was quite beyond the limits of available labour and time ; and I might have remained in doubt as to the real character of the scanty remains, had I not subsequently become acquainted with the circumvallations of the ancient forts at Endere and Ak-sipil exactly resembling this in shape and construction. Beyond the north-east segment of wall a small bit of ground left clear of sand showed fragments of much-decayed timber, with some broken pottery and little heaps of burned earth of a red colour. Elsewhere near this ruin the strips of exposed ground, which became more extensive and frequent towards the west in the direction of the previously mentioned well, showed no pottery débris. I conclude from this that the small circular fort—to such I now think it safe to attribute the remains I could trace—was situated beyond the closely inhabited area of the settlement.
Also when I proceeded subsequently to the small ruin about 4 â miles due north of my camp which I had noticed on my first arrival, the intervening ground, where bare of sand, showed no pottery débris beyond the limits approximately marked on the plan. The clearing effected by my labourers first laid bare here the remains of a small cella (D. xvii), io feet square. The walls had completely decayed on the south and east, while on the other sides, too, only about a foot and a half of them had been left standing. Débris of frescoed stucco lying within showed that treasure-seekers had been at work here, probably not very long before my visit. Turdi declared that he had noticed the little ruin for the first time on his recent prospecting trip, and that he had then secured here the small piece of inscribed fresco, D. T. 017 (see Plate LVIII). I could not ascertain its exact position, but as it shows what looks like the front part of the L. foot of a figure standing over a floral design, it may well have come from the low portion of wall still standing. Dr. Hoernle's reading of the inscription, in cursive Brâhmi and in the Eastern Irânian language, will be found in the list below. About 8o yards to the north-east the foundation beams of a single-roomed small structure could be made out in the ground, the walls above having completely disappeared, and a little to the south-west of this a heap of rough posts and brushwood with layers of dung seemed to indicate the position of a sheep-pen.
Having now completed the description of the ruins explored by me, and of the antiquarian finds they yielded, I may briefly indicate the conclusions which the observations above recorded seem to justify as to the history of the site. Whether our interest in the latter is purely archaeological, or of a more general kind connected with the geography and cultural past of the whole region, the chronological question as to the date of the ruins and the abandonment of the site must claim our attention first. Fortunately the finds of the dated Chinese documents discussed above, together with the collateral evidence of the coins, enable us to answer this question with full assurance. We have seen that those documents dating from the years 781790 A. D., by their very character and the conditions in which they were found, prove the closing years of the eighth century as the period when the ancient settlement was abandoned by its last inhabitants. In entire accord is the evidence of the Chinese copper coins, of which altogether seventeen were picked up at the site either by myself or my men. Leaving aside those which are too effaced for identification, we have among them six coins clearly bearing the legend of the K`ai-yüan period (713-742 A. D.), two of the Chien-yüan period (758-760 A. D.), while six are without legend and of the type which was current under both the Han dynasties, and probably for some centuries later b.
In the present state of our knowledge it is not to be expected that the manuscript finds in Brâhmi script and the manifold art remains should yield chronological indications of equal
s For reproductions of specimens of these coins, see Plate LXXXIX, 23, 25, 27 and Appendix D.
0 0 2