________________
JULY, 1917]
A THIRD JOURNEY OF EXPLORATION IN CENTRAL ASIA
139
After a stay which reunion with the kindest of friends, Sir George and Lady Macartney, and the glorious autumn season had combined to render most pleasant, I left the Kashgar Consulate-General on October 9 for my first winter's work in the desert. Its main goal was the region around the dried-up Lop-nor, in the extreme east of the Tarim Basin, and the whole length of the Taklamakan, that great sea of drift-sand over 600 miles in a straight line, separated me from it. A variety of considerations obliged me to revisit Khotan, and once there I was bound to proceed by the only possible route which skirts the southern edge of the Taklamakan. Much of the ground to be traversed there was already familiar to me from my previous expeditions, and for this reason I was all the more eager to use whatever chance of new routes the limits of time left me on my way to Khotan.
This induced me to move first due east to the oasis of Maral-bashi along the foot of the steep and barren mountain chain which forms here the southernmost rampart of the Tienshan. It had in its main part remained so far unsurveyed, but reports, previously collected, seemed to indicate that an old route, now but vaguely remembered in local lore, had during earlier periods of Chinese domination skirted the foot of that chain and been in use for traffic instead of the present high "road," recte caravan track, leading much further south along the actual course of the Kashgar River. The accurate survey now effected confirmed that tradition and proved the existence of a series of small ruined sites echeloned along a line of some 160 miles and dating from pre-Muhammadan times. The ground occupied by them on the gentle desert glacis of the hill chain is now wholly without water. There were also other physical observations of interest to be gathered, clearly pointing to desiocation within historical times, and not explainable by the fact that the winding bed and inundation marshes of the Kashgar River were found to have at one period, perhaps relatively recent, approached that desert glacis in places more closely than they do at present. We had met with serious trouble about water, no drinkable supply having been found on three Successive marches. This served as a suitable preparation for difficulties to be faced on our desert travel ahead.
By October 18 we were glad to reach the fields and fruit-gardens of Maral-bashi. The survey of some badly injured Buddhist ruins in the vicinity and of the curious canal system by which the oasis obtains the major portion of its by no means abundant irrigation from the Kashgardarya here approaching its end, ocoupied me for a few days. But in the main my short stay was taken up with careful preparations for the attempt I planned to make my way to the desert hills of the Mazar-tagh on the lower Khotan River by a short cut through the Taklamakan. I knew well the formidable obstacles and the risks presented by the wide intervening belt of absolutely waterless drift-sand desert. But by sending all baggage, except an absolutely necessary minimum, to Khotan by the caravan route vid Yarkand; by reducing in the same way my camp to a few indispensable followers, and keeping most of our fine camels for the transport of water in my six galvanized iron tanks and the very numerous goatskins I had brought from India, I could hope safely to overcome the difficulty about water. The advent of the cold season would help onr brave camels to face a long fast from grazing and water.
Apart from the attraction presented by the short cut and the fascination of such a desert cruise, there was an important geographical task to justify the enterprise. Our surveys of 1908 had shown reason for the belief that the Mazar-tagh hills, thed traced for some 20 miles into the Taklamakan, belonged in geological structure to an ancient range which started at an angle from the outermost Tien-shan near Mara-bashi and once extended across the
