^{page 16 note 1} Arch. J. XI (1854), 49.Google Scholar

^{page 16 note 2} Archaeologia, XLV (1880), 462.Google Scholar

^{page 16 note 3} Coins from the hoard are in Rougemont House Museum, Exeter (5) and Somerset County Museum, Taunton (11).

^{page 19 note 1} Farrar, R. A. H., ‘The Techniques and Sources of Romano-British black-burnished ware’ in A., Detsicas (ed.), Current Research in Romano-British Coarse Pottery (C.B.A. Research Report 10, 1973).Google Scholar

^{page 19 note 2} Wheeler, R. E. M., Maiden Castle (1943), p. 232, fig. 72.Google Scholar

^{page 19 note 3} Proc. Devon Arch. Soc. i, pt. 2 (1930)Google Scholar, pl. XXV and Fox, A., South-West England (1964), pl. 75.Google Scholar

^{page 19 note 4} P.P.S. XXXIV (1968), 124, figs. 15–23.Google Scholar

^{page 19 note 5} Fox, A., Roman Exeter (1952), figs. 14–16.Google Scholar

^{page 21 note 1} Listed in Appendix 1, pp. 37 ff., and fig. 13

^{page 21 note 2} I am very grateful to Sheila Pollard for inviting me to collaborate in the publication of the mirror and to John Brailsford, Keeper of the Department of Prehistoric and Romano-British Antiquities for facilities to study it and other objects in the collection, and for his personal interest and encouragement. I am alsoindebted to Kenneth Painter and to Catherine Johns for many helpful discussions in the Department. The Dorset Natural History and Archaeological Society kindly gave permission for the Verne mirror fragment to be drawn at the British Museum. The quality of the illustrations is due to Philip Compton, draughtsman in the Museum Department, and to Robert Turner, former technician in the Archaeology section, History Department, Exeter University.

^{page 21 note 3} Fox, Cyril, Pattern and Purpose (1958: hereafter cited as P. and P.), pl. 30, a.Google Scholar

^{page 21 note 4} Arch. Camb. C (1948), 35.Google Scholar

^{page 22 note 1} P. and P., p. 24. Type III A is a Y-shaped loop: probably the two forms are distinct: see Beds.Arch. J. v (1971), 11, where M. Spratling argues that the single Y-loop is a British invention, whilst the multi-looped handles were derived from Italianpaterae.Google Scholar

^{page 22 note 2} P. and P., pl. 56b, q.

^{page 22 note 3} Archaeologia, XLVIII (1884), 116Google Scholar, pl. 6; see now Antiq. J. LII (1972), 298 and fig. 4, 9.Google Scholar

^{page 22 note 4} Spratling, M., Beds. Arch. J. V (1971), 10.Google Scholar

^{page 22 note 5} Megaw, J. V. S., Art of the European Iron Age (1970), p. 263. The studs are now thought to have been red originally.Google Scholar

^{page 22 note 6} Antiq. J. XLI (1961), 186, pl. xxxiiib.Google Scholar

^{page 22 note 7} Antiq. J. XLIII (1963), 27, fig. 2a and b.Google Scholar

^{page 22 note 8} Cornish Arch, VI (1970), 5Google Scholar, and B.M. Quarterly XXXV (1971), 151 where stippling is discussed.Google Scholar

^{page 22 note 9} Arch. J. CXIII (1956), 80.Google Scholar

^{page 22 note 10} P. and P., fig. 59, b.

^{page 23 note 1} Other people see the image as a long-eared owl, but there is no beak, essential to a bird of prey.

^{page 23 note 2} The subject has recently been discussed by Megaw, Vincent in ‘Cheshire Cat and Mickey Mouse’, P.P.S. XXXVI (1970), 261.Google Scholar

^{page 23 note 3} Megaw, J. V. S., Art of the European Iron Age (1970), no. 252.Google Scholar

^{page 23 note 4} Massow, W. von, Die Grabmaler von Neu-magen, I (1932), pls. 34–8.Google Scholar

^{page 23 note 5} Arch. J. CXX (1963), 43 and 55. The mirror handles are Iron Age types.Google Scholar

^{page 23 note 6} Propertius II, xviii, 23–32. A friend has supplied the following translation.

