[1075] Trans. Hon. Soc. Cymmrodorion, 1898-9 (1900), p. 20.
[1076] See Proc. Soc. Ant. Scot., xxix, 1895, pp. 131, 149-50.
[1077] B. G., vii, 22.
[1078] Proc. Soc. Ant. Scot., xxv, 1891, pp. 428, 438, 440, 444-5.
[1079] Ib., xxxiii, 1899, pp. 15, 20-3, 26-32; xxxiv, 1900, p. 74. A similar method of fortification was practised by the Dacians (Congrès archéol. de France, 1874 [1876], p. 444), ‘in the Danne-werk at Korborg, near Schleswig’ (A. Pitt-Rivers, Excavations in Cranborne Chase, iii, 254), and in Nassau (Rev. de synthèse hist., iii, 1901, p. 45).
The well-known camp on Herefordshire Beacon is interesting because, like Old Sarum (Sorbiodunum), it contains a citadel. Though it is locally described as a ‘British camp’, its date is at present uncertain. While most of the objects which have been found in it are comparatively late, Pitt-Rivers (Journ. Anthr. Inst., x, 1881, p. 331) pointed out that the pottery seemed to indicate its Celtic origin; but the citadel presents a difficulty. Was it a later addition? See also F. J. Haverfield, Archaeol. Survey of Herefordshire, 1896, pp. 3-4.
The ‘vitrified’ stone forts of the British Isles demand a brief notice. There are none in England, but many in the northern and western counties of Scotland and some in France. It is very doubtful whether any exist in Wales or Ireland (Archaeol. Journal, xxxvii, 1880, pp. 227, 234; D. Christison, Early Fortifications in Scotland, pp. 187, 190). The question is whether the vitrifaction, which was due to fire, was accidental or designed; and in some cases the only way of settling this is to ascertain by excavation the extent of the vitrifaction (ib., p. 192). The best authorities have concluded that when the vitrified part of the fort is small the phenomenon may be safely ascribed to accident,—perhaps to a beacon fire; but that when it may be traced almost all round the rampart it was intentional (ib., pp. 186-7; Archaeol. Journal, xxxvii, 1880, pp. 240-1; R. Munro, Prehist. Scotland, pp. 382-3). Probably the builders intended to give cohesion to the walls and make it impossible for assailants to demolish them (L’Anthr., xiv, 1903, pp. 330-1); or when the vitrifaction was confined to the upper surface the defenders would have secured firm foothold while the assailants would have stumbled over loose stones (D. Christison, op. cit., pp, 186-7). [See Addenda]
[1080] Reports Archit. Soc. of ... Lincoln, &c., xviii, 1885-6, pp. 53-61; Archaeologia, lii, 1890, pp. 382-4; Vict. Hist. of ... Northampton, i, 147-9, 151-2. At Beansale and Claverdon in Warwickshire there are camps which in many respects resemble that of Hunsbury, but have not been excavated (Vict. Hist. of ... Warwick, i, 350).
Professor T. McKenny Hughes (Archaeologia, liii, 1892, p. 484) suggests that Offa’s Dyke may have ‘belonged to the defensive system of the Britons’. All we know is that those dykes which have been excavated—Bokerly Dyke and Wansdyke—were Roman or post-Roman (A. Pitt-Rivers, Excavations in Cranborne Chase, iii, p. xiii); and it is in the last degree improbable that earthworks which extend over territory that belonged to several tribes should have been constructed at a time when tribes only combined for brief periods and in the presence of urgent and common peril. Cf. F. J. Haverfield, Archaeol. Survey of Herefordshire, 1896, p. 7, and Eng. Hist. Rev., xvii, 1902, pp. 628-9.
[1081] See p. 93, supra.