over approximately . 500 miles of route, and for the last 180 miles or so, no intersections could be obtained on it owing to the absence of all landmarks. It was :hence very reassuring to find that the difference° between our positions as shown by the plane-tables at the point of junction amounted only to about half a mile in longitude and less than a mile in latitude.
After completing my successful excavations at and near the Dandan-oilik site I pro
21, 1901. Valuable antiquarian information obtained here led me
northward for five marches through the jungle belt along the bed of the dying Niya river. Beyond it in the area of bare dunes I discovered the widely scattered ruins of an ancient settlement abandoned to the desert sands since the third century A. D. 28 During the very fruitful explorations which kept us busy here for over a fortnight Ram Singh was fully employed on a detailed survey of the extensive site and on reconnaissances into the neighbouring desert. 2° From the termination of the Niya river we then traversed the wholly unsurveyed desert eastwards for a marching distance of over a hundred miles to the site of reported ruins not far from where the Endere river is lost in the sands. S0
Here the easternmost limit of my first expedition was attained, and after exploring with
interesting results the ruins of an ancient fort and other remains, we Return from Eudere. commenced our return journey by February 26. It led us first back to
Keriya along the desert track which since early times has served for caravan traffic along the southern edge of the Taklamakan from Khotau to the Lop-nor region and to westernmost China beyond. Favourable weather conditions allowed the great rampart of the snow-covered K'un-lun range far away to the south to be sighted and in parts to be sketched on the plane-table.
A rapid expedition down the Keriya river for seven long marches from Keriya brought us to a point known as Kara-don', where, near the head of the desert
Surveys b low Keriya, delta of the dying river, the remains of an ancient fort required explo-
ration. 31 Then from a point higher up the river we struck across to the west and surveyed the deceptive desert to the north of the oases of Domoko, Gulakhma and Chira. The ample evidence this ground retains of a much greater extent of the once cultivated areas and of their shifts in position during historical times gives it a special geographical interest. 32
The marches thence to Khotan offered opportunities for surveying similar areas of
early occupation now abandoned to the desert north of the Hanguya canton. Two weeks later excavations carried on at the important ruins of Rawak and surveys of other ancient sites in the desert to the north of the Yurung-kash tract were successfully completed just before the increasing heat and sand-storms closed the season for sustained work on such trying ground.
Our rapid return journey to Kashgar along the great caravan route via Yarkand'
afforded no opportunity for fresh surveys, except from Kizil to Kashgar.
Return to Kishgar. There I parted from Ram Singh who on the whole of this journey
The topographical results of this journey found their first cartographic record in the 'Map of portions of Chinese Turkistan, surveyed under the direction,'
Map of first expedition and with the assistance of, M. A. Stein, Ph.D., by Sub-Surveyor S.—R., surveys.
28 See Sheet No. 19. B. 1, 2.
. .25 Cf. Ruina of Khotan, pp. 334 sqq.; Ancient Khotan, i. Chapter xi ; ii. Plans xxVI1-xxxll.
ae See Sheet No. 19.. C, D. 1; Ruina of Khotan, pp. 389 sqq.
31 See Sheet No. 13. D. 3, 4; Rains of Khotan,
pp. 405 sqq. ..
3^- cf. Ruins of Khotan, pp. 414 sqq.;Ancient Kho. tan, i. Chap. xIII. sec. ii, iii.