Sec. iv] TEMPLE REMAINS NORTH OF KHOTAN 1279
would have involved the shifting of huge masses of sand and a far greater expenditure of time than any likely results would have justified. The abundant moisture permeating the lower sand strata was bound to have affected any other structures at this site quite as much as those already laid bare. The condition to which this moisture must have reduced any smaller relics of perishable substance was illustrated by the small fragment of a wood-carving, Ta. i. oo8, almost wholly rotten, showing a seated Buddha. The other small finds made included decorated terra-cotta pieces, Ta. 004-5 ; i. 003, very closely resembling those from the Yôtkan débris strata. Among other pottery remains some fragments of a vase with green glaze, Ta. 003 ; i. 001-2, may be specially mentioned.
In view of the interest attaching to the relics brought to light here of Khotanese pictorial art, Chrono-
and also in view of the position of the site far away from the present limits of the oasis, any indica- logical evi-
tions as to the date of the latter would be specially welcome. The safest evidence at present site.
available is that of the coins. The two found at the ruin itself are the Wu-clzu piece already mentioned and an uninscribed copper coin from the foot of the wall ii. In addition eight poorly preserved Chinese coins were picked up on a small and completely eroded ` Tati ' area to the south of the site. They are either much worn Wu-chu pieces or else uninscribed.15 The collective evidence of the coins thus points to the site having been abandoned before Tang times. In the present state of our knowledge no closer date limits can be safely deduced either from the style of the paintings or from the palaeographic character of the two inscriptions in cursive Central-Asian Brâhmi found by their side. But if my tentative reading of three characters on a small fragment of coloured wall-plaster which was found in loose sand at the first trial excavation, Ta. or, as written in Kharosthi should prove right, we should have to push back the upper date limit for the occupation of the site to about the fifth century A.D.18 In respect of the position of the site it should be noted that it lies almost Abandon-
exactly in the same latitude as the northernmost of the Kine-tokmak ruins, about 13 miles away to ment before
the east on the other side of the Yurung-kâsh. For these the probability of abandonment before period.
the Tang period has been shown above,17 and this lower date limit is certain in the case of the Rawak
Vihâra, which lies only 3 miles or less further south.18 . In any case we have thus archaeological
proof that the area of the main Khotan oasis on either side of the Yurung-kash River must have
extended much further north than it does at present.
On April II I moved down by the river and by a long march reached the oasis of Islâmâbad Oasis of (Map No. 27. B. 2), the northernmost outlying settlement of the Khotan district. Even here there Islamabad.
was striking evidence of the rapid expansion which Khotan cultivation had been undergoing for some time past ; for owing to the construction of a large new canal about 1896 the seven to eight holdings formerly cultivated by settlers from Tawakkél on the opposite bank of the river had grown into a large colony consisting of about 400 households and capable of supporting many more. A day's halt at Islamabad was needed for drying and packing the fresco pieces brought away from Mayaklik, and also enabled me to secure welcome information about the ruins at the desert hill of Mazar-tagh for which I was bound. What I had heard about them in 1 goo-1 and also on my subsequent visits to Khotan had sounded extremely vague."
But now ten narrow slips of wood, inscribed in Tibetan like those found in the Miran fort, were brought to me by Tawakkél men who the year before had prospected for ` treasure ' at a ruin on the hill-top. Thirteen more were put into my hands by Kasim Akhan of Islamabad, the son of my old guide Ahmad Merghen,20 who had died a few months earlier. The old hunter had secured them from
15 See Appendix B. 18 Cf. Ancient Kholan, i. p. 451.
16 The frescoed fragment showed unusual hardness at the 18 My doubts about the provenance of the coin batch said
time of discovery. Its original provenance could not be to have come from Mazar-tagh and detailed in Ancient
traced. Kholan, i. p. 579, were fully justified.
" See above, p. 131. R0 Cf. Ancient Kholan, i. pp. 237 sq.
Finds obtained at Islâmâbad.