² During excavation for the Department of Greater London Archaeology, Museum of London, directed by Bruno Barber and David Bowsher who together provided details. See above p. 309.

³ Licetus as a cognomen or ending of a cognomen appears to be unmatched. In line 5 -retio would appear to be the ending of the name of the deceased's origo since reference to Kajanto, The Latin Cognomina, shows that all of the dozen names ending -etius listed in the reverse index of cognomina in Solin and Salomies, Repertorium Nominum, are excessively rare. Arretium was enrolled in the voting-tribe Pomptina and supplied recruits to the Roman legions, e.g. CIL iii 14349/1 in leg II Adiutrix c. A.D. 86–100; ibid. 2840 and 6418 in leg XI before AD 42; ibid. 2071 and 8764 in leg VII, AD. 9–42, so that it is not surprising to find it as an origo. It is, however, unusual for the voting tribe to follow, rather than precede the cognomen, though there are a few cases, e.g. RIB 482,508, 525.

⁴ During excavation directed by Dr Ann Ellison for the Committee of Rescue Archaeology in Avon, Gloucestershire and Somerset. See Britannia xix (1988), 485, No. 2, with ibid., n. 4. Eighty-seven lead tablets from the site bear some sort of inscription, and are the subject of an interim report by RSOT in the final report on the excavation, forthcoming, by Dr Ellison (now Woodward). This item is published there (No. 5) as here; the next item (No. 43 there) is published here for the first time. The interim report also discusses the condition of the Uley tablets, their similarity to those from Bath, the god worshipped at Uley whose Celtic name or title they preserve, and some of the categories collected in the introduction to Tabellae Sulis (see next note): formulas, language, stolen goods, handwriting, and names.

⁵ Commentary

(Tab. Sulis refers to RSOT'S publication of the Bath curse tablets in Tabellae Sulis (1988), by page (p.) or tablet-number, printed as a fascicle of B.W. Cunliffe (ed.), The Temple of Sulis Minerva at Bath, II: Finds from the Sacred Spring (1988). The published tablets from Uley are cited as ‘Uley’ and by their number in the interim report (see previous note), where Nos. 1–3 = Britannia x (1979), 340–5, Nos. 2–4 (revised), and No. 4 = Britannia xix (1988), 485, No. 2.)

1ff. nomen juris (etc.): for the formula, cf. Tab. Sulis 16, nomen furis qui [involaverit] donatur.

2. frenem: the reading is unavoidable, neither E (and certainly not the second) resembling V elsewhere in this text. The scribe surely intended frenum (‘bridle’), but either his eye was caught by the second syllable of nomen immediately above, or, more likely, he confused his declensions in a way typical of ‘Vulgar’ Latin.

3–4. These formulas are common in British curse tablets: see Tab. Sulis, p. 67.

4–5. donator: the reading is certain; since the (legal) term donator (‘donor’) is excluded by the syntax, and since the formula requires donatur, this is clearly what was intended. The divergent spelling may be due to a tendency by British Celtic-speakers to confuse Latin ō and ū: for other instances, see Tab. Sulis 102 and ibid., p. 75 [continued p.330].

5. duas partes: duas has been written over an earlier text, presumably a mistake. The reference to ‘two parts’, i.e. two-thirds (cf. 6, tertia), suggests a formula by which a proportion of the stolen property is ‘given’ to the god to secure his intervention: cf. Uley 2.10–11, tertiam partem [d]onal (repeated, 15). The formula is not found at Bath, but also occurs in RIB 306 (Lydney) (one-half), Britannia xv (1984), 339, No. 7 (Pagans Hill) (one-half), and JRS liii (1963), 123 (Ratcliffe-on-Soar) (one-tenth). However, the syntax here is defective.

6. AFIMA sua: presumably a copying error, letters omitted or mis-read; a femina sua (‘from his wife’) is a plausible restoration, but is not certain.

6–7. tertia: this again ought to be the proportion of the stolen property ‘given’ to the god (cf. 5, with note), but the syntax is defective. The scribe seems to have miscopied a formula (see also next note).

7–8. ad sanitatem: as it stands, this (like the whole of 5–8) makes no sense. The scribe may have confused two formulas (for which see Tab. Sulis, pp. 65–6 and 68), the interdiction of health (sanitatem) to the thief until his return of the stolen property to the temple (ad templum).

