Sec. i] A DESERT ROUTE TOWARDS KAN-CHOU 511
me to make regarding the geographical advantages offered by its position in relation to the natural highway along the northern foot of the Nan-shan, and regarding the antiquity of the site occupied by the present city.6 A renewed visit to the great Buddhist temple known as Ta fo-ssû, with its colossal images of Buddha in Nirvana, which are referred to by Marco Polo,' revealed no remains that could definitely be assigned to an early date, apart from some small stucco images stored on the upper floor of the shrine which by their superior modelling suggested attribution to Sung times. But in view of the tenacity of local worship, there seems no reason to doubt that the shrine of the ` Big Buddhas', together with the adjacent Stûpas (Fig. 268), occupies the same ground as when it was seen by Marco Polo and later travellers from the West.
SECTION II.—TO NAN-KOU-CH`ÊNG AND THE EASTERN HEAD-
WATERS OF THE KAN-CHOU RIVER
Information received years before from my lamented friend Professor L. de L6czy, the distinguished geologist and geographer, who had visited Kan-chou in 1879 as a member of Count Széchenyi's expedition, made me eager to visit the small town of Nan-kou-ch`êng and the cave-shrines of Ma-ti-ssû (Map No. 46. B. 3, 4), before moving into the mountains. It was easy to satisfy this wish as I found that we should have to follow at first the route leading across the mountains towards Hsi-ning, and that it was possible to approach Hung-shui (Map No. 46. B. 4), where our transport animals were to be supplied, by skirting the foot of the Richthofen range to the west of it via Nan-kou-ch`êng. In order to fix the pcsition of the spot where the Kan-chou river debouches from the mountains I sent Lai Singh on July 6th to Nan-kou-ch`êng by a more devious route leading first to the south-west, while I myself intended to reach the place by moving due south. This intention was, however, frustrated as the carts with our baggage were, by mistake, taken on ahead by the road leading straight south-eastwards to Hung-shui.
This mistake, if it was one, was rectified after we had reached our first halting-place at Camp clxviii, in the midst of the wide gently sloping steppe which on this side divides the Kan-chou oasis from the belt of submontane cultivation. The march to the latter allowed me to realize even better than might otherwise have been possible that we had now reached a dividing line of distinct geographical interest. However flourishing cultivation might appear within the big Kan-chou oasis and however fine the tree-growth along its roads, it was everywhere dependent on irrigation. We had found the same in all the cultivated tracts visited, both on this occasion and previously in 1907, between Kan-chou and Su-chou, whether in the plain or along the foot of the mountains. A very striking difference now forced itself upon my attention as we ascended the steadily rising ground towards the south with the picturesque foot-hills of the Nan-shan coming ever nearer. After about eight miles' march from the lonely rest-house where our Camp clxviii had stood, the steppe of sand or fine gravel gave place to a belt of fertile loess soil, where abandoned fields were covered with abundant grass and flowers. When, three miles farther on, the edge of present cultivation was reached at the quaint walled village of Nan-kou-tai-tzû (Fig. 266), it was seen that its terraced fields were being cultivated without the aid of canals. Indeed the ground beyond was much cut up by dry ravines and flood-beds, which would have made the construction and maintenance of irrigation channels very difficult.
It was the same wherever the winding cart-track took us. Whether we passed fields covered with the young green of growing crops or terraces abandoned to a luxuriant growth of grass and wild flowers strangely recalling a European country-side, everywhere it was obvious that rain and
e Cf. Strindia, iii. p. 1X31 sq. 7 See Yule, Marco Polo, i. pp. 219, 221. -