conceals the Celtic name-element -varus. Varicillus is found in CIL XIII.4301, Elvorix Varicilli f(ilius), where it is clearly a ‘Celtic’ name.

6. Minura: the reading is certain. Holder notes Appian, Hisp. 71, Minurus a friend of Viriathus in Lusitania; but we have not found the instance of Minurius attributed by Mócsy Nomenclator to Gallia Belgica.

7. Atavacum: again the reading is certain, but the name Atavacus seems to be unattested. If it is a ‘Celtic’ name, it is not necessarily connected with Atavus/Atavius in central Italy (CIL IX.5022, X.6347), but Holder cites a gold coin of the Arverni reading ATAV.

8–9. Inscribed in a different script and now rather faint; there are some recognizable letters (e.g. V and N in 8).

¹² With the three previous items. See Woodward and Leach, The Uley Shrines, 128, No. 33.

¹³ Commentary The tablet names two persons, a man and a woman, each identified by a second name in the genitive case, probably their patronymic (2) or matronymic (5-6). For another matronymic see Tab. Sulis 30.1, Severianus fil(ius) Brigomall(a)e. The names are all of Celtic etymology.

1. Lucilia: not the Latin nomen, but one of the Lucius names which conceal a Celtic name-element, popular in Britain and Gaul; for Lucilia as a cognomen see RIB 958, CIL XIII.8820, V.5444.

2. Mellossi (filia): Kajanto, Cognomina, 284, notes two instances of Mellosus, both in Africa; but Mellossus here is more likely to be cognate with Mellonius which is well attested in CIL XIII (and note RIB 1665, Melonius Senilis from Upper Germany). Mellossi (genitive) is ambiguous: Mellossus could be Lucilia's father or her husband, but in view of the matronymic in 5–6 and the possibility that Lucilia and Minuvassus are husband and wife, it is better to take Mellossi as a patronymic.

3. Damaged by the corrugations, but initial A and final -MO are certain. To judge by this ending, and the absence of a patronymic, the word is not a personal name; perhaps aestimo ('I reckon') was intended, but this verb is not found in curse tablets.

4. Minu(v)assus seems to be unattested, but Minus/Minius/Minutus is found in Gaul and Britain, while Celtic vassos (‘vassal’) is a frequent name-element. The intervocalic consonantal u [v] has been lost next to u (vowel), as often in Vulgar Latin.

5–6. Only LL is uncertain. The name Senebellenus/a seems to be unattested, but again it unites two common Celtic name-elements: Seno- (as in Sennius, Saenus, Senovarus, and especially Senecio etc.) and Belenos/Bellinus.

¹⁴ In demolition debris at Bath or Ilchester, according to the Bath antiques dealer from whom it was bought by Simon Bendall. The stone is a coarse shelly limestone containing crinoidal debris but no ooliths, which has been identified as coming from Ham Hill (5 miles south-west of Ilchester), not from Bath, by John Senior of the Department of Geology, University of Durham. The lettering is ancient, and there is no reason to doubt that it is genuine; the best reason for accepting the uncertain attribution to Bath is the link with RIB 153 (see next note). Mr Bendall made the altar available to John Casey, who provided photographs and other details, and discussed it with RSOT.

¹⁵ In 1, M is ligatured to E, and the ligature of I to N is indicated by the lowering of the diagonal stroke. It should really have been attached to the third stroke of N, forming a cross with the cross-bar of T. The line-width can be deduced from the formula in 2, from which it follows that filius was abbreviated to FIL (as in RIB 153) and posuit to POS. The ligatures in 1 confirm that the inscription was condensed.

The text is closely related to RIB 153 (Bath), NOVANTI FIL I PRO SE ET SVIS I EX VISV POSSVIT, and is the fifth instance from Britain of the formula ex visu (the others being RIB 760, 1778 and JRS lvii (1967), 203, No. 5). The deity is not named, and the dedicator is identified only by his patronymic CLEMENTI FIL, just as in RIB 153 the dedicator is only NOVANTI FIL. (RIB 153 is justifiably thought to be part of a larger text, but the present item raises the possibility that it is complete after all.) The name Novantius in RIB 153 is a ‘late’ (third-century?) formation like Clementius, a name in -tius formed from a present participle or an adjective resembling it; but the parallel is not exact, since Novantius may be Celtic in origin (cf. CIL XVI.160 (Leicester), Novantico) whereas Clemens is a common Latin cognomen.Google Scholar

¹⁶ During excavation for Tyne and Wear Museums directed by Dr N. Hodgson and Mr P. Bidwell, for which see above, p. 344. Dr Hodgson sent photographs, a squeeze, and full details.

