Hostname: page-component-cd9895bd7-hc48f Total loading time: 0 Render date: 2024-12-24T18:31:38.740Z Has data issue: false hasContentIssue false
  • English
  • Français

The Earliest Roman Lead-Mining on Mendip and in North Wales: A Reappraisal

Published online by Cambridge University Press:  09 November 2011

G. Clement Whittick
G. Clement Whittick
Affiliation:
39 Glisson Road, Cambridge
Get access Rights & Permissions [Opens in a new window]

Extract

Reliance on CIL vii 1201 and 1202 has made it customary to quote the year A.D. 49 as a firm date by which Roman lead production had already been fully established on the Mendips. But of these two items one is highly doubtful, while the second on fresh scrutiny proves to involve a misreading of the inscriptions that it carries: neither can justify the confidence that has been placed in it for precise dating, as the following comments seek to show.

Type
Articles
Information
Britannia , Volume 13 , November 1982 , pp. 113 - 123
Copyright
Copyright © G. Clement Whittick 1982. Exclusive Licence to Publish: The Society for the Promotion of Roman Studies

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 So Hübner in CIL, loc. cit., p. 220. cf. among many others, O. Davies, Roman Mines in Europe (1935), 149 n. 5a; J. W. Gough, The Mines of Mendip (1930), 20 ff.; Sheppard Frere, Britannia (revised ed., 1978), 322 – ‘two’ [sc. Leland's tabula and CIL vii 1202, the ingot from Blagdon, = item 2a in the text above] ‘carry inscriptions dating them to A.D. 49’. Note especially Haverfield's comments on Leland's tabula, ‘possibly merely the top surface of a pig’ (VCH Somerset I (1906), 340.4) and ‘plainly a lead pig from the Mendip mines. Possibly enough the letters DE BRITAN may have been on the side and the rest on the face if it was a perfect specimen’ [a view which he later abandoned, since in EE ix (1913), 642, he says that Leland gave the inscription in a single line omnino recte: cf. VCH, loc. cit., no. 4]. H. E. Balch however, (Wookey Hole (1914), 96: cf. Mendip, the Great Cave of Wookey Hole3 (1947), 39: ‘a lead tablet’) follows John Ward, (Phil. Trans. Royal Soc. xlix (1756), 699) in calling it a ‘plate’ of lead, ‘a sort of trophy or monument’. W. Stukeley (The Medallic Hist.… of … Carausius (1757), 176) wrongly says that Camden describes it as a ‘pig’. The dimensions and weight of this object alleged by Hiibner in Rhein. Mus. Phil. NF xii (1857), 352, were due to confusion with the lost ingot from Bruton (CIL vii 1211) and the error was tacitly corrected in CIL vii 1201. For the most recent confusion regarding this object, see Britannia ix (1978), 484 n. 96, describing it as ‘the earliest surviving lead weight from Britain’.

