National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

Sec. 1]   THE TERMINAL STATION ON THE WALL   633

west from T. vii on the top .of a narrow gravel-covered plateau, and that, when it has attained at T. iv. a the westernmost continuation of this in the shape of an outlying clay ridge, it turns sharply to the south and, after a very short stretch, terminates on the edge of marshy ground.

The explanation why this particular point was chosen for the end of the wall is supplied by a broad geographical fact clearly recognizable in the map (No. 74. B. 3). The Limes has reached here the extreme north-east corner of the great terminal basin of the Su-lo Ho, filled with lakes and, during most part of the year, quite impassable marshes, which extends westwards to about 92° 55' long. (Map No. 7o. D. 4), and which may correctly be described as comprising between 300 and 400 square miles.' Thus the Limes could here rest its flank securely for a distance of about thirty miles, in a straight line south-westwards, upon ground which was impracticable for mounted men. Further south this great natural defence was continued by an utterly barren belt of gravel

`'Sai', and beyond by the area of huge dunes fringing the glacis of the Altin-tagh north of

A nambar.

We shall see how well even here this flank was watched, against any possible turning attack, Ground

by a line of detached posts and signal-stations thrown out as far as the southern edge of the terminal favours

watching of

basin. The defence, moreover, was greatly aided by the fact that this basin, as the map shows, S.W. flank.

extends a succession of long but narrow inlets south-eastwards into the gravel plateau beyond.

The flood-beds to which these inlets clearly owed their original formation must have been as dry throughout historical times as they now are, except on the rare occasions when rain falls on the

barren mountain slopes. But they bring down some subsoil moisture, which rises to the surface

in rare springs by the edge of the basin and, together with the abundant desert vegetation it maintains in those inlets, renders human existence possible. The long narrow ridges of clay

which the plateau projects like fingers towards the terminal basin, and which separate those inlets, rise to considerable height, some of them standing 200 feet or so above the level of the basin. Thus the posts established on them commanded an exceptionally wide outlook,. assuring the further advantage that a small number of them would suffice effectively to watch the flank of the far-flung line of China's westernmost border.

Turning now to the wall of the Limes proper we see clearly from the map that its extreme Extreme western stretch, from T. vii past T. iii to T. iv. a, follows the line of the ridge which separates the w stesternetchr northernmost of the above-mentioned inlets from the one next on the south. In the latter, water of wall line.

could certainly be obtained even now by sinking wells, and reed grazing and fuel are plentiful in it. Starting from the very end of the wall at T. iv. a north-westwards, the narrow bed of the actual

terminal course of the Su-lo Ho is reached at Toghrak-bulak, our Camp 154, after less than three miles, by going first across the extreme north-eastern corner of the great basin and then crossing the narrow tongue-like plateau which here divides it from the Su-lo Ho bed. This plateau is less than half a mile across here and, being only about 50-60 feet high, could be made practicable for carts coming from T. iv. a without difficulty.'

Natural defences on S.W. flank of Limes.

° My journey to Tun-huang in r914 afforded opportunities for further surveys in the western part of this basin. They have shown that the terminal lake actually filled by the Su-lo Ho at that time (third week of March) stretches for about six miles rom north to south approximately in the longitude of 92° 57' to 93° 2'.

The area annually inundated in the spring and summer is vastly greater and probably extends south as far as lat. 40° 9' or thereabout. Eastwards it is likely to cover most of the depression left blank in Map No. 74. A. 3, 4. The conjectural watercourses from the south shown there should be

1374

deleted. The channels formed by the water that rises in the springs along the eastern and south-eastern edges of the terminal basin lose themselves in the marshy area inundated by the Su-lo Ho.

These additions and modifications of our surveys will be duly shown in Sheet No. 35 of the new atlas (r : 500,000) now in preparation.

3 The fact that the Su-lo Ho approaches here so closely to the north-east edge of its terminal basin, whereas at present it empties itself into that basin only after a further course of over twenty-five miles to the west, is of geographical interest.

4M