⁶ See Tab. Sulis 10. The only significant difference is that the Bath 5 is formed in the same way with three distinct strokes which however do not intersect. It will be seen from the tabulation on p. 91 of Tab. Sulis that the script is unique at Bath, which makes its presence at Uley still more remarkable. The identification is reinforced by the language: not by the coincidence of the somnum permittere formula, which is quite common, but by the coincidence of maximo leto adigas (see next note), which is unique to these two texts. If the same man wrote both, as seems likely, it raises an important question of authorship. It was argued in Tab. Sulis (pp. 98–101) that the diversity of hands suggests that petitioners inscribed their own tablets rather than resorted to a professional scribe. At first sight it now looks as if the same ‘scribe’ was working both at Bath and at Uley, but it must also be noticed that Tab. Sulis 10 was addressed to Sulis by ‘Docilianus Bruceri (films)’ whereas this tablet was addressed to Mercury by ‘Docilinus’. ‘Docilinus’ and ‘Docilianus’ are almost identical and may well have been used by the same man, a little vague about Latin terminations. Names cognate with docilis were popular at Bath, and it is likely that they were romanised forms of the locally-attested Celtic name Docca: thus the man may have thought of himself as ‘really’ Docca. On the other hand, docilis-names are too common for this to be certain. But it would be quite a coincidence to find the only pair of tablets from the same hand (Tab. Sulis 95 and 96 being virtually a continuous text) were written by an anonymous scribe for two men with almost identical names. If, therefore, the two tablets were indeed written by the same man on his own behalf, they provide an interesting link between the great urban shrine of Sulis and the rural temple of Mercury.

⁷ Commentary

2. Docilinus: see previous note.

QVAENM: the tail of Q is unusually long and might belong to a lost L; there is a possible space after N, but no sign of any lost letter; the final M is damaged. Otherwise the reading is certain, but it makes no sense, nor does the easiest restoration, quaen(a)m. The context demands a patronymic (cf. Tab. Sulis 10.2) or a verb like queritur, but neither can be discerned here. In any case, the syntax is faulty, there being no main verb of which Varianus (etc.) would be the subjects.

3–4. Three common Latin cognomina.

4. pecori: with its elaborate serif, P superficially resembles B. The ‘beast’ might be sheep, goat, pig, or bovine. Unpublished tablets refer to the theft of wool and cows.

5–6. dolum malum intulerunt: the phrase has not occurred previously on a curse tablet, but enough survives of the dotted letters not to doubt the reading, and dolus malus is a common legal phrase (see Vocabularium lurisprudentiae Romanae (1933), s.v. dolus). It is thus in keeping with the quasi-legal language of British curse tablets (see Tab. Sulis, pp. 70–71). This tablet, however, is unusual in not being inspired by theft. The ‘harm’ done is not specified (unless in 6–7), but was presumably an ailment blamed on persons known to bear Docilinus a grudge, whether it was thought to be due to poison or to witchcraft. (The word venenum, ‘poison’, covers both senses, and reminds us that the distinction was not necessarily made.)

6–7. INT.RR[.] prolocuntur: this phrase surely amplifies or explains the ‘harm’ done, which makes its obscurity unfortunate. The first word(s) is damaged by corrosion. PRO is to be taken with LOCVNTVR, since a passive locuntur (‘are placed’) is inappropriate to a complaint of the villainy of Varianus (etc.); instead, this is easily taken as a ‘Vulgar’ spelling of proloquuntur (‘are pronouncing’), cf. com[o] (quomodo) and cus (quis) (Tab. Sulis, p. 76). The preceding word(s) is most easily restored as in t[e]rr[a] (‘on earth’ or ‘in earth’), but the phrase in terra proloqui seems to be unparalleled. Possibly it was a technical term, used of casting a magic spell.

8. max[i]mo [le]to adigas: this formula is otherwise found only in Tab. Sulis 10.10–12, where the text must now be restored as ut [e]um dea Sulis maximo letum (an error for leto) [a]digat.

8–10. For the interdiction of health and sleep until restitution is made, see Tab. Sulis, pp. 65–6.

11ff. The formula(?) has not occurred before, but the idea of ‘buying back’ with one's blood a ‘gift’ (of thief, or object stolen) is common: see Tab. Sulis, p. 66, s.v. redemat. 11 and 12 are both damaged by corrosion, but the traces are sufficient for the restoration proposed. The tenses could be future perfect or perfect subjunctive: see Tab. Sulis, pp. 69–70.

