AND THE PROBABLE DATE OF STONEHEXOE. 110 ditch made by the BclgcX we must suppose, tluit altliou"li the invaders were strong enough to capture such a fortress as OKI Sarum, they were not jiowcrful enougli to j)ossess themselves of the valleys which it commanded — an inference which at once shows us the falseness of the premiss that led to it. With respect to the connection supposed to have existed between Combe-bank and Bokerly-ditch, it may be right to state, that I have not examined the course of ]5okerl3^-ditch west of the Roman Road, and only cursorily tiie line of country which intervenes between the two earth- works. Combe-bank still crosses the down, in fine preser- vation, from the neighbourhood of Winterbourne Clenstone to Col-wood. For some distance it forms the boundary of this wood, and then enters it and disappears. My guide ^ professed to trace the bank to the north of Mapperton, but I nuist confess that to my eyes it was invisible. Its course, however, when I last recognised it, pointed eastward in the direction of Badbury, which was full in sight, and about four miles distant. I felt a strong conviction that the information given to Leland (according to which it went to Lytchet Maltravers) Avas erroneous. It seemed to me clearly in- tended as a boundary to separate the Winterbourne valley from the bleak and swelling downs to the north-eastward, and to be as clearly connected with the great fortress, which lifted itself aloft on the other side of the Stour directly in our front. As Badbury commands the valley, where lay Vindo- gladia — which existing remains, as well as the Itineraries, point out as the capital of the district — and as Bokerly-ditch was obviously intended as the northern boundary of this valley, it seems difficult to escape the conclusion that both Combe-bank and Bokerly-ditch were constructed as parts of one design, by the same people, and at the same, or nearly the same period. That people we may conjecture to be the Belga}, and the period five or four, or, it may be, only three centuries before the Christian era. I The general consent of our antiquaries has fixed upon the jWansdike as the last of the Belgic boundaries. Were it called the last frontier of the Belgic province — understanding ' His testimony must not be altogether proclamations duly made on this ancient rejected, as he has for years " cried the earth-work, the courts are held in the Courts " at the bank, and, therefore, may valley at an old manor house, which lies be considered as familiar with all the some two miles from tlie bank, lircumstances connected with it. After VOL. VIII. X