2 CIL vii, ad loc.

3 Graham Webster and Donald R. Dudley, The Roman Conquest of Britain A.D. 43–57 (rev. ed., 1973), 104, 108 n. 120.

4 EE ix, p. 642: F. Haverfield and G. Macdonald, The Roman Occupation of Britain (1924), 256: M. Besnier (‘Le Commerce du Plomb a l'Époque romaine…’ Rev. archéol. (5) xiii (1921), 50 ff.) rejected Haverfield's evidence and preferred to interpret as de ceangi (tanis metallis), an unfortunate suggestion recently revived by Mme F. Laubenheimer—Leenhardt (‘Recherches sur les lingots de cuivre et de plomb a l'Epoque romaine …’, Rev. archéol. de Narbonnaise, Suppl. 3 (1973), 198 n. 8). The four ingots which thus carry DECEANGL on their front face are CIL vii 1204 and EE vii, p. 341, no. 1121 (both dated A.D. 74) (Grosvenor Mus., Chester); also CIL vii 1205 and EE ix 1264, the pair from Hints Common (Staffs) dated A.D. 76 (British Mus. and Tamworth Castle Mus.). Confusion over the precise form of the tribal name stems from the textual problem in Tac. Ann. xii. 32 (ductus inde Cangos exercitus M: in Decangos Bezzenberger: in Ceangos Andresen). As cast on all these ingots the third letter of the name is certainly a cand not a G (cf. R. P. Wright and I. A. Richmond, Catal. of the Roman Inscribed and Sculptured Stones in the Grosvenor Mus., Chester (1955), 47-8, nos. 196, 197). Among the numerous inaccuracies regarding these ingots note in particular R. Cagnat, Cours d'Épigr. lat.4 (1914), 337 (recording no more than DE CEA on one only of the Hints pair: Orth, R.E. Suppl. iv (1924), s.v. ‘Bergbau’ 147, confusedly mentioning ingots with de Clea(ngis) and de Brig(antis) (this latter of the Domitianic pair from Yorkshire; see note 7 below)).

5 CIL vii 1203 (Brit. Mus). See item 2C below, in text.

6 Typical is C. Roach Smith, Arch. Assoc. Journ. v (1850), 227. Hübner (Rhein. Mus. Phil., NF xii (1857), 365) explicitly equated DE BRITAN, DE CEANG, EXKIAN and BRTT/BR(it) as all denoting origin. R. H. Morris, however, (Journ. Chester Arch. Nat. Hist. Soc. ns iv (1890-1) 71 ff.) argued against acceptance of DE as a pre-position on the Flintshire ingots and against any reference to the Kiangi on the Stockbridge specimen. This latter view was supported by Haverfield (VCH Hants I (1900), 323 n. 2). (Morris and others did not discuss the use of DE on Leland's tabula.) Ingots (number unspecified) bearing Domitian's name and the locality mark DECEANG are recorded by Camden as being among the twenty discovered near Runcorn (CIL vii 1206: Stukeley, op. cit. (note 1), 177; Horsley, Britannia Romana (1732), 316; Watkin, Roman Cheshire (1886), 294). We do not know whether on these Domitianic ingots DECEANG formed part of the main inscription (though Stukeley, Horsley and Watkin assumed so, as does F. H. Thompson (Roman Cheshire (1965), 102), or whether (much more probably, by comparison with the extant Flintshire group and with the Yorkshire pair of the same reign: see note 7) it appeared on the front face (as Besnier, op. cit. (note 4), no. 30, suggests and R. F. Tylecote, Metallurgy in Archaeology (1962), tab. 34, nos. 35-54 asserts: cf. Webster, Graham, ‘The lead-mining industry in North Wales in Roman times’, Flints. Hist. Soc. Trans, xiii (1952–1953), 20Google Scholar , no. 23). Camden's report - though mistrusted by Horsley (op. cit., 34, 316) as based on hearsay - is quite acceptable, since his description accords well with the appearance of the normal British type of ingot.

7 CIL vii 1207 (British Mus. and Ripley Castle, N. Yorks): Sheppard Frere, Britannia (rev. ed., 1978), 322 mentions only one ingot of this pair.

8 The Elder Pliny in such contexts seems to avoid the use of de; e.g., [insula Ictis] in qua candidum plumbum proveniat (HN iv. 104): plumbi nigri origo duplex est, aut enim sua vena provenit … aut etc. (ibid., xxxiv 159); cf. his use passim of fieri or nasci with in or apud. CIL xv 1 (lateres) shows that instances of ex (praediis, figlinis) considerably outnumber those of de, and this is even more noticeable in the lists given by H. Bloch (Harvard Studies Class. Phil., lvi (1947), 1 ff.: cf. also his comment on later usage, ibid., viii (1948) 10). Add to these the fact that a number of ingots from Britain carry the mark EX (not DE) ARG (entariis).

