National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

October 9th I was free to set out for my first winter's work in the desert. The region
around the dried-up ancient Lop sea was its main goal, and for various reasons that eastern-
most corner of the Tārim basin had to be reached by me via Khotan and beyond it by the
already familiar route skirting the southern edge of the Taklamakān. Opportunities for
topographical work on new ground were thus confined to the journey from Kāshgar to
Khotan, and the time available for it was limited.
I first moved due east to the oasis of Marāl-bāshi by an unsurveyed route which local
Desert route to tradition vaguely remembered as having been in use for traffic during
Marāl-bāshi. earlier periods of Chinese domination, instead of following the present
high 'road' along the lower Kāshgar river. The route led beyond the
outlying oases of Āstin-ārtush and Kalta-yailak closely along the foot of the steep and
barren hill chain which forms here the southernmost rampart of the T'ien-shan. ⁴ The fact
that most of the desert glacis of this hill chain is now wholly without water added to the
geographical interest of the series of small ruined sites and dry river beds we succeeded in
tracing near the old route. Considerable changes within historical times in the course of the
terminal Kāshgar-daryā were indicated also by the survey made on a reconnaissance which
took me from Marāl-bāshi to the detached hills of the Bél-tāgh and Lāl-tāgh in the desert
north-eastwards. ⁵
Our surveys of 1908 seemed to justify the belief that the Mazār-tāgh hill chain
Search for eroded traced then for about 20 miles from the west bank of the Khotan river
desert range. bed in the Taklamakān was in geological structure but a remnant of
an ancient range starting at an angle from the outermost T'ien-shan
near Marāl-bāshi and once extending in a south-easterly direction across the Taklamakān. ⁶
That the bold island-like hills which rise from the desert plain to the east and north-east of
Marāl-bāshi and which mark the north-western end of that assumed ancient hill range, have
been carved out and isolated by the action of wind-driven sand, a most potent physical
factor throughout the Tārim basin, was clearly proved by the observations made on the
above reconnaissance. The same process prolonged through geological ages would obviously
account for much bigger breaks in the continuity of the range in the great drift-sand desert
further south. To test this hypothesis on the ground by a 'short cut' through the Takla-
makān in the direction of the Mazār-tāgh on the Khotan river, I set out on October 25th
after careful preparations at Marāl-bāshi.
Crossing the Yārkand river we reached by three marches the end of the last of those
Progress stopped by sand-scoured hills, known as Chok-tāgh. ⁷ From a lake near it Dr.
high sand ridges. Hedin had started in May 1896 on that bold journey eastward which
ended with the destruction of his caravan and his own narrow escape.
Having taken water supply there we forced our way for three more trying marches into the
sea of bare dunes. ⁸ The ridges or 'Dawāns' into which they were heaped grew steadily higher
and rose invariably in a line diagonal to our intended direction. This bearing and the
almost total absence of level ground between the endless succession of 'Dawāns' made progress
very slow and exhausting with heavily laden camels. By the evening of the third day the
animals hired to form a 'supporting party' for our own had broken down. Assuming that
our rate of progress could be maintained, there still remained some eleven marches to the
nearest point of the Mazār-tāgh hill chain previously sighted. The same forbidding expanse
of huge sand ridges spread before us on the fourth day, and I realized from previous experience
the difficulty of steering a correct course or even of recognizing low eroded offshoots of that
hill chain from a distance.
These considerations together with others concerning the work ahead forced me to
Observations in turn back, however reluctantly. Compensation was afforded by two
dune-covered area. interesting discoveries which had already attended the effort. On the
third march between the high dunes we had again and again come upon
patches covered with minute but easily recognizable fragments of slatey rock flakes, the