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

Chap. IV   •NOTES ON SHEET No. 34   91

historical topography is fully discussed in erindia, iii pp. 1147 sqq.

The area comprised in the sheet falls into two climatically well-defined regions, divided by the easternmost extension of the Tien-shan. To the north of it stretch the plateaus and open valleys of Dzungaria where the amount of precipitation received maintains sufficient grazing for nomadic existence and in places facilitates cultivation independent of irrigation. Owing to these conditions Barkul has throughout historical times played the part of a gate for nomadic invasions threatening the extreme northwest of China.

While on the northern slopesof the snowy part of the range, known here as Barkultâgh, conifer forest is abundant (A,B.1, C, D.2), the southern side of the range and the wide glacis of piedmont gravel along its foot are extremely arid. Leaving aside the small patches of cultivation to be found in the narrow and deep-cut valleys of the Karliktagh eastwards (D.2,3), most of the cultiva-

Astronomically observed latitudes.

tion in the fertile but small oases scattered along the foot of the gravel glacis from Toghucha-Lapchuk (B.3) to Hoang-lungkang (D.3) is possible only through the use of subsoil water (kara-sn) coming to light in the usually dry flood-beds.

The latter all terminate in the drainage-less basin of the Shona-nôr (A.3), found in November, 1914, completely dry as far as M. Muhammad Yaküb's survey extended. To the south of the line of oases and the adjacent narrow belt of loess steppe with desert vegetation, the utterly barren Pei-shan 'Gobi' stretches its stony or gravel wastes.

Corrections. A. 3. Omit the latitude observation symbol at Oda-niir.

B. 1 In the environs of Lake Barkul much difficulty was experienced in distinguishing real marsh areas from sandy ground with vegetation under temporary inundation.

D. 3. The name of Ara-tam near ruins should be printed red.

NOTES ON SHEET No. 35. (SU-LO-HO DELTA)

The surveyed part of this sheet shows the ground along the eastern portion of the desert route from the Lop region to Tunhuang, together with the geographically and archœologically interesting area containing the terminal course and marsh basin of the Su-to-ho. The caravan track leading from Tun-huang to the southern shore of the dried-up Lop sea bed was followed both in 1907 and 1914, while the numerous detailed surveys to the north and south of it were also made on the second and third expeditions.

In compilation . the several traverses along the desert track were fitted upon the positions adopted for Mirân and An-hsi. The longitude values of Bésh-toghrak, as derived thus from the'-route from Mirân, and as obtained from the traverse from Altmishbulak agreed within a few minutes. The mean adopted between them, cire. 92' 46' 40', falls short by only two minutes of the longitude shown in Sheet No. 70 of the

1906-08 map ; but account must be taken also of the correction, already referred to under Sheet No. 30, which the reconsideration of the triangulation of 1913 indicates for the position of Mirân. Owing to prolonged explorations along the ancient Chinese Limes, comparatively numerous latitude observations were available for the

- ground near the terminal course of the Su-to-ho.

In view of the special geographical interest which, as stated above, pp. 30 sq., attaches to the well-defined valley stretching down from near Bésh-toghrak to the easternmost bay-like extension of the dried-up Lop sea bed, a line of exact levelling was carried from a point north of Kum-kuduk, marked by C. xcviii (Sheet No. 32. D. 4), to the western edge of the ancient lacustrine basin showing strings of Mesas .and wet sand to the east of Bésh-toghrak (B. 3). The heights shown along this line of levelling against Camps xcviu—cII, as well as the