on a fourth-century site as survivals. It is not credible that coins of Cunobelin, Augustus and Nero should do that. At Reculver therefore they indicate some early occupation. They do not indeed reveal the character of this occupation, nor can it be easily conjectured from the general probabilities of the case. But they may perhaps indicate that it was brief. Unless our details mislead us, it would seem that coins of the second and early third centuries have been found comparatively seldom at Reculver, and not in sufficient quantities to show a continuous use of the spot during those years. With these faint hints we must be content. Only the spade can tell us more.

3. RICHBOROUGH.²⁸

   Richborough is situated a little north of Sandwich, near the eastern coast of Kent, two miles from the present shore. It stands on a low isolated hill, hardly a quarter of a mile in length and breadth, and hardly higher at its summit than 60 ft. above sea-level. Round it lies the marsh. North and north—west are the levels that divide it and all Kent from the Isle of Thanet, and those levels spread out east and south, where the Stour winds out to the sea. Only on the west and south-west is the belt of marsh narrow, and the higher ground of inland Kent comes nearer. It is a strange remote place, the last point of vantage (as one might say) towards the sea and the marsh. In Roman days it was different, but no less remote. Then the marsh lay open to the sea. Salt water flowed between Kent and Thanet. The tides came up to the foot of Richborough,²⁹  and its hill was an island, save on its inland side, to the south and west. Its ancient position repeats in several respects the position of Reculver, with added features and on a larger scale. Reculver is the northeast extremity of the Blean upland towards Thanet and the sea; Richborough is the farthest outlier of the Canterbury downs towards the same. Both spots were in Roman days almost isolated by sea or salt marsh, and together they blocked the channel behind Thanet and the approach to Canterbury along the valley of the Stour. But Richborough was more than a guard-house. The sea, which once washed the foot of its hill, provided to the east and north a spacious haven, sheltered (as geologists think) by the natural breakwater of Stonor beach. Reculver had probably, as we have seen, a safe anchorage
   ²⁸  Leland, ed. Hearne, vii, 1. 138; Camden (ed. 1590), pp. 265, z66; Battely, Antiq. Rutupinae (Oxon. 1711; ed. 2, 1745, here cited), sec. 10—26, pp. 21—48, with far fewer details than he gives about Reculver; Stukeley (ed. 1),p.118 and plate 97; W. Boys, Hut, of Sandwich (Canterbury, 1792), pp. 835, 865, important; King, Munimenta Antiqua (London, 1801), ii, 2—22, with plan and plates, partly based on information from Boys; T. C. Bell, Arch. Aeliana, first series, ii, 372; Thos. Wright, Archaeological Album (London, 1845), pp. 13—18, partly repeated Wanderings of an Antiquary, pp. 87—100; A.J. Dunkin, Brit. Arch. Assoc. Canterbury Meeting (London, 1845), unimportant; C. R. Smith, Antiquities of Richborough, Reculver and Lymne (London, a chief authority; short account by Lewin, Arch. xlvi, 43’; G. E. Fox, Arch. Journ. liii, 352, good remarks on structural questions. The earlier excavations are described by Dowker in Arch. Cant. viii, 1—14, xviii, 6—14, xxiv, 201—19, and J. Garstang, ibid. xxiv, 267. The systematic excavations begun in 1922 are being published by the Society of Antiquaries as Reports of its Research Committee (1926, 1928, etc.); see also Ant. Journ. viii, 523. Other books and articles describe Richborough, but contain nothing that is both new and noteworthy. For articles on details see the following notes. The objects found in recent years are mostly preserved in a small museum on the site. Other objects, especially those excavated by Rolfe about 1846, are in the Mayer collection in the Liverpool Museum; unfortunately, the provenances of many objects in this collection have been forgotten. The ruins of Richborough Castle were purchased in 1894 by the Kent Archaeological Society, vested in a body of trustees and fenced in; they subsequently became the property of the Nation.
   ²⁹  Boys states that ‘a sandy seashore,’ on which lay a shoe with a metal fibula in it and human bones, was found in the marsh in digging the foundations of Richborough Sluice, not far from the north-east corner of the fort (Hist. of Sandwich, p. 865).