⁸ During excavation for the Dean and Chapter of Carlisle Cathedral directed by Mr G. Keevill, with assistance from Carlisle Archaeological Unit, which made the stone available by the agency of Tim Padley.

⁹ The first O of Concordiae was ligatured within the C, but no trace of it now survives. There was no suprascript bar above the numerals. The dedication to Concord implies that detachments of both legions were stationed at Carlisle, cf. RIB 1125 (Corbridge). Britannia xiii (1982), 410, No. 13, a fragment of a dedication slab found at Annetwell Street, Carlisle, which reads V[…], probably records a legionary vexillation. The evidence is discussed more fully by RSOT in Cumb. Westm. lxxxix (1989)Google Scholar , forthcoming, but may be summarised here. Stamped tiles of the Second and Twentieth Legions were found at Annetwell Street, in association with barrack-like buildings from the (military?) redevelopment of the fort (Britannia xiii (1982), 83, and Current Archaeology No. 101, 176–7). Both legions are also attested at Carlisle by (undated) inscriptions and carved stones, the Second by an inscribed fragment (Britannia xix (1988), 491, No. 8) and a possible Capricorn (CSIR I 6, No. 535), the Twentieth by a fragmentary altar (RIB 954), a fragmentary dedication slab from Annetwell Street (Britannia xvii (1986), 437, No. 11), and carved stone boars (CSIR 16, Nos. 524–6). Detachments of the two legions seem to have assisted the garrison of Netherby with building in A.D. 219 (RIB 980 with A.R. Birley, Fasti, 187), perhaps identical with those active at Maryport (RIB 852). It may therefore be conjectured that the tribune of the Twentieth who dedicated the altar which is the next item was the commanding officer of the two detachments (whence the clasped hands on his altar, the symbol of concord); and that in the third century detachments of both the legions of Upper Britain were stationed at Carlisle, as well as at Corbridge, in the Lower province. After the division of Britain in c. A.D. 213, when the northern frontier lay within the Lower province, its garrison, the Sixth Legion, may have been over-extended; a shortage of legionaries, and perhaps under-employment elsewhere, is suggested by the use of legionaries from the Upper province as beneficiarii consularis (RIB 745, 747(?), 1696), and by the presence of German legionaries commanded by a centurion from Upper Britain at Piercebridge in c. A.D. 217 (JRS lvii (1967), 205, No. 16, with RIB 1022, 1026, Britannia xvii (1986), 438, No. 20). The transfer of detachments to Carlisle and Corbridge would tend to equalise the legionary garrisons of the two British provinces, more in line with the relative importance of the Welsh Marches (two legions) and of the north (one legion only). This might be reflected in the archaeology of Chester and Caerleon.

¹⁰ By Richard Annis in the course of a detailed survey by Carlisle Archaeological Unit on behalf of English Heritage. The stone has now been disengaged, and is on display in the room where it was found. It was made available by English Heritage, and other help was provided by the Unit. There is a description of the stone and its discovery by MrAnnis, in Cumb. Westm. lxxxix (1989)Google Scholar , forthcoming.

¹¹ There is a detailed commentary (‘A Roman altar from Carlisle Castle’) by RSOT in Cumb. Westm. lxxxix (1989)Google Scholar , forthcoming, whose salient points are summarised here. The deities, which are Roman and ‘Official’, can be restored from other dedications of the early third century. There is no exact parallel, but they suggest the vota ceremony of 3 January (see the Feriale Duranum of c. AD. 223/7), when the commanding officer, notably at Maryport in Britain, dedicated an altar on behalf of his unit. The dedicator's name contains only the second or third example of a ‘pseudo-tribe’ to be found in Britain (see RIB 506 and perhaps 1292), where the imperial gentilicium of the city of origin supplants the voting-tribe. This is a convention favoured by soldiers of Danubian origin, especially legionaries promoted into the Praetorian Guard in A.D. 193 and later. It identifies the city of origin as Ulpia Nicopolis (ad Istrum), already well attested as a recruiting area in the second and third centuries, and it strongly suggests that Syrio had been an evocatus of the Guard, like other officers already attested in the north-west in the third century (RIB 966, 988, 989, 1896). His career recalls that of his contemporary, the emperor Maximinus ‘the Thracian’. The presence of this Thracian-born legionary tribune in Carlisle in AD. 213/22 is interesting. Like other tribunes detached from their legions (e.g. RIB 1132 with Britannia iii (1972), 363), he is likely to have been the commanding officer of a legionary detachment; in fact, as the relief of the clasped hands suggests, the officer commanding a pair of detachments, from the Second and the Twentieth Legions. The evidence is collected and discussed in the note to the previous item, a dedication to the concord of these two units.