¹⁷ Dr Hodgson points out that the defences were extended during the Severan conversion of the fort into a supply- base, and that this is the second centurial stone of the legion to be associated with the extension. The other is JRS lii (1962), 193, No. 13, leg(ionis) VI Vic(tricis) | (centuria) Paterni.Google Scholar

¹⁸ During building-work, when it was sawn into three pieces for ease of removal. It was subsequently identified by Peter Ryder of Riding Mill, in whose garden it was examined with the help of Georgina Plowright. It will go to Corbridge Roman Museum.

¹⁹ The reading of I is uncertain: there is a diagonal followed by a curving stroke, appropriate to part of VO. In 2 the horizontal of L is lost. AEL suggests the name Ael(ius), but note RIB 1131, dedicated to Caelestis.

²⁰ In the garden rockery, from which it was excavated by Lindsay Allason-Jones and gardeners of Newcastle upon Tyne University. It is now in store at the Museum of Antiquities, Newcastle upon Tyne.

²¹ When the cellar floor was lowered by 0.50m. The spoil removed was mostly the black organic silt typical of first- and early second-century deposits in Carlisle, and contained a dupondius of Vespasian and a sestertius of A.D. 107. The spearhead remains in the possession of Messrs Simtor of Scotby. Ian Caruana sent a drawing and full details.

²² Presumably the abbreviated name of the owner, containing the Celtic name-element sen- found in Senecio, Senilis, etc.

²³ During excavation for the Department of the Environment directed by John Casey, for which see Britannia v (1974), 413–14, and vi (1975), 235. This and the next item were made available by Alexandra Croom of South Shields Roman Museum, in advance of the final report.Google Scholar

²⁴ A feminine personal name, e.g. [Iu]lia (compare RIB 1483); but there are other possibilities.

²⁵ A feminine personal name developed from Claudius.

²⁶ By metal-detector. It was bought from a London antiquities dealer by the present owner, who made it available to RSOT. We are grateful to Michael Still, who is writing a University of London thesis on leaden sealings, for his comments on this item and No. 27, Stert (?), below.

²⁷ Reading and interpretation were prompted by two leaden sealings from Carnuntum published by G. Dembski in Römisches Österreich iii (1975), 59, Nos 19 and 20. Clockwise round Victory in the same pose they read: AAGG – NN. The pose resembles that of Victory on coin-reverses of Septimius Severus: see H. Mattingly, Coins of the Roman Empire in the British Museum, V, Pertinax to Elagabalus (1950), esp. pl. 54. 9, 10 and 19, VICTORIA BRIT. The only Victory sealing previously recorded from Britain (RIB 2411.32) is not comparable, but see now No. 27, Stert(?), below.Google Scholar

²⁸ After Severus' promotion of Caracalla to Augustus and Geta to Caesar (28 January 198) but before Geta too became Augustus (October/November 210); for the dating see A.R. Birley, The African Emperor: Septimius Severus (1988), 130, 186 and 218. Stylistic considerations apart (see previous note), this 13-year period is far more likely than the few months in 238 and c. 258 when there were again, briefly, two Augusti and a Caesar.

²⁹ During excavation by the Museum of London Archaeological Service directed by Nick Bateman, for which see above, p. 361. Mr Bateman provided details and a drawing by Damian Goodburn, whom we thank for discussing this item with us.

³⁰ The impressions were so deeply struck (up to 5 mm) that it seems impossible to interpret (a) as [T]ICLV, Ti(berius) Cl(audius) V(…), the absence of the initial letter being explained by insufficient pressure when the die was applied. It is just conceivable that the absence of a T is due to damage to the die. The iron stamps would have been similar to RIB 2409.17, the die (not a branding-iron) found in Tokenhouse Yard, London EC2.

³¹ By Mr Ray Ludford with a metal detector. David Gurney of Norfolk Landscape Archaeology provided full details, including a photograph by David Wicks and the drawing by Sue White.

³² During excavation for Tyne and Wear Museums and English Heritage directed by N. Holbrook, for which see Britannia xxiii (1992), 269. Details from Paul Bidwell were sent by Alexandra Croom. Georgina Plowright made it available at Corbridge Museum.Google Scholar

³³ Not T(iti) D(…) D(…). Most, if not all, of the reverse-legends of the ala sealings in RIB 2411.81-90 are abbreviated names with D for ‘decurion': this is explicit in 89, VAL DEC, and is certain in 90, FL SUM D, cf. 81, D I L VAL S; 83 MCI | D; while 84, HRD and 88, IPD, must also be explained in the same way since they cannot be the initials of tria nomina. (To complete the list: 82 and 86 both end in D but present problems of reading, while 85 is anepigraphic, and 87 with D on the obverse looks like two decurions, V(aleriorum) Q(… et) V(…).) Many of the reverse-legends of cohort sealings which are not explicitly centurial also terminate in D, perhaps the abbreviated names of decurions in cohortes equitatae, but the question cannot be pursued here. For other sealings of the ala II Asturum see RIB 2411.82 (Corbridge) and 83 (Carlisle). It was the garrison of Chesters by the reign of Commodus (RIB 1463), and remained there throughout the third and fourth centuries.