9 Obv. Ti CLAVD CAESAR AVG P M TR P VI IMP X: head of Claudius; Rev. Triumphal arch with inscription DE BRITANN (Mattingley, H. and Sydenham, E. A., The Roman Imperial Coinage l (1923), 125Google Scholar , no. 8. Horsley, op. cit (note 6), 328, suggested that the text of Leland's tabula had been copied from this issue).

10 Dio Cassius, lx 21-2; Suet., Claud. 17. For discussion of the significance of the details of the triumph and of other honours, cf. Webster and Dudley, op, cit. (note 3), 78-85; Dudley, D. R., Birmingham Hist. Journ. 7 (1959), 1–17Google Scholar.

11 The two fragments from Charterhouse-on-Mendip (EE iii 121e; iv, p. 206) (Taunton Castle Museum). These prove the double-line setting of a main inscription moulded on the base (as cast) of ingots of Marcus Aurelius and L. Verus (= A.D. 164-9), so supporting the validity of the record of the lost Bruton ingot (CIL vii 1211), which carried the same inscription of 18 + 18 letters. (For another possible example from the Wells area, see JRS xli (1951), 141.7). By the time of Septimius Severus it had been decided to remove such lengthy double-line main inscriptions from the restricted area available on the base (as cast) and to mould them instead on a long side face - a radical change to an inferior technique which had unfortunate results, as the very poor state of the inscriptions on the Chalon-sur-Saone ingots proves (CIL xiii 2612, involving as many as 22 + 25 letters: the CIL record is inadequate).

12 The CIL ad loc. says 1859. But A. Way's report to the Archaeological Institute was made on 5 May, 1854 (Arch. Journ. xi (1854), 278) and gave the date of the find as August, 1853 (cf. Gents. Mag. ns 42 (1854), ii. 58). Details were published in Arch. Anzeiger xi (1853), 485, no. 12 and later also in Rhein. Mus. Phil. NF xii (1857) 352.

13 cf. Roach Smith's comment recorded by Way (‘Enumeration of blocks or pigs of lead …’, Arch. Journ. xvi (1859), 27). Comparison of the various illustrations of this ingot over the years is instructive and proves that the supposed reading FI has been consistently forced upon us through pictorial channels in addition to the many printed references. Way's report to the Archaeological Institute (see previous note) gave the reading BRITANNIC … AVG. FI, or, as an alternative … AVG. IMP. The woodcut ad loc. however clearly shows AVG II. It was reprinted by Yates (Proc. Som. Arch. Nat. Hist. Soc. viii (1858), pl. 1) without modification, even though he and C. Roach Smith (Coll. Antigua, iii (1855), 258) were ‘satisfied that the last letters are FIL, … and not II or IMP’. In 1859 Way again published the same woodcut in his ‘Enumeration’ (op. cit.), p. 24; but his text mentioned only the reading communicated to him by Augustus Franks of the British Museum, viz., BRITANNIC … AVG F … (sic). Much later, still clearly showing AVG II, the woodcut was reproduced, with six others similarly borrowed from Way's ‘Enumeration’, in the British Museum's Guide to the Antiquities of Roman Britain (1922), 30, fig. 21 (where the text made no comment on the reading). The same full-page illustration showing these seven ingots was re-published by Haverfield and Macdonald in The Roman Occupation of Britain (1924), 257, fig. 66. But on close scrutiny it will be seen that the block has there been very slightly retouched, in this detail alone, in order to show no longer AVG II, as in the past, but AVG FI. In the only published photograph known to me (F. A. Knight, Heart of Mendip (1915), 517 - listed as ‘Photographed for the author’) the alleged reading AVG FI has been arbitrarily delineated in chalk. W. Gowland's drawings of twelve ingots broke the tradition, since they were specially prepared for his paper of 1901 (‘The early metallurgy of silver and lead: Part I, Lead’, Archaeologia lvii (1901), pls. lvii, lviii and p. 400); no. 14 therein shows FI in dotted outline (though in Tab. iii, ad loc., it appears as AVG. F.I.), and five years later Haverfield gave the same treatment to a somewhat smudgy sketch (not Way's original) of the ingot in VCH Som. I (1906), 341, fig. 96. Hübner (CIL, loc. cit.) had printed FI, but he expressed doubt; his comment ‘litterae sunt optimae et alte exstantes’ can only be true of BRITANNIC, and is strange in the light of his own discussion. The reading FI, or even in some records FIL, has of late not been questioned, though A. R. Burn has reverted to AVG II without comment (The Romans in Britain (1932), no. 10; (ed. 1969), no. 11).

