¹ I gratefully acknowledge the invaluable help I received from Professor A. L. F. Rivet and Professor Colin Smith, both when I communicated my reading to them and later a draft of this note, and from their splendid book The Place-Names of Roman Britain (London, 1979), which is cited hereafter as P.N.R.B.Google Scholar

² J.R.S. LVI (1966), 223, no. 33Google Scholar with pl. x, 7. An uninscribed portion of the tile has since broken off and been lost. Another fragment, originally published as Britannia, IV (1973), 333, no. 35Google Scholar (b) may belong to line 4. It reads: […]ML.[…]. Mr. Wright's reading has been republished and discussed in Todd, M., The Coritani (London, 1973), p. 40Google Scholar; Mc Whirr, A. (ed.), Roman Brick and Tile, B.A.R. Int. Ser. 68 (Oxford, 1979), pp. 238Google Scholar and 246; P.N.R.B., p. 220, s.v. Coria Soliliorum; Whitwell, J. B., The Coritani: some Aspects of the Iron Age Tribe and the Roman Civitas, B.A.R. 99 (Oxford, 1982), p. 55Google Scholar. See now Britannia, XIV (1983), 349–50.Google Scholar

³ Britannia, X (1979), 347–8, no. 20 with fig. 24; McWhirr, op. cit. (note 2), pp. 241–3 with fig. 11.4. Dr. Webster informs me that the Tripontium tile was found with others, unstratified, in a scatter of building debris containing fourth-century artefacts from an earlier building of uncertain date. Thus the archaeological context is no help in dating the graffito.Google Scholar

⁴ McWhirr, op. cit. (note 2), pp. 231–51.

⁵ There is no instance of -cesa (etc.) in the reverse index of Gradenwitz, O., Laterculi Vocum Latinarum (1904). A remote possibility is diocesa, a form of dio(e)cesis occasionally found in Medieval Latin and, as diocisa, as early as the first half of the fifth century (Eucherius of Lyons, cited in Corpus Glos-sariorum Latinorum, vi, s.v. dioecesis). The possible fragment of this line (see note 2) does not preclude this restoration, but makes a place-name more likely.Google Scholar

⁶ Todd, op. cit. (note 2), ch. 1, esp. p. 18.

⁷ P.N.R.B., p. 324, s.v. Coritani. Eight other civitates lay within the Fosse Way, all named (and two mis-spelt) by Ptolemy; five (allowing for variations in spelling) are attested by the Antonine Itinerary or the Ravenna Cosmo graphy, and four by both; the other three are attested by inscriptions and/or literary sources.

⁸ Geography, 11, 3, 11 =P.N.R.B., p. 143. The two civitates of the other eight (see note 7) which are mis-spelt are the Catuvellauni (Katueukhlanoi, etc.) and the Iceni (Simenoi). Thus Ptolemy's spelling is not always reliable.

⁹ Ed. Schnetz (1940), p. 106, line 47 = P.N.R.B. p. 207.

¹⁰ P.N.R.B., ch. 5, esp. 189–200.

¹¹ P. 106, lines 46–50 = P.N.R.B., p. 207, whose identifications are adopted here (but see note 13), including Lactodurum (Towcester) for Iacio dulma.

¹² The pages of the three MSS. relating to Britain are reproduced in Archaeologia, XCIII (1949), pls. II–X.Google Scholar Vatican and Paris are written in double column (but with different line-divisions) and place a stop after each place-name; Basle is written across the page in a single column and begins each place-name with a capital letter. The errors in word-division noted below occur in all three MSS. and thus go back to the archetype and, presumably, to the Cosmographer himself.

¹³ P.N.R.B., p. 468, s.v. Tamus. Professor Smith, who noted the coincidence between the text of the Cosmography and my reading of the Tripontium graffito, informs me that he would now withdraw the identification proposed in P.N.R.B. of Tamus with Eltavori/Eltanori: ‘while at the time it seemed logical to think this El- name another mistaken Fl- (river) name, in line with several others in this text, it is now clear that this must be abandoned and that Eltavori/Eltanori must be considered as part of the name of Leicester’. Eltanori was read by Richmond and Crawford (Archaeologia, XCIII (1949), 18)Google Scholar, but in their Com mentary (ibid., 33) they printed Eltavori. Schnetz read Eltauori, noting only that previous editors had read Eltanori. It can be seen in Archaeologia, XCIII (1949), pls. IIGoogle Scholar (third column, line 19), IV (second column, line 14), and VIII (lines 15–16), that the Vatican and Basle MSS. read Eltauori, and that in the Paris MS. u and n are often almost indistinguishable. However, other place-names where u is correct are also written with an n-like u; and the distinction seems to be that the scribe of the Paris MS. finished the first stroke of u, but not of n, with a serif to link it with the second stroke (somewhat in the manner of the Vs in the Tripontium graffito), and began the second stroke of both u and n with a serif that often extended indifferently as far as the first stroke, whether of u or n. Thus it would seem that Schnetz was right to deny the ‘variant’ Eltanori any MS. authority.

¹⁴ P. 106, line 20 = P.N.R.B., p. 206; dupli cated by Noviomagno (no tribal name) at p. 106, line 17. See P.N.R.B., pp. 190 f., and esp. 192 and 199–200.

¹⁵ P. 106, line 22=P.N.R.B., p. 206; p. 106, line 54 = P.N.R.B., p. 208 (which reads Venta Cenomum, but all three MSS. write it as one word).

¹⁶ P. 106, line 27=P.N.R.B., p. 207; p. 107, line 7 = P.N.R.B., 208; p. 106, line 36 =P.N.R.B. p. 207. Cf. Devionisso and Statio (p. 105, line 51), Cironium and Dobunorum (p. 106, line 31), Utriconion and Cornoviorum (p. 106, line 40), all of them written as two place-names by all three MSS. (see note 12). For the Cosmographer's errors in word-division see P.N.R.B., pp. 191 and 204.

¹⁷ For the ‘Romance speech-habits’ of the Cosmographer see P.N.R.B., pp. 200 f., which notes that -on forms occur only in his place-names, not in his narrative prose. Although this can be explained, one cannot be sure that he introduced all the ‘vulgar’ spellings himself: e for ae is very common in the epigraphy of Roman Britain (see Britannia, II (1971), 220), and the Tripontium graffito now provides examples of -om for -um.Google Scholar

¹⁸ ‘Coritavi’ was preferred by Grundy, G. B. in his edition of Murray's Small Classical Atlas (London, 1904).Google Scholar

¹⁹ I leave it to philologists to determine whether this throws any light on the difficult etymology of ‘Leicester’, and whether civitas Corieltauvorum could have become Legorensis civitas by 803 (see Jackson, K., Language and History in Early Britain (Edinburgh, 1953), 45, who posits *Ligora).Google Scholar