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

longitude further west than the latest maps had shown it. ³⁶ As the scale of our plane-table
sheets was not sufficiently large to record all topographical details of interest from an
archæological or geographical point of view, I supplemented them by numerous sketches.
The additional materials thus secured were afterwards embodied in a half-inch map of the
westernmost Limes which in turn served for the 'Detailed Map of the Ancient Chinese
Limes west of Tun-huang' reproduced in Serindia. ³⁷

Subsequently weeks of most fruitful archæological labour were spent by me in ex-
ploring a great hoard of ancient manuscripts and art relics discovered at the 'Caves of the
Thousand Buddhas' south of Tun-huang. During this time no fresh topographical work was
possible as the surveyor's impaired health demanded consideration. But when on June 24th
I left An-hsi, after depositing my archæological spoils, etc., at the district headquarters,
some months became available for geographical work in the western and central Nan-shan.

First an extensive ruined site was surveyed near Ch'iao-tzu between the two outermost
Survey in westernmost hill ranges of the former. Then we turned into the mountains due
Nan-shan. south and mapped the high snowy chain separating the headwaters
of the T'a-shih river from the elevated plateaus of Tsaidam.³⁸ Passing
along the northern slopes of that chain and crossing the Su-lo-ho near the hill oasis of
Ch'ang-ma, we made our way by unsurveyed routes to the famous defile of Chia-yü-kuan
near Su-chou. ³⁹ There we struck the highway which since the earliest historical times has
been the main line of communication between China and Central Asia.

Great efforts were needed to secure needful transport and official help at Su-chou for
explorations southwards in the Central Nan-shan. But by July 28th
Surveys in Central we were able to set out and after crossing the Richthofen Range reach-
Nan shan. ed the high plateau, nearly 13,500 feet above sea-level, separating
the valley of the Hung-shui-pa river from the wide uplands at the headwaters of the
Kan-chou river. ⁴⁰ No guidance was obtainable beyond the small gold-mining camp here
encountered, nor were any humans again sighted for nearly a month. Fortunately the
well-defined character of the four great ranges in which the Central Nan-shan rises towards
the uplands of the Koko-nör and Khara-nör region and the open character of the great
valleys between them facilitated systematic survey work.

By marches aggregating over 400 miles we managed during August to cross and
survey the three northernmost ranges, all rising to snowy peaks of 18,000 feet or more,
between the approximate longitudes of 98° and 100°, together with a portion of the outer
spurs of the Richthofen Range further east towards Kan-chou. In the course of these surveys,
all rivers descending to the oases from Su-chou to Kan-chou, as well as the Su-lo-ho, were
traced to their snow-fed headwaters. Wherever possible we travelled by routes and passes
different from those taken by the Russian explorers, MM. Potanin, Obrucheff and Col.
Kozloff, who had first visited parts of this mountain region.

Excellent stations for the plane-table and for photographic panoramas could be
climbed above the passes, over 15,000 feet in height, by which we
Exploration of crossed the To-lai-shan and Alexander III ranges. ⁴¹ The magnificent
Su-lo-ho sources. glacier-clad range which divides the headwaters of the Su-lo-ho from