¹² By Alan Whitworth, in the course of his measured drawing of the Wall. He provided photographs and full details.

¹³ See Britannia xix (1988), 494–5, Nos. 14–22, and the items cited ibid, in n. 28.

¹⁴ By the Vindolanda Trust, in widening a doorway of the (1746) farmhouse, now part of the Roman Army Museum, where the stone remains. Robin Birley made it available.

¹⁵ Retrieved by John Moore of the Oxford Archaeological Unit. Full information and photograph from Susan Read and Leslie Cram of Reading Museum and Art Gallery where the object is now on display; see above, p. 319. For comparable lead tanks see Britannia xii (1981), 271–6.

¹⁶ During excavation by Carlisle Archaeological Unit on behalf of the Department of the Environment and Carlisle City Council directed by Mr J.A. Dacre. It was made available by Tim Padley, with whom and with Ian Caruana we have discussed the reading.

¹⁷ Commentary

Obv. 1. The nomen Gabius is rare (e.g. CIL v 1225; 2631), and it is possible that Gabiniana was meant, Gabinius being much more common. The adjectival form indicates the century's late commander, the command being temporarily vacant: see Birley, E., Roman Britain and the Roman Army (1953), 128–9.Google Scholar

2. lulii Suriti: the final I of lulii is crossed by a horizontal stroke which seems to be casual damage, since it apparently extends leftward, and the sequence -TS- is hard to accept. Initial Ts- is found in one or two unique cognomina, but it is easier to understand Suriti as the cognomen; Suritus seems to be unattested, but cf. Suricus (CIL v 4856) and Surica (v 5618), one of several cognomina cognate with Surus.

3. AR is presumably an abbreviation, but otherwise obscure.

Rev. The reading is uncertain, and the meaning obscure, apart from the denarii sign and numerals, which are found on other lead tags (e.g. Britannia vi (1975), 291, No. 39(a)).

This tag is not just a label of ownership, like those of centuries found at Caerleon URS xvii (1927), 216, No. 23) and Chester (JRS xxi (1931), 250, No. 12). In Britain it resembles the collection found at Usk (Britannia vi (1975), 291, No. 39), but unlike them, it lacks a reference to sarcina (‘package’) or to weight. A closer parallel is the latest tag from Caerleon (below, No. 60). Lead tags have been found widely in military contexts in the western provinces, but the reading is often difficult, and published readings are sometimes at variance with the drawings or photographs which accompany them. It is likely that tags were used for various purposes, as yet imperfectly understood. This tag arguably identifies clothing or equipment belonging to a soldier, for which a fee is payable on collection, whether for transport, repair, cleaning (understanding loto as perfect participle of lavo, ‘washed’), or whatever. (1½ denarii seems too small a sum to be the value of the item concerned.) Somewhat similar tags from forts in Raetia have been interpreted as labels attached to clothing which has been tailored: see Jahreshefte des Öst. Arch. Inst. xlvi (1961), 185–97 (R. Egger).

¹⁸ During excavation by the Carlisle Archaeological Unit directed by Ian Caruana. Tim Padley made it available.

¹⁹ During excavation by Exeter Museum's Archaeological Field Unit directed by P.T. Bidwell and J. Pamment. Information on this and the following item from Neil Holbrook of the Royal Albert Memorial Museum.

²⁰ -abri or -abr(i)i could theoretically be part of any name ending in -aber, -abrus or -abrius, in the genitive case, but reference to Mócsy, Nomenclator, and Kajanto, The Latin Cognomina, shows that Cantaber is statistically much the most likely. It occurs 23 times in CIL, with 11 examples from CIL ii (Spain). It is already attested in Britain (RIB 1568).

²¹ During excavations by Exeter Museum's Archaeological Field Unit directed by J. Pamment. See Britannia x (1979), 324; and Neil Holbrook, Roman Finds from Exeter 1971–9, forthcoming.