So now, you crazy girl, you're taking your style from painted Britons? Having your fun, dolled up with a foreign tint on your gleaming head? Nature's look is always right: a Belgic colour is disgusting on a Roman face. If a silly young thing dyes her hair to change it, I hope she'll suffer horribly in the Underworld! Throw the stuff away: you'll look lovely to me anyway— lovely enough, if you come often enough. Really, if Miss X likes to dye her temples blue, does that mean that the ‘Blue Look’ is nice?

^{page 24 note 1} The technique of mirror engraving is currently being studied by Richard Savage and Philip Lowery and also forms part of Mansel Spratling's University of London Ph.D. thesis. All three have indicated that the designs were based on compass construction. My brief account is based on Mr. Savage's contribution to the Society's meeting at which the paper was given, and on friendly discussions and correspondence whilst this article was being prepared for publication.

^{page 24 note 2} I put the problem to Mrs. Lane, who wrote as follows:

Corrosion can form in two ways: it can either follow the contours of the metal surface exactly, thus preserving any design in the remaining metal, or it can attack in such an uneven way that no design remains. If there is no design on the present metal surface that has been attacked by corrosion, there is no way of telling visually whether there was originally a design or not. Even if it were possible to examine the surface metallographically where there was no design visible, one could not say with any certainty whether there had never been any design or whether the corrosion had removed the layer of metal beneath it which would retain the ‘shadow’ of the design. Sometimes the corrosion layers retain the design even when it does not remain in the metal surface. In the Holcombe mirror the corrosion layers only retained the design where there was design remaining in the metal. So I do not think that where there was heavy corrosion one can form any conclusions; where there was little corrosion one might be more certain.

^{page 27 note 1} Antiq. J. XLVIII (1968), 47: it is derived from a crescent shield used by Thracians and other eastern tribes.Google Scholar

^{page 27 note 2} Antiq. J. XXVIII (1948), 131, fig. 7.Google Scholar

^{page 27 note 3} Ibid. 135–6 and fig. 9.

^{page 28 note 1} Antiq. J. XLI (1961), 194.Google Scholar Mansel Spratling has recently questioned this in discussing the dates of British mirrors, Beds. Arch. J. V (1971), 14.Google Scholar

^{page 28 note 2} R.C.H.M. Dorset ii, p. 605Google Scholar.

^{page 28 note 3} P. and P., pl. 58.

^{page 28 note 4} Antiquity XXXIV (1960), 207.Google Scholar

^{page 28 note 5} P. and P., pl. 60.

^{page 30 note 1} Archaeologia LXI (1909), fig. 3, p. 333Google Scholar, and Spratling, M., Beds. Arch. J. V (1971)Google Scholar, pl. 2. Two of the same motifs are also on the St. Keverne mirror roundels. (Fox, A., South West England (1964), pl. 76).Google Scholar

^{page 30 note 2} This was done at the British Museum in 1950, when the mirror was sent for conservation from Holland. A new detail, not shown on Dunning's drawing in Arch. J. LXXXV (1928), 70 is the crescent at the top of the pattern.Google Scholar

^{page 30 note 3} Britannia, I (1970), p. 136Google Scholar, n. 42. Dr. J. Bogaers had previously suggested that it was brought back by a legionary of the Ninth Legion in the early second century. This is less likely because the legion had been stationed in the north-east at Lincoln and York (Studien zu den Militärgrenzen Roms (1967), p. 75 and Taf. B).Google Scholar

^{page 30 note 4} P. and P., fig. 60 and p. 97. The mirror was destroyed in the last war but fortunately had been sent for conservation to the British Museum in 1939, when it was photographed. The negatives have disappeared but a print recently came to light in the Dept. of Prehistoric and Romano-British Antiquities which is reproduced as pl. VI. It was not available in 1956 when Cyril Fox re-drew the design from the original publication by Spence Bate in Archaeologia, XL (1866), pl. xxx. The photograph shows that the top portion of the uppermost roundel was filled with basketry and presumably was without the crescent shown in the previous drawings.Google Scholar

^{page 31 note 1} Archaeologia, LXI (1909), 340, fig. 8.Google Scholar

^{page 31 note 2} P. and P., fig. 51, p. 86.

