for although I failed to penetrate the valley from below, I could nevertheless look down into it from the heights of Baibhárgiri. The two ranges of Baibhár and Sona are thus seen to approach each other, but without meeting. At the point where the distance is least is a high hill close to, but isolated from, the Baibhár range; a short way to the south-west of it is a small detached hill close to Sonagiri; close to this I could see a small tumulus having precisely the appearance of an old stûpa. The position corresponds accurately with Hwen Thsang's description, and I have no doubt in my own mind that it is the long and vainly-looked for stûpa; but all my attempts to get to it were vain, for on descending the hill only a short way, the dense jangal hides the smaller hill and the stûpa, and from the valley through the pathless jangal I found it impracticable to get to it.
There is a large cave, but quite inaccessible, on the side of Baibhárgiri, near the source of the northern tributary which joins the Saraswati before it enters the inner city. The cave appears to be natural, but from its position it appears too far east to coincide with Hwen Thsang's Asur's cave. A second small cave, and equally inaccessible, exists further east, also on the southern slope of Baibhárgiri.
If, however, the necessary time could he spared to cut down a portion of the jangal in the valley, I have no hesitation in saying that the stûpa described by Hwen Thsang could be got at. Of its existence I had ocular proof, and I think it very unlikely that I could have mistaken a natural mound or hill for a stûpa, the more so as I carefully looked at it with a powerful binocular.
Beyond this, but quite invisible even from the top of the Baibhár hill, are the bot springs of Tapoban; a thin vapour over the spot where they exist declare their position. I did not see them. The outer wall of the great fort, beginning at the north entrance and going eastwards, ascends the Vipulagiri to its summit, then descends down a spur in a southern direction and ascends Ratnagiri. From the summit of this hill two branches diverge; one descends southwards, merging into the Nekpai embankment across the eastern defile leading to Giriyak, and emerging on the opposite side it ascends Udayagiri. The other branch stretches away towards Giriyak; and tradition says it goes right up to, and embraces within its circuit, the Giriyak hill. Whether it stretched unbroken right through or not I have not been able to ascertain, but certain it is that a line of walls stretches westwards from