³⁴ During excavation by Birmingham University Field Archaeology Unit directed by Peter Leach: see Britannia xxii (1991), 279–81. Jane Evans made this and the next item available to RSOT. They will be published with three other graffiti too brief for inclusion here, in the final report on the excavation.Google Scholar

³⁵ Only the horizontals survive of F, with possible trace of the vertical stroke; there is damage to the left of it, but it may be the first letter. The graffito is presumably the potter's signature, part of a personal name or place-name.

³⁶ During rescue excavation directed by Tom Blagg for the Department of the Environment. Fiona Seeley of the Archaeological Section of Suffolk County Council sent details and drawings (by Donna Wreathall) of this and the next item.

³⁷ For this inscription compare a gold double ring in the British Museum from the Franks Bequest, inscribed on one bezel DVLCIS and on the other DVLCI(S), the letters apparently produced with small punches as here. See F.H. Marshall, Catalogue of Finger Rings … in the British Museum (1907), 100, No. 586. AMOR DVLCIS, ‘sweet love’, is found on the bezel of a silver ring from Castel Collen (RIB 2422.19).

³⁸ By Messrs R. Gearing and H. Jeffries with a metal detector.

³⁹ The letters probably would have been filled originally with red and green enamel, like the other examples of this class of object. There are three of them, two from South Shields (RIB 2429.13 and 14) and one from Chester (RIB 2429.15), and they also are incomplete. From RIB 2429.14 and 15 it can be seen that one end of the terminal was furnished with a flattened loop. The present item is the first to be found in Britain where the other end of the terminal, furthest from the loop, is preserved. But the words utere and felix are not followed by crescents, unlike in RIB 2429.13, where alone the complete legend is preserved.

⁴⁰ With the next three items during re-excavation directed by M r F. Aldsworth, former Archaeologist for West Sussex. David Rudling, Director of the Field Archaeology Unit of University College London, provided details of this and the next five items, including drawings by Jane Russell.

⁴¹ Other pedalis bricks with the finger-drawn graffito LCC were found by Samuel Lysons during the original excavations of 1811–18: see RIB 2491.102. So too was RIB 2491.41, reading […]CCI. Since these and the present items all seem to be variations on the same ‘LCC text, it cannot be simply explained as the initials of tria nomina or a batch-total (‘250’).

⁴² There is a vertical below and prolonging the vertical of the first L, which could be interpreted as the top of I or L, but may not be significant at all.

⁴³ The imprint of a nailed shoe or sandal partly overlies the final C.

⁴⁴ During trial excavations by the Field Archaeology Unit of University College London, directed by David Rudling for the West Sussex County Council and the Trustees of Bignor Roman Villa.

⁴⁵ During excavation by the Field Archaeology Unit of University College London, directed by David Rudling.

⁴⁶ And brought to Devizes Museum, where the Curator, Paul Robinson, sent a drawing and other details to RSOT. Its present location is unknown. For other examples see RIB 2421.5–40.

⁴⁷ By metal detector. The finder brought it to Devizes Museum, where the Curator, Paul Robinson, made it available to RSOT. It has now been bought by the Museum.

⁴⁸ The only parallel from Britain is the large sealing from Ancaster(?), No. 12 above, which is probably Severan (198–210). The pose and legend of the Stert(?) sealing are reminiscent of a reverse coin-type of Elagabalus (218–22): see H. Mattingly, Coins of the Roman Empire in the British Museum, V, Pertinax to Elagabalus (1950), pl. 85.14 and 15, and pl. 87.11, VICTOR ANTONINI AVG. The Stert(?) sealing might read VICTOR-[IA…], but this would mean breaking the legend at the bottom of the die; it is more likely that the die-cutter reversed the coin-legend, making it read perhaps (clockwise from 7 o'clock) [AVG ANTONINI] VICTOR.

⁴⁹ During field-walking by the artist Helen Anderson, who made it available to RSOT. It will probably go to the Grosvenor Museum, Chester.

