SECTION I.—TO THE GATE OF THE GREAT WALL'
WITH our start from the ' Myriad Buddhas ' on July 3, 1907, there commenced those explorations across and along the high ranges of the Western Nan-shan which during the next two months kept my attention fixed mainly on geographical work. The results of the extensive and successful surveys then carried out over an area which, with the additions made on the return journey to An-hsi, amounted to close on 24,000 square miles, are presented in Maps Nos. 82-94. Whatever observations of a direct or indirect bearing on the geography of this great region I was able to make will be found detailed in Chapters LXXIV—LXXIX of my Personal Narrative.' Hence I may in the present chapter restrict myself to an account of those localities which offer an antiquarian interest, and to some remarks on those general physical features which have had a distinct bearing on the history of this region and of the tracts immediately adjoining it northward.
The surveys carried out along those portions of the Nan-shan ranges which stretch to the west of the Tu-ta-fan 2 showed me that physical conditions must prevent the valleys which drain them - from having ever possessed in historical times any importance beyond that attaching to the routes which descend through them from the south. The huge triangular trough at the head-waters of the T'a-shih River, though immediately adjoining the main range, is almost throughout a barren waterless waste. The only place in it which appears ever to have been occupied for any length of time is the ruined fort of Shih-pao-ch`êng.s It guards the defile upon which the different routes from across the high range south converge. Though its remains look recent, the position is one which is likely to have already claimed military value in earlier times.
Proceeding eastwards for close on fifty miles over open plateaus of stone or gravel, we find, indeed, the flourishing small oasis of Chang-ma at•an elevation of over 7,000 feet above the defile in which the Su-lo Ho breaks through the continuation of the Richthofen Range towards Yü-mênhsien.4 But the course of the river higher up lies through wholly impassable gorges, while the main range south of Ch`ang-ma, rising to peaks 19,000-20,000 feet high, affords no practicable route and descends with extremely barren slopes devoid of any but the scantiest surface drainage. Thus Ch`ang-ma can never have been a place of any consequence. Notwithstanding the favourable season, difficulties about water were also experienced on our way through the mountains east of the Su-lo Ho. In this connexion it is of interest to note that north of the Shui-ch`ü-kou Pass I found the ruins of two posts, meant to guard the route descending the valley, to be situated at an appreciable distance from the points where water can be found at present.& This fact seemed to me evidence of desiccation since these little stations were erected.
' See Desert Cathay, ii. pp. 262-333.
2 See Map No. 86. D. 2, circ. 97° 50' long., 39°34' lat.
' See Map No. 84. A. I. The ruins consist of a walled enclosure in stone about r 8o feet square, with defensible quarters and a keep in one corner.
' See Map No. 84. D. r ; for views of Chang-ma, Desert Cathay, ii. Figs. 218, 22 1. I am unable at present to verify the exact spelling of the name, which, as pronounced locally, sounded Chong-ma.
6 See Map No. 86. B. I.