474 RETURN T O TI I E M I R A N SITE [Chap. XII
the efforts which the Chinese made during the reign of the Emperor Hstian-tsung (A. D. 713-62) to defend their Central-Asian dominion against the attacks of the Arabs from the west and of the Tibetans from the south, our Chinese historical sources show that Chinese statecraft relied greatly on the help of the various Turkish tribes over which their political control extended. Among these tribes employed as auxiliaries we find repeated mention of the Karluks, of whom we know that they had become attached to the U igurs.14 Yet it is significant that when towards A.D. 790 the Tibetans succeeded in capturing Pei-t`ing, Karluks with some other Turkish tribes are mentioned among those who helped them in reducing this last foothold of Chinese power north of the Tien-shan.'
It is impossible for us to determine with any certainty to what circumstance the presence of this single Turkish document at the Miran fort, among the thousand odd Tibetan pieces, is due. It may be the solitary relic of a success which brought a detachment of Turkish auxiliaries employed by the Chinese into temporary occupation of a fort previously established and garrisoned by the Tibetans and subsequently retaken. Or the Runic Turkish record may not have originated at Miran at all, but may have been brought there from some post, possibly about Tun-huang or An_ hsi, which those Turks were still holding for the Chinese. Instead of discussing such explanations, which for the present must remain purely conjectural, it will be more useful to emphasize what the archaeological indications furnished by the ruined fort teach us. They all combine to prove that it was built and tenanted during the Tibetan domination, which began in the second half of the eighth century A. D. and came to a close about a hundred years later.
The irregular plan and construction of the fort, as well as the total absence of any Chinese records, are strongly opposed to the assumption that it could date back to a period when Chinese political control and cultural influence still asserted themselves in the Lop tract. The amount and uniform character of the Tibetan records, and of the refuse-heaps, too, in which they were embedded, conclusively prove a long occupation by Tibetan soldiers and officials. It is equally important that of the nine copper coins found at or near the fort all but two show the type with the legend Kaiyiian rung pao, introduced by the first Tang Emperor Kao-tsu (A.D. 618-27) and continued by his successors for more than a century. Of the remaining two coins, one shows the nien-hero Chêng-ho (A.D. I I I I—I 7) and the other the modern regnal name Kuang Hsü of the late Manchu Emperor (1875-1908). Both were found on the surface in the central area of the fort, having been dropped there by some visitor to the ruin. Finally, it must be recorded that neither in nor around the Miran fort could I trace any evidence that the site was inhabited in any permanent fashion during the rule of the Uigurs, who about A.D. 86o supplanted the Tibetans as the predominant power in the Tarim Basin, or during the succeeding Muhammadan period.
Thus everything points to a rapid abandonment of the site soon after the Tibetan occupation. There is certainly no obvious reason for attributing this abandonment to the effects of desiccation, i.e. to failure of the water-supply, whatever changes progressive desiccation may have since worked on this ground in the course of many centuries. Even now the river, passing in its present bed within three miles to the west of the ruined fort, carries water sufficient for the irrigation of such an area of cultivation as would supply the food necessary for whatever garrison the small fort can have held and for a commensurate agricultural settlement.1° Whether the same statement would
11 Cf. Chavannes, Tures occid., p. 142, note 2; p. 297 ; also ibid., pp. 94, 286.
15 See M. Chavannes' extracts, Ancient Khotan, i. p. 536.
la During both my longer stays at Mïrän in the winters of 1907 and 1914 the river was hard frozen, and all the water that was reaching its bed there was spreading itself in a wide ice sheet, as explained in Desert Cathay, i. p. 438. Hence
I was unable to take any measurements then of the volume of water carried by the Miran River. Nor would the volume which I might have ascertained in December r906 have furnished any real guidance as to its irrigation capabilities; for the Jahan-sai, as the ,river is known higher up where it debouches from the mountains, is fed mainly by the permanent snow and ice of the high Chimen-tagh, and its supply of