visible. At the time of its partial uncovering in 1861, it was thought to be hexagonal on plan, with a side 12 ft. to 14 ft. in length. This description needs confirmation, which, it may be hoped, will be forthcoming now that the redoubt is obsolete. Unlike the eastern pharos, the western was built in part of re-used materials, and contained tiles stamped CL. BR.
   Besides the sea, two roads gave access to Dover. The one, coming from Canterbury, climbed the high downs near Barham, descended into the valley of the Dour and reached Dover by way of Buckland, Chariton and Biggin Street. The other, coming from Richborough, ran due south from Each End across the hills, and appears to have finally wound its way down by Charlton Cemetery and Bridge Street into the main Dover valley, thus joining the Canterbury road 700—800 yards north of the Roman settlement. It must have served purely military or local uses. Even the Canterbury road, with its steep gradients and high ascents, cannot have been so easy for commerce or travellers as the more level route from Canterbury to Richborough.
   It remains to note the smaller, portable objects belonging to Dover. The list is surprisingly scanty, and most of the items have been already mentioned. For inscriptions we have many tiles marked CL BR, Classis Britannica ; another marked MD probably AND (?Anderida); a ring from St. Martin’s church engraved with a horse and the name Heraclides ; Samian ware stamped HABI?lSF, FRONTINI, CNNTOS (?), C.IN.T.V.SS.A and OSIN (?); a black saucer with IVI?OF and a pelvis with SAV. For pottery, there are Samian, Up-church, and commoner wares, nearly all found in burials and containing nothing earlier than the second century.⁸⁰  Other objects are two or three fibulae, an intaglio of ‘Maecenas’ on a cornelian, and the oolite statue and head, already mentioned.⁸¹  Few coins have been noted, though many must have come to light. Some were found about 1899 or 1900 at the market-place end of Cannon Street, a Second Brass of Nerva, a First Brass of Trajan, and several of A.D. 250—400, including 1 Valerian, some Tetricus, 2 Carausius, 1 Valens, and a few minims. The Priory Hill burials yielded a few coins of Severus, Postumus, and Constantine, and a bronze coin of Licinius was found in 1904 at Tower Hamlets, not far off. From the Castle Hill Stukeley notes a Diocletian; and 7 coins, which I understand were found on the spot in 1862, are preserved in the Castle church—1 each of Hadrian, Commodus, Elagabalus, Valerian (?), Gallienus, Tetricus, and Constantine.
   To these surviving or recorded remains of Roman Dover we can add its Roman name and a detail from the ‘Notitia.’ The name was Dubrae—or perhaps rather Dubra—as the Antonine Itinerary, the Peutinger Map (P1. VI), and the similarity of ancient and modern name-forms adequately prove.⁸²  At Dubrae the ‘Notitia’ places a fort of the Saxon Shore, garrisoned by an auxiliary regiment of milites Tungrecani, levied in Northern Gaul. But no
   ⁸⁰  Camden (ed. 1607), p. 243; Stukeley, Itin. (ed. 1, 1724), p. 122; Clayton, Arch. Journ. xix, 87; Puckle, Hist, of Dover (1864), p. 7; Tate, Arch.Ael. vi (1865), 183; Knocker, Grand Court of Shepway, 28 August, 1861 (London, 1862), p. 47. For the inscribed tiles see Vaillant, Classis Britannica (Arras, 1888), p. 353. He says one is in the museum at Cambridge, but inquiry there has not confirmed his statement.
   ⁸¹  Fibulae, Arch. Cant. v, p. xli, Num. Chron. 2nd Ser. x, Proceedings, p. 9 intaglio, Soc. Ant. Minutes, 24 March, 1752; statue and head, see pp 43 and 44, above.
   ⁸²  Itin. Ant. 473 ;, Duroverno ad portum Dubris mpm xiiii. Dubris is the only case of the name that occurs (Itin. Ant.; Peutinger Map; Ravennas, 428, 3; Notitia, 0cc. xxviii); it is dative or ablative, as usual in these lists. The name was apparently a neuter plural, and is connected with the Welsh dwfr and Irish dobor.