14 So, e.g., Hübner (CIL, ad loc.), unjustifiably claiming that ‘nominis Britannici elementum ultimum periit ictu securis vel mallei’. Haverfield (VCH Som. I (1906), 341, no. 5) also indicates a missing letter; but he did not himself examine the ingot (loc. cit.). Besnier (op. cit. (note 4), no. 22) mistakenly recorded BRITANNIC I.

15 Hübner (CIL, ad loc.) gives only V·ETP·, but the C is clear, as also is the stop after the P. There is no stop after the T. Haverfield (VCH. Som., loc. cit, note 14) inserts one, but in later references he omits it. His ‘inverted s’ instead of C is no more than the effect of a casting-flaw which close inspection reveals to be in fact detached from the c (cf. EE ix, p. 642). The die, as is commonly found when incuse stamps are involved, was carelessly struck while held in a different plane from that of the angled surface of the ingot, so producing a deeper impression of lettering at one side, or end, than at the other, or even, in the worst instances, failing in part to contact the ingot's surface at all. This explains why the c has been thought ‘not in character with the rest of the stamp’ (Graham Webster, op. cit. (note 6), 32 n. 13).

16 Probably from January to April, since suffecti had succeeded them by the 23rd May (CIL xi 6236 = ILS 5540). For the evidence and for the form of the names, see A. E. Gordon, ‘Quintus Veranius, consul A.D. 49’, Univ. Calif. Publ. Class. Arch. 2 No. 5 (1952), 239 ff.: R.E. Suppl. (1955), s.v. ‘Q. Veranius’, 3.947, vi. cf. P. Gallivan, Classical Quarterly n.s. xxviii (1978), 407 ff.

17 Note that A. E. Gordon is unwilling to accept without question the single letter v on the ingot as a reference to Q. Veranius (op. cit. (note 16), 241, and R.E., loc. cit. (note 16)).

18 From his analyses of the Mendip ingots J. A. Smythe concluded (private record) that the Romans in their early days there had exploited a vein or veins of arsenical lead-ore fairly rich in silver, as witness the analyses of the Blagdon and St Valery ingots, which leave little doubt that desilverization had not immediately been introduced; subsequent ingots surviving from the area, with the exception of one (perhaps both) of the small fragments now in Taunton Castle Museum (see note 11 above) have almost certainly been desilverized, the average of their silver-content being slightly less than 11 dwt (=0.017 per cent) to the ton avoirdupois of lead, a remarkably low figure which is quite consistent with desilverization, even allowing for the varying and generally not very high silver-content of Mendip ores. R. F. Tylecote's comments (op. cit. (note 6), 91) are in general agreement with this, and a similar conclusion was reached by Bucknell, E. T., An Historical and Chemical Investigation of the Metalliferous Ores … in the Mendip Hills (Thesis: Univ. Library, Bristol (1924))Google Scholar : cf. Friend, W. Newton and Thorneycroft, W. E., Journ. Inst. Metals, xli (1929), 116Google Scholar.

19 Sheppard Frere, op. cit. (note 1), 103 and n. 39, 106 and n. 40, dating the revolt to A.D. 60, the date of the Bossington ingot.

20 op. cit. (note 13), tab. iii, 402-3, nos. 22, 27, 50, 404, 406.

21 CIL vii 1205 (British Museum): Gowland did not know of the existence of the Tamworth (Castle Museum) specimen also from Hints Common.

