Sec. i] AMONG THE RUINS OF ANCIENT KAO-CH`ANG 595
merit, remarkable for the lifelike ease with which the animated poses are rendered, and for the freedom of the drawing in all details. There is nothing to afford a definite indication of date ; but judging from external conditions the print was probably found in a tomb. From tombs may possibly also have come the two bronze mirrors, of and 034 (Pl. LXXI), the former showing on its back a Chinese landscape in very low relief and the latter four Chinese lapidary characters. The small stucco head of a woman, 024 (Pl. LXIX), is probably from a similar source. Among small stone carvings the jade ` Netsuki ', 02 (Pl. LXXI), representing a reclining dog, and the jade fish, 013 (Pl. LXXI), are manifestly of Chinese workmanship. The steatite mould of a heart-shaped leaf, 016 (Pl. LXXI), probably served for producing metal ornaments. Bronze objects of interest are the mace-head, 028 (Pl. LXXI), resembling one found at Yôtkan, and the prism-shaped ornamented weight, 031 (P1. LXXI). The fragments, oil (Pl. LXXXVII), of a fine bamboo matting, interwoven with a figured silk, closely resembles the manuscript-roll covers of similar make from the ` Thousand Buddhas ' and may have served the same purpose.
During my first stay at Kara-khôja I took occasion to test a persistent rumour that ancient ruins never yet explored were to be found in the rugged outer range overlooking the oasis from the north-west. I had already heard of them in 1907, when they were said to be hidden in the barren hills to the south-west of Murtuk. Again, on my passage through Pa-no-pea, Ahmad, the ill-fated rebel, stated that he had seen them once, while hunting in the hills about the gorge known as Kuruk-aghiz. He had offered to guide me to them—once he had made his peace with the Chinese local authorities. Several men of Kara-khôja, no doubt eager for a new ground to search for treasure or antiques, offered themselves as guides to Kuruk-aghiz. So I let Afrdz-gul proceed to that valley, provided with a sufficient supply of water to permit of a prolonged search in that barren wilderness of eroded hills. The valley was found to debouch to the north of Sai-langar, a resting-place on the road from Turfân town to Sengim-aghiz (Map No. 28. c. 3), and to be formed higher up by a maze of deep-cut and extremely narrow gullies 1Ô But two days' thorough search of them, carried right up to the watershed above Murtuk, failed to reveal any trace of former occupation.
In the light of subsequently gathered information it became clear, in fact, that the alleged ruins were but the local pendant of those legendary ` old towns ' in the desert that haunt the imagination of people in certain of the smaller oases along the northern edge of the Taklamakan. Those ` old towns ' once sighted by adventurous hunters, &c., can never be found again ; for the magic of evil spirits is supposed to hide them, when searched for. Just as in the case of the Kôtek-shahri, in search of which Musa Iiâji of Korla had led me in January, 1908, into the desert north of the Inchike-darya,11 the motive prompting our Kara-khôja informants had obviously been the hope that my own European magic would prove superior to the wiles of the local demons and reveal to them a new and rich site to exploit.