²² During excavation for the Trust for Wessex Archaeology and HBMC directed by Miss S.M. Davies and Messrs P.J. Woodward and A.H. Graham. Miss Davies made this and the next nine items available to RSOT, by whom they will be published, together with more fragmentary graffiti, in the final report. For an interim report on the excavation see Britannia xvi (1985), 306–7, and Proceedings of the Dorset Natural History and Archaeological Society, cvi (1984), 99–106.

²³ The traces that survive of two letters are consistent with I and A respectively. Between 5 and V there is trace of the decorative motif that marked the end of the text. The motto has already occurred, with similar but not identical decoration, on a beaker from York: RCHM, The City of York, I: Eburacum, 135, No. 151 (f). It is also well attested in CIL xiii 10018.206 (seven instances).

²⁴ Only the top half of the letters survives, and it is not certain that the initial letter is preserved. The spacing of the letters supports the reading of the two Ls, and the other traces are sufficient for the reading adopted here. The name would be a variant spelling of the feminine diminutive of the cognomen Albus. Albillus is found in CIL xiii.

²⁵ The I is so brief as not to be certainly a letter at all, even though it looks deliberate. But to disregard it and read Ula (which seems to be unattested) is harder than to understand the graffito as a barbarous Iul(i)a or Iul(l)a.

²⁶ Nutrix means ‘wet-nurse’, in which sense it is frequent in the epitaphs of CIL vi. It is rarely attested as a cognomen (e.g. CIL xii 4742), but is probably so to be understood here, since it is unaccompanied by a personal name, and there seems nothing to distinguish the flagon as ‘(Property) of the (or a) wet-nurse’. The elaborate ligatures were forced upon the scribe by his failure to begin far enough to the left. It is interesting to see them on a flagon dated by the Trust for Wessex Archaeology to the later first century; for another early example from Britain, see RIB 200.

²⁷ The second S is incomplete, but of much exaggerated height, as if it were the final letter of the text. Some distance to its right (the width of c. 4 letters but underneath the handle and thus probably uninscribed) is the top of another exaggerated descender, probably I or L (but H and K cannot be excluded), which would seem to be the initial letter of a text which ran round the shoulder of the jar. The cognomen Tacitus is common and its restoration almost certain, since the only other possibilities in Mócsy, Nomenclator, the cognomina Capacitus and Citus, are both very rare. The conjunction of two cognomina, Tacitus and Primus, must mean that two persons are being named; and it would seem that the original text was a list of c. 6 personal names.

²⁸ The first letter is damaged, and just possibly could be a heavily serifed V. Only the last letter is of cursive form, and could also be A, but this is excluded by the preceding vowels. A verbal ending in -eor seems unlikely, even [t]ueor (‘I preserve’); likewise a personal name in -eor, since Mócsy, Nomenclator, offers only the unique Resteer. Perhaps it is a ‘Vulgar’ spelling (i>e) of a comparative adjective used as a personal name, e.g. fortior.

²⁹ With the next five items by two divers, Bob Middlemass and Rolfe Mitchinson, directed by Raymond Selkirk, who made them available. They evidently belonged to a votive deposit of objects, including coins (A.D. 70–402), figurines and pottery, dropped from the bridge.

³⁰ AVE and AMA have been found separately on rings quite often (see CIL xiii 10024.39fr.); in Britain, on gold rings, AVE MEA VITA at Ribchester (CIL vii 1306) and AMA ME at Carlisle, (JRS liii (1963), 163Google Scholar, No. 23).

³¹ For other votive rings inscribed to Mars (MAR), one of them found in the Rhine at Mainz, see CIL xiii 10024.13. To them may be added the silver ring fromBenwell, DMT (JRS xxxvii (1947), 181, No. 12, as emended by RPW in 1954). Mars was probably the god of the confluence at Piercebridge, Mars Condates, cf. RIB 1024.

³² The same die has been found at Brough-under-Stainmore, but is now lost: Roach Smith, C., Coll. Ant. vi (1868), pl. xvi.9.Google Scholar The next three dies have not been found in Britain, and seem to be unparalleled elsewhere.

³³ During excavations for the Winchester Excavation Committee directed by Martin Biddle, see Britannia iii (1972), 349. Information on this and the following item from Mrs Prue Skinner of the Winchester Research Unit.

