^{page 222 note 1} For the numbering of the insulæ, see fig. 1.

^{page 222 note 2} Cirencester Urban District Council, Mr. H. Pitts, Messrs. J. Jeffries & Sons, the Memorial Hospital Management Committee, Mrs. J. Bailey, Mr. W. L. Yearick, and the Cirencester Housing Society.

^{page 222 note 3} Grants were received both in 1967 and 1968 from the Society of Antiquaries of London, Bristol and Gloucestershire Archaeological Society, and the Ministry of Public Building and Works.

^{page note 4} Site supervisors were, Miss P. Cullen, Messrs. S. G. Austin, M. W. C. Hassall, D. Viner, M. W. Westcott, B. E. Vyner, and C. Gingell. The pottery shed was looked after by Miss E. Waite and Mrs. H. McWhirr, whilst the photography was done by Messrs. R. A. Fagence and C. E. Birchell. Mr. W. A. Blythe once again acted as site treasurer, and Miss J. Barker shouldered most of the background organization. Mr. J. Real continued to make us welcome at the museum and to help us in a number of ways.

Mr. R. Reece kindly examined and reported on the coins found during the excavations, and Dr. F. S. Wallis carried out an analysis of the tesserae from the mosaics.

We were fortunate in having the services of Mr. D. S. Neal in drawing the mosaics and Dr. D. J. Smith who has written the second part of this paper.

^{page 222 note 5} Antiq. Journ. xlvii, 191–2, fig. 4.

^{page 224 note 1} A similar rear ditch was found behind the fort rampart at Leaholme; Antiq. Journ. xlii, pl. vii, section BC.

^{page 227 note 1} Antiq. Journ. xlv, 105.

^{page 227 note 2} For the two phases of walling, and the greater depth of foundation of the narrow wall, Antiq. Journ. xlvii, 190.

^{page 227 note 3} A ‘Roman Well’ recorded on the O.S. map, some 25 feet to the east, is probably medieval also.

^{page 229 note 1} A similar block was found in the north-east section, Antiq. Journ. xli, 67, pl. XVIII.

^{page 229 note 2} J.R.S. lxiii, 206, no. 3.

^{page 229 note 3} Sections here have recorded both narrow and wide walls; Antiq. Journ. xxxvii, 206; xliii, 22; xlvii, 191.

^{page 229 note 4} Ibid., 188.

^{page 229 note 5} Ibid., 190.

^{page 229 note 6} This raises the question of whether other bastions exist between those plotted on the plan, fig. 1.

^{page 229 note 7} Ibid, xlii, figs. 3–5.

^{page 229 note 8} Ibid., pl. vii.

^{page 229 note 9} Ibid., 9, fig. 5, pl. vii.

^{page 233 note 1} By Messrs. J. Real and P. Grace who kindly provided notes and a plan.

^{page 233 note 2} The southern limit of these rooms was marked by a wall parallel to the corridor walls and uncovered during the restoration of the site early in 1969.

^{page 234 note 1} These views may have to be modified in the light of discoveries made during the restoration of the site.

^{page 234 note 2} For the detailed discussion on the mosaics see Part II (p. 235).

^{page 234 note 3} Antiq. Journ. xlii, 3.

^{page 235 note 1} Ibid. xlvi, 244 and pl. XXXVIII.

^{page 235 note 2} Ibid. xlvii, 194, fig. 5, pls. XXXIIIa, XXXIVb.

^{page 235 note 3} It is proper to say that I did not myself see the remains of these mosaics. I am indebted to Mr. A. D. McWhirr not only for inviting me to comment on them but also for providing notes and colour transparencies together with photographs of the meticulous and convincing scale drawings by Mr. David S. Neal.

^{page 236 note 1} See Smith, D. J., ‘Three Fourth-century Schools of Mosaic in Roman Britain’, in La Mosaïque Gréco-Romaine (Centre National de la Recherche Scientifique, Paris, 1965), pp. 95–116,Google Scholar and id., in The Roman Villa in Britain (ed. Rivet, A. L. F., 1969), pp. 97–102, 112–13, 115–16Google Scholar; hereafter respectively abbreviated to Smith (1965) and Smith (1969).

^{page 237 note 1} Gonzenbach, V. von, Die römische Mosaiken der Schweiz (1961), no. 56-IV, pls. 22–3 (misnumbered III on pl. 22)Google Scholar.

^{page 237 note 2} See Smith, D. J., ‘The Labyrinth Mosaic at Caerleon’, Bull. of the Board of Celtic Studies, xviii (1959), 304–10Google Scholar(inc. a list of all the labyrinth mosaics then known to the writer).

^{page 237 note 3} Lysons, S., An Account of Roman Antiquities Discovered at Woodchester (1797), pl. XII (lower)Google Scholar; id., Reliquiae Britannico-Romanae, ii (1817), part 1, pl. XXIIGoogle Scholar.

^{page 238 note 1} Lysons, S., An Account of … Woodchester (1797), pl. XIGoogle Scholar.

^{238 2} id., Reliquiae Britannico-Romanae, ii (1817), part 1, pl. XIXGoogle Scholar.

^{page 238 note 3} No published illustration.

