Page:Archaeological Journal, Volume 8.djvu/317

 SILCHESTEK. 231 to have been about 15 feet above the interior ; tlie thtch about GO feet wide, and the bottom of it about 20 feet Ijelow the top of the rampart. Though this exterior hne conforms to the shape of the ground in some measure, it does not seem to have done so altogether, and, from its irregular outline, it seems probable that it existed before the wall was huilt ; and, from its general conformity, that there was a rampart where the wall now is before the latter was built. If we compare the whole work with some of the ancient camps in Cornwall, which are supposed to be British, such as Bury down, near Lanreath, and Castle-an-Dinas, near St. Columb Major, both of circular form, we might suppose that the original outline of Silchester was British also. This may receive some confirmation from the three large dikes which diverge from it.^ One from the north gate, points towards Pangbournc on the Thames ; one from the south gate, apparently, though not exactly, continuous with the northern dyke, leads towards Winchester, and is called by Gough, in his additions to Camden,'^ as also by Stukeley, Lomfbank and Gnmeaditcli? Another, in the direction from Andover and Old Sarum, which comes up close to the ditch of the outer rampart when it swells out to the south-west projection, which, it should be observed, is not opposite either of the gates in the wall, and is therefore probably anterior to its formation, if we suppose that at this point there was originally an entrance.'* Each of these lines of entrenchment consists of a rampart and litch — the ditch being on the south-east in the two southern lines, and on the w^est in the northern one. ! ' One of tliese dikes as it leaves the - Gough's Camden, vol. i., p. 142. (Per- •ampart, and the rampart itself, are haps, from Grim, an elf, a hay, witch. Irawn in Sir R. C. Iloare's Map, in the Bosworth's Diet.) Iiid vol. of his Ancient Wilts ; but the ^ " Farther on I crossed a great Roman ■anipart is not continued on the eastern road eomin<^ from Winchester ; they call lide, where the traces are sufKciently it Long hank, and Grime's dike." (Page trong to introduce it ; and which leads 1 6.0, Stukeley.) (Ir. Albert Way to suppose, that the ^ The only way we can suppose this Imphitheatre was originally within the line to have entered the gate in the wall, ines of defence. He say.s, " an argument either the nrst or south gate, would have 1 favour of the original continuity of the been by a traverse in the outer entrench- jxterior line of entrenchment around the nient, near each gate ; for the rampart is /■hole of Silchester, may be gathered, perfect and continuous where the line S it strikes me, from the position of the approaches the outer i*ampart. To this mphitheatre, slightly beyond, or, at all sup[)Osition the present appearance of the ^ents, in a parallel line with the face of ruins offers no difficulty, except that the »e inner work, on the .side where the outer ditch could not have been filled witii iter work is now wanting." (MS. Notes, water. ov. 1849 )