tion of the prevailing east-north-east wind, were least exposed to its destructive force. The other two sides which directly faced this had been wholly ground down and literally carried away by erosion. It was only after the experience gained of similar breaching at sites far away to the east that I fully realized what the wind had accomplished here, and was thus able on my second visit to the Lou-lan site in 1914 to recognize such scanty traces of the walls facing east and west as had not been almost completely effaced.
By the evening of December 22 our work at the ancient fortified station was completed. There remained now to be explored another group of ruins about eight miles away to to the west (Fig. 53) . These had been first discovered by Dr. Hedin. But as he had been able to visit them from the Loulan station only once and in the course of a single day, and had only five men with him to help in exploring them, there was obviously scope left for systematic excavation. The main point was whether we could retain adequate time for it. Our ice supply was diminishing rapidly. My anxiety was increased when Tokhta Akhun returned from the spring at the foot of the Kuruk-tagh with the report that its water was so salt that no ice had as yet formed on it. Yet the minimum temperatures during our stay at the ruins had fallen as low as 45° Fahrenheit below freezing-point. For the same reason the camels had refused to drink from it. Fortunately the return of the camels from our half-way depot enabled us to shift our camp to those ruins on December 23.
The excavations strenuously carried on there during the following five days with our contingent of labourers, in spite of cases of illness still counting thirty men, proved very
IIIF Color HighRes