^{page 238 note 4} Hakewill, H., An Account of the Roman Villa … at Northleigh (1826), pl. 2, fig. 1Google Scholar.

^{page 238 note 5} MacLean, J., ‘Tockington Park’, trans. Bristol, and Glouc, . Arch. Soc. xii (1888), pl. VIGoogle Scholar.

^{page 238 note 6} Ibid., pl. VII.

^{page 239 note 1} The mid and outer points of the peltae are drawn out too far in Hakewill's engraving (see p. 238, n. 4), though otherwise the engraving appears to preserve, within its limits, a very true record of this mosaic as it was when first uncovered. Hakewill's other engravings show that he did not restore missing parts of the mosaics of North Leigh even when restoration would have been both possible and justifiable.

^{page 239 note 2} Hakewill, , op. cit., fig. 2.Google Scholar In the Haverfield Library, Ashmolean Museum, are tracings of fragments of mosaic found at North Leigh of which one showing this motif is annotated as ‘Filling of W angle’ of Room 8.

^{page 239 note 3} Smith (1965): id. (1969).

^{page 239 note 4} Wacher, John S., ‘Cirencester, 1964: Fifth Interim Report’, Antiq. Journ. xlv (1965), 102–104, pl. XXXIIIaGoogle Scholar.

^{page 240 note 1} Hakewill, , op. cit., pl. 2, fig. 1Google Scholar.

^{page 240 note 2} Unpublished unfinished pencil and watercolour drawing (by Lysons?) in the Gloucestershire volume of the ‘Topographical Collections’ of the Society of Antiquaries of London.

^{page 241 note 1} Cf. Smith (1965), p. 110.

^{page 241 note 2} Archaeohgia, lx (1907), pl. XLI. This accurate coloured lithograph shows just over half of the design, but the mosaic itself is preserved in entirety at Stratfield Saye. I am indebted to H.G. the Duke of Wellington for allowing Mr. David Baker to photograph it for me.

^{page 241 note 3} On the discovery at Trier of an unquestionably ‘Corinian’ pattern and the questions that this raises concerning the period and eventual fate of the Corinian school see Parlasca, K., Die römische Mosaiken in Deutschland (1959), p. 50, pl. 50Google Scholar, Smith (1965), pp. 111–14, and Smith (1969), pp. 115–16.

^{page 241 note 4} Cf. Smith (1965), p. 110.

^{page 241 note 5} O'Neil, Helen E., ‘Whittington Court Roman Villa, Whittington, Gloucestershire’, Trans. Bristol and Glouc. Arch. Soc. lxxi (1952), 35–6, 47, pl. 11Google Scholar.

^{page 242 note 1} O'Neil, Helen E., ‘Whittington Court Roman Villa, Whittington, Gloucestershire’, Trans. Bristol and Glouc. Arch. Soc. lxxi (1952), p. 45Google Scholar, and information kindly communicated to me by Mrs. O'Neil.

^{page 242 note 2} R. E. M., and Wheeler, T. V., Report on the Excavation of the Prehistoric, Roman and Post-Roman Site in Lydney Park, Gloucestershire (Soc. Ants. London, Research Report IX, 1932), pp. 65–7, Frontispiece and pls. XIX-XXIIGoogle Scholar.

^{page 242 note 3} Smith (1965), p. 106; id. (1969), pp. 97–8.

^{page 242 note 4} Except for the instances cited, the saltire-design with the characteristic pelta-motifs is unknown in Britain or abroad. Mention must be made, however, of three inter-related Romano-British mosaics which were ultimately derived from a saltire-design and incorporated other filling-motifs such as have been noted in association with the pavements here attributed to the Corinian school. One is the mosaic of (Old) Broad St., London (R.C.H.M. Eng., London, iii—Roman London, pls. 39, 48), wherein sixteen-petalled flowers with guilloche-knot centrepiece framed by interlaced squares of guilloche immediately recall one of the pavements of Cirencester (? Lysons's drawing) and that of Room XII at Tockington Park. The same can be said of a pavement from Kenchester, Herefords. (Report of tie Research Committee of the Woolhope Club, 1916, pl. 23), which as well as this motif incorporated square panels of guilloche mat, thus recalling also Mosaic 2 of Cirencester 1968 and Room 1 at North Leigh, while each quarter of the entire mosaic was basically a repeat of, again, the design of Room XII at Tockington Park. The third is the pavement of Bishopstone, also in Herefords. (R.C.H.M. Eng., Herefordshire, iii—N.W., pl. 88 = F.C.H. Herefordshire, i, fig. 14, facing p. 192), in which, together with interlaced squares of guilloche and the sixteen-petalled flower with guilloche knot centrepiece, the pelta-motif without triangular ‘foot’ but otherwise of the type characteristic of the Corinian saltire-designs appears in the angles, again as in Room XII at Tockington Park, This pavement must derive from that of Kenchester and that in turn from the design of Room XII at Tockington Park, while even the mosaic of London appears to owe its repertory almost wholly to the repertory of the Corinian school. In fact, a mosaic of Dyer Street, Cirencester (1777) could be the earliest of this series; see Lysons, , Rel. Brit.-Rom. ii (1817), part 1, pl. vGoogle Scholar.