536 A STRANGE OLD LAKE BED CH. XLVIII
depression. The aneroid readings at our camp within the lake basin showed clearly that the bottom of the latter lay at least 200 feet above Besh-toghrak.
Our survey had revealed no surface formation which might be considered a permanent drainage barrier between the lake basin and the valley. The belt of dunes which we had found intervening, nowhere exceeding fifty feet in height and only two and a half miles wide where we crossed it, was not a barrage in any sense. Instead of being the cause of the Su-lo Ho waters coming to a standstill in the present terminal basin, it was only a sign and result of the latter getting partially dried up ; for manifestly the sand materials contained in those dunes were but the silt deposited by the dying river in the basin and the erosion products scooped out from the dry portions of the latter, both carried away by the prevalent east winds and heaped up on the opposite lake shore. It is exactly in the same way that I invariably found the dried-up terminal courses of rivers in the Taklamakan flanked and headed by ridges of sand or Dawans particularly high. These, too, were built up with the silt which the dying rivers could carry no farther, and to some extent also with the products of the wind erosion proceeding over the area once protected by moisture and vegetation.
So, if my interpretation be right, I was confronted here at the very threshold of our new ground of exploration by evidence of that dominating factor among physical changes in the Tarim Basin, desiccation. I had learned by practical observation how much its effects could alter the surface aspects of an area within the relatively short space of historical periods. But at the same time repeated experience has taught me how clearly subsoil drainage often preserves, even in these parched-up regions, distinct traces of river courses which have long ago ceased to carry life-giving moisture into otherwise arid wastes. I was thus
prepared to give due weight to the evidence which the relative abundance of desert vegetation between Kum-
kuduk and Besh-toghrak afforded as to drainage still finding its way underground from the present terminal lake bed of the Su-lo Ho down the valley leading to Lop-nor.
IIIF Color HighRes