³⁴ During excavations for the Winchester Excavation Committee directed by Martin Biddle; see Britannia ii (1970), 283.

³⁵ During excavations directed by Professor W. Manning, Mrs R. Niblett and Mr C. Saunders for the Verulamium Museum and Cardiff University; see above, p. 300. Information from Miss Birgul Biktimir, Keeper of Conservation at the Verulamium Museum, who submitted the object for inspection.

³⁶ By Mr Alan Harrison who retains possession. The discovery was made during the construction of a chicken farm outside the scheduled area of the known Roman site, together with metalwork, pottery and coins of first- to early second-century date, including Claudian copies in proportions suggestive of a Neronian fort. Information from John Creighton of the Department of Archaeology of the University of Durham, who also supplied a photograph. For the site see Stead, I.M., Excavations at Winterton Roman Villa and Other Sites in North Lincolnshire, (1976), 1–19.Google Scholar

³⁷ For epigraphic occurrences of the genius of private persons, see ILS iii i, index viii (p. 530). For Deccius, probably a single peregrine name rather than a nomen as classified in the indexes of RIB, see RIB 1805 (Carvoran).

³⁸ During excavations directed by P. Bennett under the overall supervision of T. Tatton-Brown for the Canterbury Archaeological Trust; see Britannia x (1979), 334. Information on this and the following six items from Marian Green of the Trust, who submitted the sherds for inspection.

³⁹ The nomen Lal(l)ius, though rare, is attested in Rome: CIL vi 200; Italy, ix 2727, 2228 ff., 3073; xiv 1222; and Gallia Narbonensis, xii 3119.

⁴⁰ During excavations directed by P. Blockley under the overall supervision of T. Tatton-Brown for the Canterbury Archaeological Trust; see Britannia xiv (1983), 334.

⁴¹ There are many possible restorations of which statistically the most likely are [Aur]el(i)i and [A]el(i)i or cognates.

⁴² Presumably the genitive case of a personal name -miv(i)us, but no example is found in Mócsy, Nomenclator, and in Solin and Salomies, Repertorium Nominum, only the rare Mivius (attested twice) and Omivius(?) (once).

⁴³ During excavation directed by K. Blockley under the overall supervision of T. Tatton-Brown; see Britannia xii (1981), 366.

⁴⁴ If this interpretation is correct it is likely that the trader's name, for whom the figurine had presumably been made as part of an order, was written on the side of the figurine above the first surviving line of the text. For negotiators cretarii, see CIL xiii, index 13; BRGK xxvii (1937), 104, No. 188 and ibid, xl (1950), 124, No. 3; and for negotiators cretarii Britanniciani, merchants in the pottery trade specifically with Britain, ILS 4751 (Domburg) and AE 1973, 370 (Colijnsplaat).

⁴⁵ This could be part of a name ending in the suffix -anus, but other expansions are of course possible.

⁴⁶ See n. 40 above.

⁴⁷ If inverted, the last letter could also be an F, H, M, N, P, or R.

⁴⁸ During excavations directed by P. Bennett under the overall supervision of T. Tatton-Brown for the Canterbury Archaeological Trust; see Britannia viii (1977), 423–4.

⁴⁹ In line 2 the tops of the letters only survive and the last letter could also be a D, F, P or R.

⁵⁰ During excavations by Leicester City Museums (now Leicestershire County Museums, Leicestershire Archaeological Unit) directed by Miss J.E. Mellor with the aid of the Ministry of Public Building and Works; see JRS lvi (1966), 203. Information, rubbing and xerox of drawing from P.N. Clay of the Leicestershire Archaeological Unit.

⁵¹ This interpretation assumes the numeral was written retrograde.

⁵² During excavations by Leicester City Museums (now Leicestershire County Museums, Leicestershire Archaeological Unit) directed by Miss J.E. Mellor with the aid of the Ministry of Public Building and Works; see JRS lix (1969), 215. Information on this and the following three items from Dr R.J. Pollard of the Leicestershire Archaeological Unit. Dr Pollard also provided a drawing of the present item and copies of drawings and rubbings of the other three.

⁵³ In line 2 the last letter could also be an N. A is another theoretical possibility but is unlikely following the letter O. The first line almost certainly represents the end of one of the many personal names in -alis.

