^{page 85 note 1} For abbreviated references used in the inventory and in the body of this paper, see p. 102.

^{page 85 note 2} Clark, , Prehistoric Europe, 213–5Google Scholar.

^{page 85 note 3} Piggott and Daniel, 16–17, Pls. 28 and 29.

^{page 85 note 4} Ant. J. XX, 1940, 53 ff.Google Scholar

^{page 85 note 5} op. cit., 213.

^{page 86 note 1} Bulleid and Gray, fig. 98, p. 327.

^{page 86 note 2} e.g., Zurich-Alpenquai. Clark, op. cit., fig. 116, p. 215.

^{page 86 note 3} Arch., LII, 1890, 360–4Google Scholar.

^{page 86 note 4} Bulleid and Gray, fig. 65, p. 313.

^{page 86 note 6} Sir Arthur Evans's suggestion, when first publishing the Aylesford tankard (Arch. LII, 1890, 357Google Scholar), that the origin is to be sought in native woodwork, and citing as evidence a tankard found in a bog in Northern Ireland, does not appear to fit the facts. This vessel from Ireland and now in the Ashmolean Museum is carved from solid wood, its handles of one piece with the rest. Miss Joan R. Kirk of the Ashmolean, in a letter to the writer, was unable to give a date but compares other vessels of similar construction with an 18th century date stamped on them. Mr A. T. Lucas of the National Museum of Ireland puts some of the Ashmolean type a couple of centuries earlier, but no more. He does not believe that any of this type are of prehistoric date. One may note however certain pottery mugs of the Early Bronze Age which appear to be skeuomorphs of wooden vessels carved from the solid (Antiquity, IX (1935), 34–8)Google Scholar.

^{page 86 note 6} Allen, Romilly, Arch. Cam., 1896, 214Google Scholar.

^{page 86 note 7} A similar insertion of metal to bind together a circle of wooden construction was used in the nave of a chariot wheel found at Bar Hill in Scotland. In this case the metal used was iron and the technique applied on a much larger scale. Macdonald, and Park, , The Roman Forts on Bar Hill (1906), 92–9Google Scholar.

^{page 86 note 8} Bulleid and Gray, fig. 65, p. 313.

^{page 86 note 9} Maiden Castle Report, fig. 72, no. 185, p. 232–3; fig. 74, no. 227, p. 237–8.

^{page 86 note 10} ibid., 233.

^{page 87 note 1} Cardiff Naturalists' Society's Reports and Transactions, LXXX, 1948–1950, 40, pl. 4, 2Google Scholar.

^{page 87 note 2} This vessel will not be mentioned to any great extent in the general discussion. The triquetra motive is so well known in Celtic art in this country that discussion of it is not thought germane to this paper. (But see Fox, , Llyn Cerrig Bach Report, p. 46Google Scholaret seq. for a discussion of a new example of this motive, the now famous crescentic plaque).

^{page 88 note 1} The pottery mugs in the war-cemetery at Maiden Castle have handles which do not seem to have been adequate enough for grasping in the fist. Wheeler, Maiden Castle Report, fig. 72, no. 185, fig. 74, no. 227.

^{page 88 note 2} Bushe-Fox, , Swarling, 17–18, 23Google Scholar. Hawkes, and Dunning, , Ant. J., LXXXVII, 1930, 260Google Scholar.

^{page 88 note 3} Fox, , Arch. Cam. Region, p. 99Google Scholar; Arch. J., XCVI, 54, 55Google Scholar.

^{page 90 note 1} PRW, 210.

^{page 90 note 2} Ant. J., XXVII, 5Google Scholar.

^{page 90 note 3} On the other hand Neath v may have been made by an inexperienced craftsman unable to copy his contemporaries' skill, and the other articles from the hoard do not favour as late a date as Newstead or Bartlow handles which are similar to Neath v.

^{page 90 note 4} Willers, p. 19, pl. 12,8.

^{page 90 note 5} Piggott and Daniel, 9–10; Leeds, 56 (‘Flamboyant style’).

^{page 90 note 6} Hencken, , Arch. Cornwall, 110Google Scholar.

^{page 91 note 1} Cf. infra, p. 93.

^{page 91 note 2} Cf. Llyn Cerrig motive on the crescentic plaque, pp. 50–1.

^{page 91 note 3} But cf. supra, p. 90, note 3.

^{page 91 note 4} Neath IV may have been influenced by the handles from south-eastern Britain where openwork handles do not appear to have been favoured.

^{page 93 note 1} PPS, XVI, 1950, 17, 18, 27, fig. 9, 3AGoogle Scholar.

^{page 93 note 2} Navarro, De in The Heritage of Early Britain, p. 81Google Scholar and pl. 12b. ‘Late boss style.’

^{page 93 note 3} Neath III is at a transitional stage from Class la to Ib, as the idea of using the double rivet holes as an integral part of the decoration has not yet occurred to the craftsman.

^{page 93 note 4} Very much classical influence.

^{page 93 note 5} Neath IV is an exceptional piece; strictly speaking it fits into Class IIb but it is so Celtic in feeling and obviously a local treatment of the handle theme, perhaps owing much to Neath in, that it must be regarded as distinct from the other handles in Class II with their strong classical influence.

^{page 93 note 6} The Kew and Porth Dafarch handles are not true mirror-handle types in the sense that Bulbury is, but are included in this category to prevent undue complications. It might even appear that Bulbury is merely a true descendant of the Ornavasso type influenced locally by the popular mirror handle style.

^{page 94 note 1} Arch., LX, 289, fig. 35Google Scholar.

^{page 94 note 2} BMEIAG, p. 143, fig. 162.

^{page 94 note 3} Perkins, Ward, PPS, 1939, p. 175, fig. 2Google Scholar.

^{page 94 note 4} ibid., 175.

^{page 94 note 5} Cf. supra, p. 93, note 5.

^{page 94 note 6} But cf. supra, p. 93, note 6.

^{page 94 note 7} PPS, 1939, 173Google Scholar.

^{page 94 note 8} p. 61 ff.