22 CIL vii 1209c (British Museum).

23 We now know that in Hadrian's reign Cornovian lead-ore was not thought by the Romans to be worth desilverizing. The ingot from Snead (‘The Roveries’: CIL vii. 1209e), when I first examined it in 1935, was found to have two fragments of galena embedded in its upper surface (as cast in the mould), in the same way as two of the ingots found at Brough-on-Humber in 1940 also had lumps of galena embedded in their original tops (JRS xxxi (1941), 145 ff. Smythe, J. A., Trans. Newcomen Soc. xx (1939-1940), 143 ff.)Google Scholar. The presence of such galena is proof (J.A. Smythe, loc. cit.) that the ingot itself is a product ‘of ore-smelting, not of the reduction of litharge’, i.e., the lead has not been desilverized, since in this process galena would be destroyed. The galena embedded in 1209e consisted of two rounded lumps approximately 25·4 mm by 38 mm by 12·7 mm thick: these had floated one into each of the two corners of the mould remote from the pouring position, which in this instance, as surface ‘flats’ and striations prove (Whittick, , ‘Casting Technique of Romano-British Lead Ingots’, JRS li (1961), 105 ff.)Google Scholar was centrally across the width of the mould about seven inches from the AVG end, so that the stream of molten metal impinged upon the opposite wall of the mould. Smythe's comments on the presence of galena in an ingot as proof of non-desilverization apply to two ingots found at Brough-on-Humber which carry the mark EX ARG: this further evidence from CIL vii 1209e now shows that ingots not carrying the mark EX ARG may similarly be proved not to have undergone desilverization. Further, since the silver-content figures of the three surviving Hadrianic ingots from Shropshire are so very similar (1209a = 0·0082%: 1209c = 0·0070%: 1209e = 0·0079%), it is reasonable to argue that these figures represent the natural average silver content of the then-available lead veins, and therefore that none of the Hadrianic ingots from Shropshire should be regarded as desilverized. Analysis (by J. A. Smythe: private record) of the galena embedded in CIL vii, 1209e showed a content of 15 dwt of silver to the ton avoirdupois of lead, whereas the ingot itself (as was the case with the Brough ingots), showed a considerably higher figure, viz., 0·0079% or approximately 2 oz -11 dwt, (these figures should be added to the comments made by J. A. Smythe, loc. cit. and by R. F. Tylecote (Metallurgy in Archaeology (1962), 90). The Snead ingot, CIL vii 1209e, is preserved in the Mayer Collection at the Merseyside County (formerly Liverpool) Museum. Ft was bought for Joseph Mayer in 1856 from a Mr Morris of Clun (Salop) for the sum of £5 by Thomas Wright, (Brit. Library, Add. Ms. 33, 346, f. 90). Note that the statement in VCH Essex iii (1963), 187, s.v. Marks Tey, is ambiguous. Only some items of the Marks Tey collection at Liverpool were destroyed, not the Mayer Collection as a whole, and I am grateful to the Museum authorities for confirmation that the ingot has suffered no damage: its weight is approx. 190 lb (183·906 kg), not 185 lb as stated by Haverfield (VCH Shropshire l (1908), 265, no. 4) and by Besnier (op. cit. (note 4), 49, 27 f.).

24 cf. Sheppard Frere, op. cit. (note 1), 339-40 n. 4.

25 Chester Arch. Soc. Journ., xxxi (1935), 74.

26 R. F. Tylecote, op. cit. (note 6), 88, table 35.

27 op. cit. (note 13), tab. iii, 402-3, no. 27: copper 0·034%; antimony 0·006%.

28 op. cit. (note 13), no. 50: copper 0·035%; antimony 0·006%.

29 EE ix, 1264 (Tamworth Castle Museum).

30 Other modifications of Tylecote's table 35 due to wider evidence may be noted (see note 26 above): Somerset (with fifteen specimens), 2:9 ratio (as against 1:7 in table 35, with two specimens): Shropshire (with three specimens), 2:48 ratio (as against 9:4 in Table 35, with one specimen). (Figures taken from J. A. Smythe, private records, and W. Gowland (note 13 above), table iii).

