836 The Ancient Wiltsliire Tiyhes. 



know from Roman history that the Britons of the coast were a more 

 civilized race than those who dwelt further in-land ; and it will be 

 observed that the dykes suj^posed to have been made by the Belgse, 

 as they gradually expelled the British tribes who preceded them, 

 always have their foss to the northward." — Here is another incidental 

 confirmation of the feasibility of our theory, for no doubt they must 

 first of all have landed somewhere on the southern coast of Hants 

 (as the Welsh traditions point to the Isle of Wight as their first 

 place of settlement) and gradually worked their way northwards. 



I will now speak of the various dykes of which we have remains 

 in Wilts, and explain the conclusions to which, having regard to the 

 facts I have endeavoured to explain, I have come to respecting them. 



I. — BoKERLY Dyke. This is near Martin and Damerham, in the 

 southern extremity of the County, close by the Dorsetshire border, 

 and is no doubt a very ancient dyke. There is but a fragment of it 

 remaining in Wilts. It can be traced, however, southward, between 

 Holt Forest and Cranborne Chase. Dr. Guest thinks it may have 

 crossed the Stour, south of Blandford, and then have joined Combe 

 Bank, a work precisely of the same character in that vicinity. Fur- 

 ther on there is another fragment, which, as late as Hutchins' time, 

 was a "■ great ditch like Wansdyke, in the road from Bindon to 

 Weymouth, extending for several miles." Its mutilated remains 

 are still to be seen between Oare and Bindon, and at one time it was 

 continued no doubt to the coast. The bank of this dyke is to the 

 eastward, so that it seems no improbable supposition that it was a 

 portion of the western boundary of an early, perhaps the first, Belgie 

 conquest, and that all those fragments are portions of what was one 

 line of boundary. Have we any traces of the line of this dyke to 

 the north ? Though all traces, as far as I know, of earth- works have 

 disappeared, yet I think there is a name which indicates the existence 

 of such at one time. About seven or eight miles to the north-west 

 of this dj'ke, just to the north of Hindon, we have Bockerly Hill 

 and Bockerly Coppice, the one in Pertwood, and the other in 

 Chicklade. These are in the immediate vicinity of a number of 

 ancient British remains, such as tumuli, villages, &c., and all along 



