one approaches it much more closely than the ordinary traveler or inhabitant of Turfan is likely to do. One feature alone, the little red range along the fault-line, redeems Turfan from being utterly commonplace and almost uninhabitable.
Evidences of changes of climate in Turfan are as abundant as in the Lop basin. Everywhere one encounters vast beds of dead reeds, not only in regions where the digging of kariz has lowered the water-level, but even in places where the activities of man might be supposed to tend to increase the amount of water locally. Grum-Grshimailo, a Russian explorer, who visited Turfan in 1889, says that though now the playa of Böjanti is dry most of the year, Chinese records and an old song seem to indicate that formerly there was at least a large reed-swamp, if not a lake. From this and from the evidence of ruins he concludes that the water supply has greatly diminished. The plain is dotted with ruins not only in districts which are now inhabited, but in more remote regions, where no surface water is now available and the underground supply is saline. Ten such places (Assa Bulak, Pokluck, Cholak, Böjanti, Kakshal, Olpang, Tura Kariz, Chong Assa, Kichik Assa, and Kosh-Dung) appear on the map on page 297, and I heard of still others. In some cases the only relics of human occupation are adobe " turas," as Buddhist stupas are called in this region, and a few bits of pottery. Elsewhere, as at Cholak, Chong Assa, and Kichik Assa, there are ruins of forts, houses, and lamaseries. The stupas appear to indicate the sites of villages, for they are on the prolongations of watercourses in places where agriculture would
IIIF Color HighRes