⁵⁰ This item is interesting for its variety of contemporary cursive hands, and for the very rare word sumtuaria which introduces a new type of military document. It notes the ‘expenses’ of three workmen who sign their names; whether they were auxiliary soldiers (cf. RIB 2491.96) or legionaries of the Twentieth Legion is uncertain. A line-by- line commentary follows:

1. sumtu[aria…?] cf. 5, sumtuaria […], evidently a neuter plural. There may have been a qualifying adjective, but this seems unlikely. 1 and 5 are written by different hands. Like the rest of the text, the handwriting is Old Roman Cursive, but note S in 5, which was made without lifting the writing-point and thus approaches the New Roman Cursive form (see Tab. Sulis, pp. 90 and 94).

Sumtuaria (for Classical sumptuaria) is a plural noun from the adjective sumptuarius, which derives from sumptus, ‘expenditure’; in a late-Roman military context (Vegetius Epitoma (I.20) sumptus can mean ‘spending money’, but the only other instance of the noun sum(p)tuarium seems to be the fragmentary pay-sheet of A.D. 72 from Masada (H.M. Cotton and I. Geiger, Masada II, The Latin and Greek Documents (1989), 35, No. 722). Here the sumtuarium of 20 denarii is a standard deduction equivalent to the 80 drachmas [= 20 denarii] in victum of P. Gen. Lat. 1, and is translated by the editors as ‘food expenses'. This would seem to be its meaning at Holt, or perhaps ‘expenses’ in general; at any rate, the living expenses of soldiers outposted from their base. Perhaps they received an allowance in lieu of rations.

2. Iuniu[s…]. The first letter resembles S, but the diagonal stroke was made first, and must therefore have been a casual mark already there. The second V is cut by a diagonal apparently from the line below; perhaps of an exaggerated (and thus initial) A or M. A cognomen may have followed lunius, but the nomen is found on its own, cf. RIB 2501.274–6.

3. The numeral is made by ligaturing each digit to the next; the fourth has an exaggerated downstroke as if it were the last, as one would expect. (S for semis cannot be read.)

4. Maternu[s]. The cognomen is frequent (eight instances in RIB, and more in RIB 2501.352–60), perhaps because, like Maturus, it recalled the Celtic name-element mato-. The diagonal stroke above (see note to 2) may have been the initial letter of Maternus’ nomen, but this is far from certain.

6. Bellettus. The handwriting is idiosyncratic (it is obviously the man's own signature). The first letter looks more like B than D, and Bellus and its cognates were popular in Britain and Gaul; the second L has been written above the first. The form Bellettus seems to be unattested, but Belliccus cannot be read.

⁵¹ With the next two items during excavation in advance of road-widening, directed by N. Jones for the Clwyd- Powys Archaeological Trust, for which see Britannia xxii (1991), 224. Wendy Owen provided details, and made the three items available.Google Scholar

⁵² The Celtic name Boudus is well attested in CIL XIII and is cognate with British Boudic(c)a. At the likely date of this graffito (pre-Antonine), there was only one Cohort of Asturians in Britain, the Second; in the Flavian period it was active in central Wales, at Llanio (RIB 407-8). See Jarrett, in Britannia xxv (1994), 52–3.Google Scholar

⁵³ Despite the unusual position of the graffito, this is probably still an ownership inscription: an abbreviated personal name such as Domitius or Donatus. Do, ‘I give’, is possible, but seems too allusive.

⁵⁴ The stone clearly reads AQ, which Haverfield expanded to a(t)q(ue), but there is no warrant for such a contraction; and even if a connective were needed between militavi and nunc hic sum (and it is not), ac: would have been enough. By contrast, AQ is the regular abbreviation for aquilifer in the various epigraphic lists of legionaries, as well as in the tombstones CIL VIII.2782 and 18311 (both AQ LEG III AVG). (See Speidel, M.P., ‘Eagle-Bearer and Trumpet’, Bonner Jahrbücher clxxvi (1976), 123–63, and especially the references collected at 143–5.) To be described as miles of the legion and then as aquilifer is unusual, but CIL III.15005, 1, though the reading is not quite explicit, seems to be a parallel. Only the inscribed base and sandalled feet survive of Flaminius’ sculptured tombstone, which depicted him as a legionary standing in a niche holding what RIB calls a ‘staff’; but this ‘staff’ was evidently the shaft of the eagle standard, his pose being identical with that of Cn. Musius at Mainz (illustrated in Speidel, op. cit., 137, pl. 6), who was the legion's aquilifer just before it came to Britain. RSOT.Google Scholar

⁵⁵ The only other one is Vittius Adiutor, aquilifer leg(ionis) II Aug(ustae) in Tab. Vindol. II, 189, No. 214.

