had made primarily for Richborough. Part of the Claudian army may have landed here in A.D. 43 (p. 2). Nevertheless, pottery and coins earlier than the end of the first century seem to occur there rarely or not at all, and the development of the place must have waited until the second or even the third century. The comparative frequency of fourth-century coins suggests a more definite occupation in that period, coincident with the régime of the Saxon Shore. To the end, however, Dubra remained a small military or naval station. It gave no anticipation of the importance of medaeval and modern Dover as the bridgehead of the Calais crossing.

5. LYMPNE.⁸⁶

   The village of Lyrnpne stands on the edge of a long shelf of high land which overhangs the north side of Romney Marsh. From its church and houses a grass-grown slope drops down steeply and brokenly, falling 300 ft. in 600 yards, amid much disturbance of landslip, to the military canal, which forms the landward limit of the levels, and coincides roughly with the high-water mark of the sea. Near the bottom of this slope lie the ruins of a huge Roman rampart, long known as Stutfall Castle. Its highest point is about 150 ft. above high-water mark, its lowest extremity comes down to the marsh and canal. Across the marsh, a mile and a half to the south-eastward, the sea beats on the beds of shingle heaped up by itself, and on the artificial defences of Dymchurch Wall. In Roman days it probably came nearer. Its exact ancient line cannot now be determined. But it seems certain from archaeological evidence that a harbour existed here in the Roman period. Lymne, as we shall see, was called a harbour by the Romans, and its waterside ruins, as well as the analogy of the other Saxon-Shore defences, require a harbour close to it. Geological evidence, if not very decisive, agrees with this. The shingle and blown sand near West Hythe suggest an inlet of the sea between Hythe and Dymchurch, which came close to Lympne. The only question seems to be whether this inlet ended at Lympne or formed part of a river estuary. Early documents mention a river Limene which rose in the Weald, and probably flowed past Lyrnpne, and which certainly seems to be connected by its name with Lympne. This river may be an earlier channel of the Rother, which now emerges into the sea near Rye, but which may be conjectured to have then made its way out between Hythe and Dymchurch. Leland goes so far as to declare that a memory of a tideway existed at Lympne till his time in the names Shypwey and Old Haven, and the name Shipway (whether or no it have any connexion with ships) still survives. But no other trace of such a tradition is recorded,⁸⁷  and it would be worth little, however well recorded.
   ⁸⁶ For early mentions see Talbot, ed. Hearne, Itin. of Leland, iii, 158; Leland, ed. Hearne, vii, f. 41 Lambarde, Perambulation of Kent (1576), p. 145; Camden (ed. 1607), p. 246, and Stukeley, Itin. p. 124 (ed. 2, p. 132). Hasted, iii, 442, Gough, Britton, Beauties, viii, 1136, etc., add nothing. Excavations were undertaken by Mr. Roach Smith and his friends in 1850, and to these we owe practically all our knowledge of the site: see C. R. Smith, Richborough, etc., p. 233, and Report on Excavations at Lymne (London, 1852) and Thos. Wright, Wanderings of an Antiquary, pp. 124—36.
   ⁸⁷  For the geological evidence see Topley and Drew, Geology of the Weald (Memoirs of the Geol. Survey, London, 1875), pp. 303—6. For the Limene, see Anglo-Saxon Chron. A.D. 892—4 and charters cited by C. R. Smith, Topley, etc.; the passage often quoted from Asser is not by Asser and has little value. For Shipway, see Leland, ed. Hearne, vii, f. 41: the name might, of course, mean only ’sheeptrack.’ On the genera] question of Romney Marsh in Roman times see T. Rice Holmes, Ancient Britain, 532—52.