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

Bësh-toghrak. ²⁸ Extending over a distance of 60 miles it has proved a continuously
descending slope with a total drop of 250 feet from the latter point.
Coupled with other observations, the result of this levelling has confirmed the belief
Connection of Su-lo-ho formed on my passage in 1907 that the waters of the Su-lo-ho at a
and Lop basins. period relatively recent in a geological sense had drained into the
Lop basin. ²⁹ In this connection the fresh surveys effected in the
desert area which lies east of Bësh-toghrak and north of the present terminal basin of the
Su-lo-ho, proved of special geographical interest; they showed that its depressions still
receive subsoil drainage from abandoned branches of the Su-lo-ho delta, and that its mazes
of Mesas are those typical of all lacustrine basins in this region. ³⁰ The importance of the
connection thus traced here between the drainage area of the Târim which has its western
limits on the Pâmirs, and that of the Su-lo-ho which extends as far as the watershed of the
Pacific Ocean, fully 24 degrees of longitude further east, scarcely needs to be emphasized.
Leaving Lâl Singh and Muhammad Yakûb behind for supplementary surveys within
the present terminal basin of the Su-lo-ho, ³¹ and along the river's
Chinese Limes explored course between it and Lake Khara-nôr, I proceeded to the vicinity of
to Tun-huang. the latter along the line of the ancient Chinese Limes first discovered
by me in 1907. From there I completed my detailed exploration of the Tun-huang Limes
on ground stretching eastwards which circumstances in 1907 had obliged me to leave
unsurveyed. ³²
A brief halt was necessary at Tun-huang during the last days of March to allow men
and animals to recover from the trials of our winter campaign. Then we
Survey in westernmost separated once more. While I paid a fresh visit to the famous cave
Nan-shan. temples of the 'Thousand Buddhas,' or Ch'ien-fo-tung, south-east of
Tun-huang, not without archaeological profit, Lâl Singh proceeded to the mountains due south.
Owing to deep snow he was obliged to content himself with surveying the northern slopes of
the westernmost Nan-shan near the debouchure of the river of Tun-huang or Tang-ho, before
re-joining me by the middle of April at An-hsi via Tung-ya-t'u and T'a-shih. ³³ Muhammad
Yakûb was sent north of the Tun-huang oasis by a new route and then mapped the Su-lo-ho
river along a previously unsurveyed portion of its course to An-hsi.
The task I had set myself for the spring was to trace the line of the Chinese Limes
of Han times from Tun-huang as far as possible to the east and to
Ancient Limes traced to explore whatever ruins might have survived along it. I commenced
and beyond An-hsi. this task by skirting across a belt of difficult salt marshes into the
desert north-eastwards of Tun-huang. At a point not far from where our exploration of
1907 ended, I came again upon the ancient border wall and traced it thence through to
An-hsi. ³⁴ From there, accompanied by Lâl Singh, I moved up the right bank of the Su-lo-ho
and found further remains of the Limes wall and its watch-towers opposite the low hills of
Wang-shan-tzu, exactly where our survey of 1907 carried along the left bank of the river
had led me to look for them. ³⁵
The search for the ancient defensive line which at the end of the second century B.C.
had been raised to protect China's great line of communication into
Survey to east of Central Asia from Hun raids was now successfully continued to the
Su-lo-ho bend. sharp southward bend of the Su-lo-ho southward. Here near the small
village of Shih-êrh-t'un we touched the easternmost point at which on my previous expedi-
tion I had been able to trace remains of the Limes line. ³⁶ The more careful survey of the