Page:Archaeological Journal, Volume 8.djvu/318

 232 SILCHESTER. These djkes are not so straight as the hues of Roman road, but are curved more or less in several places ; this is the more necessary to observe, because from what Gough, in his additions to Camden, says, " A military road called Longbank and Grimesdyke, pitched with flints, runs from the south gate of the town to the north gate of Winchester," it would be supposed that this bank was in some part pitched with flints, which there is no reason to believe it ever was, as it runs half a mile on the west of Latchmore Green, where the pitched way has been opened, and where it may be seen now, it is presumed, if the surface be removed. Others, fol- lowing the above writers, have said the same thing ; but, if the matter be examined, it will most likely be found that the Roman roads were straight, paved with flints, and bedded in gravel. The entrenched line, which leaves the outer rampart between the west gate and the south gate, at the projection before mentioned, runs about 200 yards in a soutlierly direction, and then turns towards the westward,^ but is not to be seen beyond the road from Silchester Common to Latchmore Green. That which leaves the south gate of the outer rampart is scarcely to be seen in the copse for about 200 yards, but, on emerging from it, the traces are seen in the fence which has been formed on it, being a broad bank raised about two feet or more, with a ditch on the south-east side ; it crosses a small rill, where it is obscure, and thence serves as a field- way as far as the road from Silchester Common to Latchmore Green, where a pond in the road seems to have been formed in the ditch of the entrenchment ; crossing the road, it forms the south side of the lane, called the Old-house Road, for about 150 yards ; thence, bending to the south, it is large and well defined as it runs towards the brook, on each side of which, for a short distance, it is not traceable, but appears again in a broad fence as it proceeds to form the east side of the wood, on the boundary of Silchester parish. Crossing the parish boundary it continues straight as it enters the wood in the parish of Pamber, and continues to form the east boundary of the wood till we come to Frog Lane. At this ■' In the Map of Silchester, given in the road from CM Sarum, and a branch is 2n(l vol. of "Ancient Wilts.," tYiK tiini continued, which we failed to notice, as a abovc-nicntioucd is described as a "Roman " Roman road from Winciiestcr."