ROMANO-BRITISH HEREFORDSHIRE over it are three layers of the same material exposed lengthwise.' They also mention an aqueduct or drain opened about 1785." Most of the older writers, as Camden, Gough, Stukeley, and Williams,^" placed Ariconium here, and Magna in various other localities. Camden and Gale sought it at Old Radnor, Williams at Cwm in the same county ; others at Ledbury, or Gaer near Brecon." The first to identify Kenchester as the site of Magna was Horsley,"** while Salmon maintained that it was Credenhill Camp." Horsley showed that his view not only removed all difficulties as to distance, but that Magna was obviously on the line of the road through Kenchester. His views were, however, for some time stoutly contested. The 'niche' or 'chair ' at the east end of the site, mentioned by several of these writers, was still standing about 18 14, but was afterwards destroyed. It consisted of a mass of rubble, stone, and brickwork, resembling part of the ' Jewry wall' at Leicester," and forming an arch or vault (see Fig. i). It was also known at one time as Becket's chair." Until the early part of the 19th century the ruins must have been considerable, but the site was then cleared for cultivation. Stukeley 's plan shows that a large part of the area was arable at that time, and it is presum- ably the south-western section which was described about 18 17 as being ' a complete wilderness of decaying walls and debris.' " A view of the site as it was in 1852 is given by Thomas Wright." The Hereford Museum is said to contain a pavement found in 1821, but the date may be an error.*' In 1840-2 Dean Merewether made a partial exploration of the site. A street was traced out by the remaining foundations of the walls on either side, which were from i ft. to 3 ft. below the surface, and were 2 ft. wide by 6 ft. deep. The base of a suite of rooms and passages, forming a house of some size, was laid bare. There were traces of decorations in red plaster on the walls, tessellated pavements, and a hypocaust. The pavements had mosaic patterns in red, blue, and white, one with scroll-patterns, measuring 13 ft. by 2 ft., being in fair preservation, and apparently forming a border (Fig. 5). Other pieces had devices of fish and sea-horses, but were imperfectly preserved." Some of these are now in the museum at Here- ford.'" At the same time he found a very perfect quern, now also in the museum, and an oculist's stamp, now lost, which is described later on ; also a fibula, bone pins, tesserae, and fragments of pottery. Mr. Hardwick, writing in 1857, refers to the placing of the site under cultivation, and to the lines of streets and foundations being visible among the crops in a dry summer. He argues that the buildings of the main street, " Beautiet of Engl, and Wales, vi, 5S4, with drawing of niche ; cf. Arch. Journ. xxxiv, 353. " Hist, of Radnorshire, 43. " Duncumb, Hist. ofHerefs. i, 22. " Brit. Rom. 465 ; Duncumb, op. cit. i, 24 ; cf. Manncrt, Geografhie, ii, (Britannia), 140. " New Sum. 699 ; cf. Gough, op. cit. iii, 73 and Williams, op. cit. 43. '-' y.C.H. Leia. i, 186. ■' For illustrations see Stukeley, op. cit. pi. 85 ; Brayley and Britton, loc. cit. ; Woolhope Club Trans. 1882, pi. opp. p. 241 ; see also Wright, Wanderings of an Antiq., 39 ; Gent. Mag. xxxvii, (1852), 126 ; Arch. Journ. nociv, 352 ; and for the legend connected with Becket, Woolhope Club Trans. 1882, p. 244. " Arch. Journ. xiv, 83. " Wanderings, pi. opp. p. 35. -' Antiij. xxvi, 245. Probably the Bishopstone pavement or Dean Merewether's (see below) is intended. '" Woolhope Club Trans. 1882, p. 245 ; Arch. Journ. xxxiv, 354 ; Wright, Wanderings of an Antiq. 40 ; Duncumb-Cooke, Hist, of Herefs. iv, 113 ; Gent. Mag. xxxvii (1852), 127. '" Cf. Arch. Journ. xiv, 83, and Duncumb-Cooke, op. cit. 114, where it is inaccurately stated that the pavement was found in 1857. 179
