52o TO KAN-CHOU AND THE CENTRAL NAN-SHAN [Chap. XIV
news of the great European conflagration. But for the telegraph line passing Kan-chou I should have escaped them much longer. During this refreshing stay I was able to make arrangements which permitted the indefatigable Lai Singh to set out again westwards, after the briefest of halts, for fresh surveys in a portion of the Richthofen range that had not previously been mapped. They were to take him up the big valley debouching at Li-yuan above the point where we had seen it in 1907 (Map No. 46. A. 3), and thence along an unknown portion of the northern slopes of the main range to the head-waters of the Hsi-ta-ho descending west of Kao-t`ai (Map No. 43. D. 2, 3). In this way our survey of the whole mountain area drained by the Kan-chou river would be . practically completed.
Journey On August 22nd I myself set out from Kan-chou on the long-planned journey which was
along river to take me across the desert ranges of the Pei-shan back to Turkestan for the work of the autumn below Kan-
chou. and winter. In order to gain Mao-mei, the appointed rendezvous for my several parties, I chose
the route leading along the right bank of the Kan-chou river which had previously been partially mapped by Lâl Singh. Most of the ground traversed by us on the eight long marches to Mao-mei was pleasant. But to me personally the journey proved somewhat exhausting, as the effort of doing it on horseback in which I rather unwisely persisted inflicted a severe strain on my leg. I had further reason to regret that the absence of boat traffic on the Kan-chou river did not allow us to do at least part of this journey by water and with the comfort that this means of progress during the flood might have offered. For the heavy rain experienced during the first few days had turned the track that our carts had to follow, within the belt of rich cultivation extending along the river down to Lo-pa (Map No. 43. D. 2), into an almost continuous stretch of bog. Farther down the brave Chinese mules and ponies harnessed to the carts experienced equal trouble from the ridges of dunes which in places, especially below the villages of Hsi-pa and Chêng-i (Map No. 43. D. 1), stretch from the barren hills eastwards right down to the river bank.
Remains of The crumbling remains of the clay-built mediaeval ` Great Wall ', a poor structure only eight
mediaeval feet thick, were first met with below the village of Ping-yu-pao. There it runs up a steep little spur
Great on the low but in places rugged hill range that flanks the right bank of the river all the way down
Wall '. P gg g g Y
to Mao-mei. It perhaps had its continuation south-eastwards along the top of the hill chain, which watch-towers crown at intervals. Or it may be that the impassable character of the chain was deemed to afford sufficient protection, along the part extending towards the point east of Kan-chou where the late border wall was seen again (Map No. 46. B. 3). Farther down, this wall cropped up again in long stretches, usually along the edge of the present cultivation, as shown by the map ; while elsewhere it had completely disappeared, and only the line of more or less decayed watchtowers remained to mark the direction along which it probably ran. This complete decay over long distances of the mediaeval `Great Wall', which runs on to, and ends at, Chia-yü-kuan beyond Su-chou, helps us the more to appreciate the time-resisting solidity which the methods of construction employed by the engineers of Han times had assured to most parts of their Limes wall on ground offering far greater physical difficulties. On the other hand, it must not be overlooked that the less arid climatic conditions of the valley of the Kan-chou river down to the defile below Chêng-i, where the mediaeval wall comes to its end on the right bank, were not as favourable to its preservation as those of the ground farther west towards Su-chou.3
Temple of Among other points along this route that deserve brief notice I may mention that, a short
Lai-lai- distance east of the village marked on the map as Sun-nai-pao,4 we made a night's halt at a large
and interesting temple. It is known as Lai-lai-miao (?) and is situated near high dunes. It seemed
3 See above, p. 403 ; Serindia, iii. pp. 1121, 1133. doubtful, its Chinese form having been recorded later at
4 The name taken from 1,5.1 Singh's plane-table seems Kan-chou.