broken pottery, some burnt wheat and burnt cordage (?), and various small objects, in particular a little votive ‘thumb’ in bronze.³⁶  Dowker inferred from the burnt wheat the existence of large granaries. But nothing in the recorded remains really resembles the granaries on other Roman sites. We may think rather of poor cottages, the homes and shops of fishers and harbourmen. It must be confessed that the evidences accord little with our ideas of a flourishing town and port.
   To the south of the fortress four sites have produced structural relics of importance. On the highest part of the hill, 450 yards from the fortress, are the buried remains of a small amphitheatre.³⁷  This has been known and recognized since Leland, who says that in his time it was called Littleborough. Roach Smith partially excavated it in 1849. It proved to be elliptical in shape, measuring 166 ft. by 200 ft., and enclosed by a flint wall 3½ ft. thick, faced on the outside with local chalk and on the inside with mortar. From the bottom of the inner face of this wall a bed of mortar or cement, 2 in. thick, spread 15 ft. towards the centre of the arena, while on this bed and against the wall rested a sloping bank of mixed

 clay and mortar, 8 ft. long at the bottom and 7 ft. high at its highest—perhaps a support for wooden seats. Lastly, three entrances were identified, a large one on the north and two smaller ones on the south and west. On the ruined wall of the western entrance a skeleton was found, ‘lying on its left side, the legs drawn up and the wrists crossing each other. The place had evidently been hollowed out for its reception : most of the bones of the hands and feet were wanting: but. where the right hand had been, a brass coin of Constans was found.’ This was presumably a post-Roman burial, possibly

of an executed criminal. The general date of the amphitheatre may be indicated by the fact that ‘the coins found among the ruins are confined to the period extending from the time of Gallienus to that of Arcadius, with the exception of one of earlier date—a denarius of Domitian.’ The period thus suggested is that of the Saxon-Shore fortress, and it may be found on further exploration that the amphitheatre was the work of the garrison—the Second Legion—which had long possessed a serviceable amphitheatre at its former station, Caerleon.
   East of the amphitheatre, during the laying of a colliery railway in 1926, the remains of two small temples were exposed and largely destroyed. The
   ³⁶ Stukeley (ed. 1724), 118; Boys, Sandwich, 868; Dowker, Arch. Cant. xviii, 5. A platform of square tiles outside the north wall is recorded by Hasted (iii, 689), but too vaguely to be used as evidence. See below, p. 128.
   ³⁷  Leland, ed. Hearne, vii, f. 138; Boys, Sandwich, 867; Nichols’ Bibl. Topogr. Brit. i, 47; C. R. Smith, Richborough, etc., 52, 161.