31 Arch. Camb. ci (1950), 83-4 and JRS xli (1951), 142 (each with photograph). The latter account mistakenly attributes the ingot to the 2nd century. (Nat. Museum of Wales, Cardiff).

32 See note 4 above.

33 The lettering on the Carmel ingot has a height of 38 mm (1·5 in.) on average (the tallest found on any British ingot); the depth of its inscribed panel is c. 13 mm (0·5 in.) (matched only by the Hadrianic ingot from Theobalds Park, Cheshunt: EE ix, 1264a); the Carmel ingot's magnificent leaf-stop, standing stalk-downwards after NIPI and virtually the same height as the lettering, is a striking instance of a feature of which large examples are rare in pre-Antonine times (Britannia ii (1971), 291 n. 4). No full analysis of this ingot has been made, but J. A. Smythe (private record) found that its silver content was 0·0037% = 1 oz - 4·2 dwt per ton avoirdupois of lead, with ‘a strong trace’ of copper and a specific gravity of 11·38. Since the average of the silver-contents of the five other ingots from the Flintshire mines is only 14·36 dwt, it might be thought that the Carmel ingot has not been desilvered. But the figures must be regarded as inconclusive.

34 Dessau ILS 8710 combines the two Chester ingots; but their inscriptions are different: cf. R. P. Wright and I. A. Richmond, op. cit. (note 4). It may be noted that since these ingots prove that imperial production of ingots was well established by A.D. 74, it antedates Frontinus’ campaign in the area.

35 The unimpressive appearance of the inscriptions on the Domitianic ingots from Yorkshire (note 7 above) is a good example of similar poverty of craftsmanship. On the other hand, absence of a frame does not necessarily exclude good workmanship in epigraphy, as some of the ingots of the Societas Lutudarensis prove.

36 See note 6 above.

37 One would however prefer to say in inferiore parte in reference to the true base of an ingot as cast.

* Lead ingots are weighty and unwieldy objects to handle, whether in a Museum or in a private home, and I wish to express my sincere thanks to the many Curators and private owners whose patient cooperation has enabled me to examine in detail the ingots in their possession, and often to obtain specimens of the metal for analyses carried out by the late Dr J. A. Smythe. Special thanks are due to the British Museum authorities, who on more than one occasion arranged for me to have access to their twelve ingots and allowed me to make squeezes from which the photographs in the above text have been taken.

38 1772–for which W. Pitts’ ‘1792’ (in Top. Hist. Staffs. (Newcastle-under-Lyme, 1817), 164) is perhaps a misprint - is given in most accounts, including the Catalogue (ed. 3, 1786) published by Richard Greene, antiquary and surgeon of Lichfield, in whose private museum in that city the ingot was for some years exhibited (Gents. Mag. lviii (1788), 847; Stebbing Shaw, Hist. & Antiqs. Staffs. I (1798), 308, pl. xxx, 331; A. Way, op. cit. (note 13), 28) until dispersal of the collection after his death in 1793. But Greene's own account of its dis-covery (letter of 17 October, 1772; Gents. Mag. xlii (December 1772), 558: cf. xliii (1773), 61) says ‘last winter’, and Haverfield (VCH Staffs. I (1908), 190) rightly dates the find to 1771, as did Pennant and Erdeswick.