⁵⁴ During excavations by Leicester City Museums (now Leicestershire County Museums, Leicestershire Archaeological Unit), directed by M.G. Hebditch with the aid of the Ministry of Public Building and Works.

⁵⁵ This numeral is found fairly commonly as a graffito cut after firing on amphorae and is possibly an indication of capacity.

⁵⁶ During excavations by Leicester City Museums (now Leicestershire County Museums, Leicestershire Archaeological Unit), directed by Miss J.E. Mellor with the aid of the Ministry of Public Building and Works; see JRS lix (1969), 215.

⁵⁷ The name seems to be unattested, but is made up of the two elements ^*mapos and rix (‘youthful king’).

⁵⁸ During excavation by Leicester City Museums (now Leicestershire County Museums, Leicestershire Archaeological Unit), directed by M.G. Hebditch with the aid of the Ministry of Public Building and Works.

⁵⁹ By Mr D. Wells in whose possession it remains. Information and drawing, by Noelle Kite, from Peter Liddle of the Jewry Wall Museum, Leicester.

⁶⁰ Until the county boundary changes in 1974 Vernemetum was in the parish of Willoughby-on-the-Wolds, Notts.

⁶¹ For a recent discussion of rings inscribed with the letters TOT, conceivably an abbreviation for the name of the Celtic deity To(u)tates, see Henig, M. and Ogden, J., Ant. Journ. lxviii (1987), 366–7Google Scholar including, p. 382 n. 53, a reference to the present example.

⁶² During excavation by the Department of Urban Archaeology of the Museum of London directed by J. Ayre; information and copies of drawings by Gillian Hale of this and the following item from Dr I. Betts of the D.U.A. See above, p. 307.

⁶³ The stamp has been classified as an impression from die no. 1 in the full study currently being undertaken by the Building Material Department of the D.U.A.

⁶⁴ Taking the long horizontal stroke to be the upright of the letter L on its side. This interpretation differs from that of Sir Ian Richmond and R.P. Wright, commenting on fragments of a tile bearing part of the same stamp found at Treasury Green, Whitehall: see JRS liv (1964), 183 No. 29, where the expansion T(egularia) Car(…) was suggested, taking Car as an abbreviated place-name. The stamp on the Whitehall tile, however, was incomplete at the bottom and, as is now clear, the monogram is certainly composed of more than four letters. For figlinae tegulariae cf. CIL i 594.2.2.24.

⁶⁵ The fact that the two owners share the same praenomen and nomen would suggest that they were freedmen who had been manumitted by the same master, D(ecimus) M(…), as plausibly suggested by Richmond and Wright, op. cit. (note 64).

⁶⁶ During excavation for the Vindolanda Trust directed by Robin Birley, who made this and the next five items available.

⁶⁷ There is a complex of letters after RVSTI on the reverse, resembling V, S and K written together, presumably intended for -CVS.

⁶⁸ The stamp is identical with that on a slipper of similar construction from York (Britannia xviii (1987), 374, No. 32). The second F, tentatively read as E on the York example, could (like it) be E or TE ligatured.

⁶⁹ Presumably a personal name in the genitive case. There are other possibilities, but an attractive restoration is [Tal]ionis, since this uncommon name has already occurred as a graffito on another samian vessel at Vindolanda (Britannia xiii (1982), 419, No. 73).

⁷⁰ Perhaps a personal name, Cimio or Cimius: cf. Britannia xix (1988), 498, No. 34 (with note).

⁷¹ During excavations for the Department of the Environment by the Staffordshire Archaeological Society directed by Mr A.A. Round, who sent details and a drawing. For the site, now identified as that of a mansio, see Britannia ix (1978), 435–6.

⁷² The first letter is uncertain and could be an R or retrograde 5. It is followed by a character like a diminutive L or V on its side. It is tempting to interpret the graffito as the alphabet written from A to Q(?) round the rim of the vessel. However, the estimated diameter of the rim, internally 0.21 m, would mean that there would have been easily enough space to accommodate the full twenty-three letters of the Roman alphabet with space left over, in which case the first letter would be the end of a different text.

⁷³ During excavation for Cadw by the Glamorgan-Gwent Archaeological Trust directed by A.G. Marvell and P.L. Wardle. Information and xerox from D.R. Evans of the Trust.

