north-east, led necessarily near the one followed by Hedin, in 1900, in the reverse direction. But there was nothing to guide us except the position of the ruins as indicated in his route-map and the compass. A notable change, moreover, had taken place in the physical aspects of this dismal ground since Hedin had traversed it from the north. The great newly formed lagoons in which the waters of the Tarim had in that year of an exceptional spring flood spread northward, had since completely dried up. The water of the rare pools left behind in the salt-encrusted beds of these lagoons was so salt that in spite of the great cold it had not yet frozen.
By the evening of December 14 we had left the last depression with dead poplars and tamarisks behind us and camped between high sand-cones covered with hoary tamarisk growth. There all the bags of ice brought on the donkeys were carefully stacked on the north side of the highest sand-cone to form a depot, and the donkeys sent back in charge of two extra men. They were then to return from Chainutköl and bring up the reserve food-supplies of the labourers.
Starting off again, we passed very soon into that zone of excessive wind erosion which constitutes so striking a feature of the northern portion of the Lop desert. It consists of an endless succession of steep clay banks or terraces, separated by sharply cut trenches (Fig. 58) . These are known to the Lopliks by the convenient designation of `Yardangs'. They have all been carved out by the winds with the help of the sand which they drive before them and which thus serves as their instrument of corrosion. These terraces or Yardangs invariably run from east-north-east to west-south-west and thus clearly mark the prevailing direction of the strongest and most persistent winds, which sweep down for the greater
IIIF Color HighRes