39 Norris, H., Tamworth Castle, (Tamworth, 1899), 5Google Scholar , dates the find to ‘about 1833’; ‘in 1830’ appears in editions of the same writer's Guide to T. Castle (Tamworth, 1924, 1929, 1930), pp. 16 ff). A Catalogue of the Museum of 1906 (ed. W. Moreton, then Hon. Curator) says ‘about 1830’. The Curator kindly tells me that a MS note in the Museum's records says ‘found about the year 1830 on Packington or Hints Moor, about two miles west of the church’. In 1885 it was presented to Tamworth Parish Church (where it was put on exhibition in the vestry) by Robert Isham Woody, of Moat House, Tamworth. How he came to possess it seems uncertain Tamworth Castle was purchased by the town in 1897 and the ingot was probably moved there as an exhibit (1898–9). In the 1900 ed. of his Guide to T. Castle, 15 ff., W. Moreton had dated the find to 1838. So also Haverfield (EE, loc. cit.) followed by Besnier (op. cit. (note 4), 32b) and Graham Webster (op. cit. (note 6), no. 22) (wrongly giving Tamworth as the find-spot); but the consensus of evidence is against this late date.

40 Therefore not recorded with its fellow, the British Museum ingot from Hints in VCH Staffs. 1 (1908 and 1968 reprint), 190, s.v. This Tamworth ingot was at first overlooked by Atkinson and Taylor (Flints. Hist. Soc. Trans, ix (1922), 8-9; but cf. x (1924), ii, no. iv, where, however, the side inscription is said to be wanting) and is omitted by Tylecote (op. cit. (note 6), tab. 33, 34), by O. Davies (op. cit (note 1), 158 n. 5) and by Ellis Davies (Prehistoric & Roman Remains of Flintshire (1949), 146).

41 H. Norris (loc. cit., note 39); Haverfield, EE ix 1264. But Norris’ wording is: ‘the inscription [sic: sc. on the British Museum ingot] is IMP VESP etc … on the top, and on one side it has DE CEA. … The second dis-covery was made about 1833 when another pig [sc. the Tamworth Museum specimen] was unearthed bearing a like inscription fsic] to the former! The inscriptions [sic] clearly indicate that the lead was moulded in the days of Vespasian and Titus in the year A.D. 76; while DE CEA presumably denotes the locality …’. There is ambiguity here: but it seems to me that the distinction specifically made between the ‘inscription’ on the ‘top’ (i.e., the base as cast), with their imperial names and dating, and the locality mark on the front ‘side’, makes it unlikely that Norris was in fact recording, as Haverfield claimed, any lettering on the front face of the Tam worth ingot; I incline to the view that, like many others, Norris failed to distinguish the inscription on the front face.

42 Besnier, op. cit. (note 4), 32b.

43 A. R. Burn, the Romans in Britain2 (1969), 25 no. 25., says in error that the side lettering is ‘scratched’ on the British Museum ingot and that it is absent from the Tamworth specimen. The rubbings shown in PL. IVB below were taken several years before World War II and since then the side lettering on both ingots has deteriorated greatly; on the Tamworth ingot only the upright of the letter D is now visible - and that barely – to the naked eye. The most recent references to the Tamworth specimen (Graham Webster, op. cit. (note 6), no. 22, and A. J. H. Gunstone, N. Staffs. Journ. Field Studies (Univ. of Keele) 4 (1964), 25) do not mention the side lettering. Gunstone gives ‘1833 or 1838’ as the find-date. But see note 39 above.

44 Hence disregard the comments in Flints. Hist. Soc. Trans, (loc. cit., note 40), 9, no. 3. The N has usually been omitted from descriptions of the British Museum specimen also, as it was from Gowland's sketch (op. cit. (note 13), 50), reprinted by Haverfield in VCH Staffs I (1908 and 1968 reprint) and more recently by Tylecote (op. cit. (note 6), no. 50). Besnier (loc. cit. (note 4), 32b) follows Hiibner's mistaken version in recording DE CEA, and Haverfield (VCH. loc. cit.) misprints as DECEAN. G.

45 Gunstone (op. cit. (note 43)) and Graham Webster (op. cit. (Note 6), no. 21) print the side inscription of the British Museum specimen as though both N and L were absent. Since all extant examples of this tribal name on ingots are now known to include the final L, its insertion in the inscription recorded from some of the Runcorn group (Graham Webster op. cit. (note 6), no. 23) may be thought justified.