350 THE LOP-NOR PROBLEM.
within its catchment area in the course of 30o to 40o years for the river to have dried up in the sand . 132 km. distant from the Tarim? It is not very probable.* Another cause, or rather several other causes, must be found for this stream having been cut off from the main river. It may in part be due to the fact that during the 30o to 400 years in question the oasis of Kerija has been more densely inhabited. Another contributing circumstance may be the forest, and the belt of vegetation which is spreading along its banks, setting up a greater drain upon its volume, though on the other hand the increased shade and the generally moister surroundings will have tended to augment the river's power of resistance to drought
Fig. i72. THICK FOREST AT KATAK, CAMP XIX, KERIJA-DARJA, 1896.
as compared with the time when these forests were scantier. Finally, it may also bq ascribed to the extension of the sandy desert and the increase in the amount of sand, resulting in the choking of the northernmost reaches of the stream; or rather in compelling it to take an underground course, an inference suggested by the presence of groups of fresh poplars along the north-north-east prolongation of the river. To these there have of course to be added certain climatic factors, at any rate in part; though what their real nature is we know but little.
Perhaps the same forces have affected the Chotan-darja. Anyway, it would be difficult to prove that the rivers of East Turkestan have on the whole suffered
* In a note to p. 79 of Vol. I. I have expressed a different opinion as to the Kerija-darja. But now, after having studied the problem more thoroughly, I find the opinion expressed above more likely.
IIIF Color HighRes