⁵⁶ Eric Birley prefers cho(rtis) […’vanianorum (in Temporini, H. and Haase, W. (eds), Aufstieg und Niedergang der römischen Welt, xviii.l (1986), 40Google Scholar, n. 166), i.e. reading CIHO in 3–4, which accords with surviving traces. This would be an unknown auxiliary cohort. There is no reference to it in P. Holder, The Roman Army in Britain (1982), or Jarrett, M.G., ‘Non-legionary troops in Roman Britain: Part One, The Units’, Britannia xxv (1994), 35–77.CrossRefGoogle Scholar

⁵⁷ Taking O to be ligatured within C., and L to be lost at the end of the line; in 4 there is sufficient trace of G. For the layout cf. RIB 1268, COL |LEGI (the worshippers of Minerva at High Rochester). Note also RIB 2102 and 2103 (the worshippers of Mercury at Birrens). For a col(legium) dei Sil(vani) in Britain see RIB 2422.52; and for many elsewhere, see P.F. Dorcey, The Cult of Silvanus (1992), ch. 5, ‘Cultic Structure and Organization'. There is no other instance of the collective term Silvaniani, but like terms are coined for other gods: Ioviani (in ILBulg. 273, brought to our notice by Iveta Mednikarova) and (in J.P. Waltzing, Corporations IV, 180 ff.), Apollinares, Dianenses, Herculani, Martenses/Martiales, Mercuriales, etc.

⁵⁸ We owe this correction to Michael Still (see note 26 above). The cognomen Titianus is not uncommon, but compare RIB 2410.7 (Chester), (centuria) Titiani, and RIB 593 (Ribchester), (centuria) Titiana. TD also occurs on RIB 2411.94, where it may be equivalent to TVD on other sealings (see note ibid.); where the cohort was not equitata, the possibility of a decurion (see note 33 above) does not arise.

⁵⁹ By Dr M.C. Bishop, who publishes it with annotation in M.C. Bishop, Corstopitum: An Edwardian Excavation (1994). 37. fig. 29.

⁶⁰ Another photograph of the fragment, with details of discovery, was published by Haverfield in Arch. Journ. xlix (1892), 199Google Scholar (whence EE ix.1270). The numeral ‘40’ apparent in the Corbridge photograph is not a reference to Site 40 at Corbridge, but is in fact ‘49’, its number in Haverfield's Carlisle Catalogue (CW xv (1899), 484).Google Scholar

⁶¹ The fragments are at present in Manchester Museum, where Dr John Prag provided information. They are being examined by Dr Paul Holder, by whom they will be published in the Bulletin of the John Rylands Library, forthcoming. We owe this preliminary account to him; for this cohort at Ravenglass, cf. RIB 2411.94.

⁶² The diploma belongs to the same issue as RMD 184, which is identical in style and date. It will be published by Margaret Roxan, to whom we owe these preliminary details. By being issued to a cavalryman, it confirms that the cohort was equitata.

⁶³ Both faces of the diploma read HIERO, so Horsley's drawing must have overlooked an I ligatured to H or E; it has certainly missed another ligatured I in DOMITIVS. The form Hieron is closer to the Greek than Hiero, but either is acceptable. Unfortunately the problem remains of lines 2–3. In 2 read COH(ORS) with Davies, Ep. Stud, xii (1981), 206–7, who rightly rejects the ‘dative of advantage’ read by RIB but omitted from its transcript. In 3 the new diploma supports the reading EQ(VITATA), but C(VI) P(RAEEST) or similar is also required.Google Scholar

⁶⁴ See JRS liv (1964), 180, No. 16, and the 1994 Annual Review of the National Art Collections Fund, 152, with colour illustration, which Professor S.S. Frere brought to the attention of RSOT. It was sold without provenance for £460 as Lot 119 in Sotheby's (London) sale of 28 October 1993 to Rupert Wace, who sold it to Dr Nicholas Reeves, from whom Louth Museum bought it for £1,750.Google Scholar

⁶⁵ For another trulla stamped with the name of a military unit see RIB 2415.39 (ala I Thracum). The Fourth Cohort of Breuci is first attested in Britain by the diploma of 122 (CIL XVI.69), and was active at Bowes in c. 130/33 (RIB 739); it was making stamped tiles in the first half of the second century at Grimescar, West Yorks., which have been found at Castleshaw, Slack, Castleford and Aldborough (RIB 2470.1–2). In the third century it was at Ebchester (RIB 1101 and 2470.3). Its history before 122 is unknown, but this trulla suggests it was active in Lincolnshire.