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

Sec. iii]   RETURN FROM THE NAN-SHAN TO MAO-MEI   5 2 I

fairly old and the sand was invading all of its numerous courts and shrines. With its elaborate and graceful wooden pavilions, balustraded passages, and bridges connecting different shrines it recalled more closely the architecture of celestial mansions as represented in the Chien-fo-tung paintings of the Western Paradise than any of the structures seen by me on Chinese ground. Its remains when completely abandoned to the drift-sand may provide an interesting ` site ' for some future excavator a thousand years or more hence. It is evidently a cherished place of pilgrimage for the inhabitants of the flourishing walled villages by the river, and I regret not to have been able to secure an account of its origin and tradition.

Below Lo-pa the line of continuous cultivation with its picturesque villages (Figs. 281, 282) comes to an end. Cultivation farther down is reduced to a succession of narrow patches between the high sands covering the foot of the barren low hills eastwards and the wide river-bed. A somewhat wider belt of cultivation was again reached at the large walled village of Hsiang-p`u, where to my great satisfaction I found Lai Singh safely arrived after crossing the river. He had completely carried out his final task in the mountains south of Li-yuan and Kao-t`ai, having received a very friendly welcome at the camps of the Shera and Sarö Yögurs, small tribes of Mongol and Turkish-speaking nomads, who graze in those mountains! By this additional effort my untiring companion had succeeded in extending the Nan-shan surveys carried out during this journey over an area quite as large as that mapped there in 1907.

About ten miles below Hsiang-p`u lies the small oasis of Chêng-i, where bare rocky ridges approach the river valley on both sides and gradually narrow it to a gorge. A military post still maintained in the little walled town is a relic of the time when the passage through the mediaeval wall, rising immediately above the point where the river enters the gorge between steep cliffs about 30o feet high, was guarded as a true kuan or barrier. This character of the locality is marked also by small shrines occupying terraces on either side of the river. The defile (Fig. 277) in which the river has here cut its way through the barren hill chain extending all the way from above Kan-chou to beyond Su-chou is quite impassable when the river is in flood, and probably impracticable for traffic at other seasons also. The road towards Mao-mei crosses the bare rocky ridge to the east by a pass, about 400 feet higher up and away from the river, and then descends again towards the latter in a narrow tortuous valley. Where the river debouches from the defile and the spur to the east of it flattens out into a gravel fan, a huge ridge of dunes, probably 30o feet in height, stretches down from the north to the right bank. The road crosses it over two steep saddles of drift-sand, and here the hauling up of our carts was accomplished only with great labour and delayed us for hours. The whole route from below Chêng-i offers a succession of strong defensive positions, and rows of watch-towers both on and above the pass show that it was duly guarded. From Ta-tzüwan, the first hamlet encountered on the north side, a long day's march, mostly over bare gravel or steppe and past straggling strips of cultivation by the river, finally brought us back on August 29th to our old temple quarters by Mao-mei. There to my relief I found the second surveyor safely returned with our camels from their summer grazing. In spite of the heat that they had experienced in the low hills beyond the terminal basin of the Etsin-gol, the camels were fit to face the hard work of the long desert journey before us.

Lâl Singh's survey south of Kao-ttai.

Passage through ` Great Wall ' at Chêng-i.

5 Regarding these interesting remnants of tribes, probably of Turk origin, transplanted from ` outside the wall ' by the Chinese administration of the seventeenth century, cf. the

remarks in Serindia, iii. p. 1129, based on an account by General C. G. P. Mannerheim, in the journal de la Société Finno-Ougrienne, xxvii (191 I).

3

X