, ALREADY on my first arrival in Khotan I had commenced local inquiries for ancient sites
in the desert particularly deserving exploration. It had soon become clear that the only hope of securing sure guidance lay in the collection of specimens distinctly traceable to specific old sites among those which ` treasure-seekers ' were in the habit of visiting. ` Tréasureseeking', i, e. the search for chance finds of precious metal within the areas of ancient settlements now abandoned to the desert, has been a time-honoured occupation in the Khotan oasis, offering to certain more adventurous elements • of the population the combined fascinations of a kind of lottery and a roving life. In recent years, owing to the continued demand from European collectors, the small fraternity of professional treasure-seekers have learned on their periodical visits to ancient sites to pay attention also to antiquities as a sort of secondary proceeds. Nevertheless, all the information that could be elicited at Khotan, even from persons who seemed reliable, was far too vague to permit a plan of systematic explorations to be based upon it. I accordingly gladly availed myself of the offer of Badruddin Khan, the Ak-sakal of Afghan traders, who from previous services rendered to Mr. Macartney was well acquainted with the fraternity, to organize and send out small ` prospecting ' parties which might secure me the needed specimens.
During the short stay I made at Khotan in the middle of November, after my return from the surveying expedition into the Karanghu-tagh mountains, these small parties turned up with their spoil. The objects brought to me from sites close to the oasis, such as Hanguya Tati, Ak-sipil, Jumbe-kum, did not seem to justify a belief that the remains there surviving were essentially different from those that might be found on ancient ` Tatis'. It was otherwise with the party which, under the guidance of Turdi, an old and, as experience showed, thoroughly reliable ` treasure-seeker ' from a village of the Yurung-kash canton, had gone out to visit the most distant of the locally known sites, called by them Dandân-Uiliq (` the houses with ivory '). Among the specimens brought back by them were two small pieces of fresco inscribed with Brahmi characters, numerous fragments of stucco reliefs which had unmistakably served to decorate Buddhist shrines, and also a small but undoubtedly genuine piece of a paper document in cursive Central-Asian Brahmi 1.
It appeared probable, on further examination of the ` treasure-seekers ', that the ruins from which they had unearthed these remains and which they described as situated nine to ten marches north-eastwards through the desert, were identical with the site which Dr. Hedin had seen on his memorable march to the Keriya Darya, and which his narrative speaks of as ` the ancient city Taklamakan.' He had reached it by another route from Tawakkél beyond the
I For a description of these finds see the List of Objects Plate LVIII, for the stucco fragments Plate LV.
below, under D. T. or—oil; for the fresco pieces see