RUINED SITES EAST AND NORTH OF KHOTAN
SECTION I.—THE SITE OF FARHAD-BEG-YAILAKI
A RECONNAISSANCE which R. B. Lai Singh had, during my work at Kara-dong, carried out in the desert north and north-west revealed no traces of ancient occupation beyond that site. Nor was information about old remains further down obtainable from any of the shepherds and hunters of wild camels whom we could find. So leaving Kara-dong on February 22, I moved up the river in order to meet as early as possible the party of my old ` treasure-seekers ', for whose dispatch from Khotan I had arranged through Sir G. Macartney's kind help before leaving Kuchâ. Following the route familiar since 1901, I was able to ascertain that the river's course above Yoghan-kum showed but very little change, a fact fully accounted for by the high ` Dawâns' which accompany it on both sides. On February 25 this concentration across a vast space of desert was successfully secured when a dozen familiar ` Taklamakanchis' under old Turdi's nephew Roze joined me at Kochkar-öghil (Map No. 31. c. 1).
The information they brought enabled me to settle my plans and move straight to the desert belt north of the Domoko oasis. There they had succeeded in tracking, hidden away amidst high ridges of dunes and belts of tamarisk-cones, an extensive but scattered series of ruined dwellings and Buddhist shrines which had escaped us in 1906, while at work at Khâdalik, some Io miles to the south-west. Of the very deceptive nature of the ground surrounding these remains I had a striking experience when, after moving from the Keriya River, as in 1901, to the terminal marshes of the Shivul stream (Map No. 31. B. 4), we made our way westwards across the sandy jungle tract to where the ruins were declared to be situated near the end of the Domoko-yâr. Though the route rightly taken with the assistance of our previous surveys had brought us at Camp 341 (Map No. 31. A. 4) to within 2 miles of the principal portion of the site, my experienced guides, not having before approached it from this side, failed to locate the ruins, scattered as they are in the maze of close-set tamarisk-cones.2 Not having found water we were obliged to move south-west, where the position was cleared up on reaching the site of ` Old Domoko'.
In Ancient Khotan I have already had occasion to discuss the very instructive archaeological evidence which the ruins of this extensive settlement, abandoned about i 840, furnish as to similar changes in the more distant past of this tract and about the peculiar hydrographie causes underlying them.3 So there is no need to refer to it again. The photographs, reproduced in Figs. 306, 308, of two of these last-century ruins may help to illustrate the initial stage through which the homesteads of ancient settlements abandoned to the desert, such as the Niya and Endere Sites, are bound to have passed in their time. The drift-sand was still very low, and wind-erosion had nowhere progressed deeper than about 6 to 8 inches. But there was an interesting change since 1901 to note. Within a mile of the southern end of the deserted village area I came upon newly