148 ROUTE ACROSS DRIED-UP LOP SEA CH. IX
once leading eastwards of Lou-lan might reveal. In order to assure adequate time for the latter rather hazardous undertaking, it was essential to effect excavations rapidly and therefore to take along as many labourers as I could possibly manage to keep supplied with water, or rather ice. What with big loads of ice sufficient to assure the minimum allowances of water for thirty-five people for a month, with food supplies of one month for all and of an additional month for my own men, and with indispensable outfit to afford protection in the wintry desert exposed to icy gales, the thirty camels I succeeded in raising, including our own fifteen, were by no means too many. It goes without saying that everybody had to walk.
By February r, 19 I 4, I had safely started this large column from Miran. Next day we took up our supply of ice packed in bags from a terminal lagoon of the Tarim. Thence four marches brought us to my immediate goal, a large ruined fort which had first been sighted some years before by Tokhta Akhun, my faithful old Loplik follower. Wind erosion had deeply scoured the ground outside and in places had completely breached the very solid enclosing rampart (Fig. 63). It was built of alternate layers of brushwood fascines and stamped clay, after the fashion observed before in the ancient Chinese border wall west of Tun-huang. Plentiful relics in the shape of architectural wood-carvings, implements, coins, etc., were recovered by clearing what remained of dwellings within. They proved occupation to have ceased about the same period as at the Lou-lan site. A well-marked dry river-course passing the fort was easily traced by the rows of fallen dead trees lining its banks. As proved by its direction, it was a southern branch of the
IIIF Color HighRes