^{page 31 note 3} P. and P., pl. 60.

^{page 31 note 4} P. and P., pl. 56a, c: the mirror is now lost, see Appendix II, p. 39.

^{page 31 note 5} ‘The craftsman is surely building up a new diffusive style in which the voids are intended to please the eye as much as the pattern.’ P. and P., p. 94; see also Antiquity, XXXIV (1960), 210 for comments on the Great Chesterford mirror, and its new diagonal balance.Google Scholar

^{page 31 note 6} Mayer and St. Keverne constitute a third group. Spratling has recently argued that all mirrors could be contemporary.

^{page 34 note 1} Archaeologia, XIV (1803), 90, pl. XXII, 2. The hoard deserves to be fully drawn and republished.Google Scholar

^{page 34 note 2} P.P.S. XVIII (1952), 47.Google Scholar

^{page 34 note 3} Ibid., p. 52 and fig. 3.

^{page 34 note 4} A harness set consisted of a pair of bits, one large terret, and four small ones: cf. the Stan wick hoard, Yorkshire (P.P.S. XXVIII (1962), 17).Google Scholar

^{page 34 note 5} The technique, though characteristic, is not confined to the south-west: patterns of punched dots occur on a terret from Stanwick (loc. cit, fig. 7, 24) and from Snettisham (Norwich Mus.). See also Megaw, V., B. M. Quarterly, XXXV (1971), 151 for references to other examples.Google Scholar

^{page 35 note 1} Camulodunum Types III and IV.

^{page 35 note 2} Antiq. J. X (1930), 154Google Scholar: Piggott, S., P.P.S. XVI (1950), 8, fig. 3, 3.Google Scholar

^{page 35 note 3} Piggott, op. cit., p. 10.

^{page 35 note 4} J. V. S. Megaw, Art of the European Iron Age, no. 257; P. and P., pl. 39, p. 49, ‘Thames School’.

^{page 36 note 1} P.S.A. Scot. IV (1861), 293.Google Scholar

^{page 36 note 2} J. Anderson, Scotland in Pagan Times, p. 127, fig. 5, 102–6.

^{page 36 note 3} Bulleid, A. and Gray, H. St. G., The Glastonbury Lake Village I (1911), p. 300.Google Scholar

^{page 36 note 4} Bulleid, A. and Gray, H. St. G., The Meare Lake Village II (1953), p. 253.Google Scholar

^{page 36 note 5} Antiq. J. I (1970), 154 and 193. Site N.Google Scholar

^{page 36 note 6} Clifford, E. M., Bagendon, a Belgic Oppidum (1961), p. 193.Google Scholar

^{page 37 note 1} P. W. Cornwall F. C. II (1960–1961), 216.Google Scholar

^{page 37 note 2} Arch. J. CXIII (1956), 43, Hut A.Google Scholar

^{page 37 note 3} B. M. Quarterly, XXXV (1971), 145.Google Scholar

^{page 37 note 4} Ibid., pl. LX and fig. 1.

^{page 37 note 5} Antiq. J. XLI (1961), 186.Google Scholar

^{page 37 note 6} Charles, Thomas (ed.), Rural settlement in Roman Britain (1966), p. 76.Google Scholar

^{page 40 note 1} Plenderleith, H. J. and Werner, A. E. A., The Conservation of Antiquities and Works of Art (2nd edition, 1971), p 250.Google Scholar

^{page 40 note 2} Museums J. LV (1955), 112.Google Scholar

^{page 40 note 3} Obtainable from W. Canning & Co., Ltd., P.O. Box 288, Great Hampton Street, Birmingham 18.