^{page 87 note 10} Trans. Bristol & Glos. Archaeol. Soc. lxi (1939), 202–23.

^{page 88 note 1} Théophile Eck, Les deux cimitières galloromains de Vermand et de St. Quentin (1891), pl. xvi, la; 6 and 9 have a narrower spacing of beading or grooving on the spindle-housing, more like that of the Sea Mills specimen. These finds are also shown in Leeds, E. T., Early Anglo-Saxon Art and Archaeology (1936Google Scholar), pl. v, lower right.

^{page 88 note 2} Boulanger, C., Le Mobilier funéraire galloromain et franc en Picardie et en Artois (1902–5Google Scholar), pl. VII, 4.

^{page 88 note 3} T. Eck, op. cit., in note 2, pi. xvi, 2b (found with 1 a).

^{page 88 note 4} For example, Faussett, B., Inventorium Sepulchrale (1856Google Scholar), pi. ix.

^{page 88 note 5} For examples from Silchester, Hants, see Medieval Archaeology iii (1959Google Scholar), pls. iii, iv.

^{page 88 note 6} For example, cf. a Lydney specimen, presumably Valentinian in date, identical with one from Silchester: R. E. M., and Wheeler, T. V., Excav. in Lydney Park, Glos. (Res. Rept. Soc. Antiq. Lond., ix, 1932Google Scholar), pi. XXVIII, 132.

^{page 89 note 1} From personal inspection.

^{page 89 note 2} Military occupation ends c. 383–8, apparently; see Nash-Williams, V. E., The Roman Frontier Wales (1954), 143Google Scholar.

^{page 89 note 3} Including several hoards. Bulletin of the Board of Celtic Studies, ii (1925), 92–97Google Scholar.