国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

Chinese
route
through Lop
Sea-bed.

enterprises towards Central Asia, across the forbidding salt-encrusted bed of the ancient Lop Sea,
and the equally desolate waterless wastes around it. How, thanks to lucky archaeological finds
and the scanty notices preserved in the Chinese Annals, I succeeded in tracing this route right
through to its eastern end, near an old terminal basin of the Su-lo-ho in the desert of Tun-huang, has
been recorded in Chapters VIII and IX.

Western-
most Chi-
nese Limes.

I had now arrived on ground which had retained a special fascination for me since the explora-
tions of my second expedition on the westernmost extension of the ancient Chinese Limes. In the
desert to the north-west of the Tun-huang oasis and then again east of it towards An-hsi I traced
portions, previously unexplored, of that protected line of Han times, the remains of its ruined
watch-towers yielding further early Chinese records on wood (Chap. X. sec. i, iii). In the course of
these labours I was able to pay a fresh visit to the 'Caves of the Thousand Buddhas', south-east of
the oasis. Besides other relics I secured there a not inconsiderable addition to the collection of
ancient Chinese manuscripts which I had recovered in 1907 from the great cache of its walled-up
chapel, in the shape of some 570 well-preserved rolls containing texts of the Buddhist Canon
(Chap. X. sec. ii).

Han Limes
explored to
Mao-mei.

For a month from the middle of April, 1914, I was occupied in tracing and exploring the
remains of the ancient Chinese Limes eastwards, first along the Su-lo-ho to its bend below Yü-mên-
hsien, and then through the desert to the Mao-mei oasis at the confluence of the rivers of Su-chou
and Kan-chou (Chap. XI, XII). Chinese records on wood found at ruined watch-posts proved
that this portion of the Limes, which we explored over a distance of more than 250 miles, was
originally constructed about the close of the 2nd century B.C. as a defence against the Huns, and was
garrisoned down to later Han times.

Explora-
tions along
Etsin-gol.

The friendly support of the Tao-t'ai of Su-chou enabled me to push from Mao-mei down the
Etsin-gol into a portion of southernmost Mongolia and to explore remains, dating mainly from the
Tangut and Mongol domination, at and near the site of Khara-khoto first discovered by Colonel
Kozlov. Besides obtaining abundant remains of manuscripts and block prints in the Tangut
language and Tibetan, as well as other archaeological finds, we were able to make interesting
geographical observations regarding the Etsin-gol valley and its delta. At all times it had served
as a highway for nomadic inroads from the north, and Marco Polo's account of it was fully verified
on the spot (Chap. XIII).

Pei-shan
ranges
crossed into
Dzungaria.

Marching south from Khara-khoto through unsurveyed desert hills, we gained the city of
Kan-chou and thence set out for the high ranges of the Nan-shan. There a serious accident
prostrated me in the middle of July ; but R. B. Lāl Singh's energy made it possible to round off,
as I had intended, our topographical work of 1907 in that great mountain region by a survey of the
head-waters of the Kan-chou river (Chap. XIV). We regained Mao-mei at the close of August, and
thence by a month's arduous travel over ground almost wholly unexplored crossed the utterly
barren ranges of the Pei-shan and the eastern extremity of the T'ien-shan (Chap. XV). Our
journey beyond, along the northern foot of the latter, acquainted me with portions of eastern Dzun-
garia which have played an important part in the history of China's Central-Asian expansion.
Finally it brought us to the ancient site of Pei-t'ing and, by an old route across the T'ien-shan
known so far only from T'ang accounts, down into the basin of Turfān (Chap. XVI).

Winter's
work at
Turfān sites.

The great number of ancient remains to be found within or close to the oases of the Turfān
district reflect its importance, as fully attested by its early history (Chap. XVII). Though these
sites had been the scene of very fruitful excavations by several preceding expeditions, there remained
scope for profitable archaeological work and also for a detailed survey of an area of great geographi-
cal interest. During our stay, which lasted through the winter, we were able to make a close