The inference drawn from the piece of the Roman road further eastward on Hakpen Hill is the same. It is perfectly distinct and quite straight for about a mile, but if it had been continued in that line, it would have passed the hill at a distance of at least 200 yards to the southward, and never have joined the other piece till long after it had passed the Devizes road. It was deflected northward in the village of Kennet, apparently to reach the bridge, and then to join the piece coming from Bath. The result of all this seems to be, that the evidence of the Roman road is inconclusive either way and must be withdrawn. Taking the point where it passes the Devizes road, and the piece which is found on Hakpen hill as fixed points, to join these it must have passed considerably to the southward of the hill; whether it did so in a mathematically straight line or in one slightly curved, was a matter for the judgment of the surveyor; but till we know his motives, it is not in our power to found any argument upon them. [Illustration: 20. Elevation of the Bartlow Hills. From the 'Archæologia,' xxx.] If, however, the Roman road refuses to give evidence in this cause, the form of the hill offers some indications which are of value. As before mentioned, it is a truncated straight-lined cone, sloping at an angle of 30° to the horizon, while all the British barrows known are domical or, at least, curvilinear in section. In all his experience, Sir R. Colt Hoare met with only one straight-lined monument of this class, which consequently he calls the Conical Barrow. Whether it was truncated or not is not quite clear. There are bushes, or weeds, growing out of the top, which conceal its form.[96] Nothing was found in the barrow to indicate its age except a brass (bronze?) spear-head, but it was attached to a British village, apparently of the Roman period, inasmuch as iron nails and Roman pottery were found in it.[97] Be this as it may, there are a range of tumuli at Bartlow, on the boundary between Essex and Cambridgeshire, which are all truncated cones, and are undoubtedly of Roman origin. A coin of Hadrian was found in the chamber of one of them, and Mr. Gage, and the other archæologists who were present at the opening, were all agreed that all the four opened were of about the same age.[98] We may therefore feel assured that they were not earlier than the time of Hadrian, though from the style of workmanship of the various articles found, I would feel inclined to consider them somewhat more modern, but that is of little consequence. The point that interests us most is, that the angle of the Conical Barrow quoted above is 45° to the horizon, that of the principal tumuli at Bartlow 37½° and that of Silbury Hill 30°. Here we certainly have a sequence not long enough to be quite satisfactory, but still of considerable value, as an indication that Silbury hill was post-Roman.