⁷⁴ During excavation for Cadw by the Glamorgan-Gwent Archaeological Trust directed by V.M. Metcalf; information and copy of drawing from R. Livens and D.R. Evans. For the site see Britannia xii (1981), 317 and ibid, xiii (1982), 334; and for a similar tag from Carlisle, No. 14, above.

⁷⁵ If the letters TR in (a) stand for the contents of the package to which the tag was attached, this could conceivably be triticum, ‘threshed grain’. In (b) we have taken the last line to be the man's origo and if this is correct, then Verona seems a possibility.

⁷⁶ During excavation by the Dean Archaeological Group directed by Bryan Walters, who made the sherd available to RSOT. It has been published by Jones, Terry in Dean Archaeology i (1988), 22.Google Scholar

⁷⁷ The first letter could be I. There is possibly the trace of a letter after E, consonant with N. The Celtic name Cunovendus is already attested in Britain: Britannia x (1979), 348, No. 20; ibid, ii (1971), 292, No. 14 (Cunovindus); and on (unpublished) lead ingots probably of British origin.

⁷⁸ Rivet and Smith, PNRB, reject the identification with ‘Mediomanum’.

⁷⁹ During excavations by the Clwyd-Powys Archaeological Trust directed by Miss J. Britnell and Mr J. Thorburn; see Britannia xvii (1986), 364–5. Information on this and the following four items from Wendy Owen of the Trust.

⁸⁰ The second digit is very much shorter than the first.

⁸¹ Ancarius is a rare but attested Latin nomen: see Solin and Salomies, Repertorium nominum. Following MICARII/ANCARII are possible traces of two further letters, perhaps an O and an I. Below this are what may be traces of a second graffito in much smaller characters.

⁸² At the beginning of the surviving text is the top part of a rounded letter, B, C, D, G, P or R. The last surviving character could be part of an A (lacking the cross bar) or M or N. The space before this and the preceding letter S is not large, but probably represents a division between words, or rather a nomen and cognomen.

⁸³ Presumably part of a personal name, perhaps Greek.

⁸⁴ By RSOT at the suggestion of Dr Graham Webster. It was made available by the Warwickshire Museum, Warwick.

⁸⁵ The object is too fragmentary to be certainly a ‘curse tablet’ complaining of theft, but this is suggested by its appearance and the mention, as in many of the Bath tablets, of an article of clothing. (The ‘Dalmatian’ sleeved tunic is usually written dalmatica, understanding vestis.) Since this is mentioned in both texts, they would seem to be related, although there is no other instance from Britain of a capital-letter text inscribed over cursive. A few texts are written both in capitals and cursive. The purpose of the doubled text is obscure: perhaps for emphasis, or to renew a worn text, or as a summary of the contents. []SEITHAVS remains unexplained. Perhaps a personal name.

⁸⁶ Trier: Kaiserresidenz und Bischofssitz (Landesmuseum Trier, 1984), Nos. 130 a and b, a parallel brought to our notice by George Boon. All three sealings are from different dies.

⁸⁷ The Chi-Rho resembles that on the bronze coinage of Magnentius, and suggests a mid fourth-century date for the sealing. At all events, the Christian symbol means that it is Constantinian or later, when Maxima Caesariensis was one of the four British provinces, with its capital at London (Frere, S.S., Britannia (1987), 198–9).Google Scholar No doubt it included Silchester. The expansion P(rovincia) M(axima) C(aesariensis) is paralleled by that of the third-century lead sealings which read P(rovincia) B(ritannia) S(uperior) and P(rovincia) B(ritannia) I(nferior). The occurrence of the legend both in Britain and at Trier, which in the fourth century was the seat of the praetorian prefect of the Gauls (which included Britain), and indeed during the 340s and most of the period 367–88 also the headquarters of the western emperor, suggests the movement of official consignments. In 359 corn was exported from Britain to the Rhineland, in what is said to have been a revival of earlier practice (Libanius, or. 18.83; Ammianus 18.2.3; Zosimus iii.5). RSOT.

⁸⁸ The excavator, V.E. Nash-Williams, and M.V. Taylor read GAIITICI, but the initial CR is certain. The cognomen Gaeticus is almost unknown, whereas Creticus is common (see TLL Onomasticon, s.v. Cretes, 712.40ff.). The sherd is now on display in the Legionary Museum, Caerleon, where it was noticed by RSOT.