Richthofen Range, which in a clear atmosphere are visible from Su-chou, more exactly determined through triangulation than had been possible before. Brilliantly clear weather on May 4th broke the succession of rainy or hazy skies and raised hopes that this might be possible. But though Lal Singh duly measured an astronomically fixed base line on open ground some distance south of the city, the return of more clouds and rain ultimately frustrated this effort.
I have already recorded in Serindia such observations as I had been able to collect on my previous visits to Su-chou regarding the geographical features that account for the economic and commercial importance of Su-chou and for the part the city and district have played in the history of China's relations with Central Asia, ever since Han Wu-ti first established the command of Chiu-ch`üan 1 I may therefore proceed at once to the account of the journey by which we reached our distant goal in the north-east. Apart from the archaeological interest attaching to the ruined site of Khara-khoto, there was a distinct geographical task to be served by this journey.
The drainageless basin in which the Etsin-gol terminates receives all the waters carried down from the high ranges of the Central Nan-shan by the rivers of Su-chou and Kan-chou and their tributaries. In 1907 we had succeeded in surveying approximately the western half of the big mountain area from which those two rivers descend, as well as the greater portion of the geographically important plateau that extends between the foot of the Richthofen Range and the hill chain bounding the deserts of southernmost Mongolia. Well watered by those rivers and occupied by the main oases of the S u-chou and Kan-chou districts, this plateau has been destined to be a true passage-land between China and innermost Asia all through historical times.
My object now was to extend the survey of this great drainageless area, towards both its terminal basin in the north and its head-waters in the mountains to the south-east. It was evident that the first portion of the task would have to be carried out before the heat on the Etsin-gol and in the deserts flanking it became too great for effective work. Then the summer months immediately following would remain available for fresh surveys in the high valleys of the Nan-shan, and also for that grazing holiday which was indispensable for our hardy camels if they were to keep fit for the explorations of the next autumn and winter.
The oasis of Mao-mei where the rivers of Su-chou and Kan-chou meet was to be our immediate objective, and for this we set out on May loth by two different routes. Lâ.l Singh was to proceed eastwards by the high road to where it strikes the Kan-chou river. He was then to descend to Maomei along that portion of the river's course which breaks through the above-mentioned hill chain bounding the northern edge of the plateau, and which had never been surveyed. My own route had necessarily to lie along the Pei-ta-ho and past the outlying oasis of Chin-tea in order to enable me to search for the eastward continuation of the Limes.
I had already followed the main route connecting this oasis with Su-chou along the right bank of the river, when returning in 1907 from my first visit to Chin-tea. So I now chose a different track leading through that portion of Su-chou cultivation which lies north of the Pei-to-ho. Here we passed again the line of the mediaeval ` Great Wall ', completely decayed where we saw it, and then camped by the edge of village lands which stretched beyond it (Map No. 43. B. 1). Next day we crossed the low hill chain that starts from the eastern end of the range overlooking Hua-haitzû and extends towards Kan-chou. Here, too, as above Yeh-mao-wan, we found its crest guarded by advanced watch-towers, apparently of no great antiquity. At the north foot of the hill chain we had to cross an area covered with drift-sand ridges rising to thirty feet in height, before we reached the bed of the Pei-ta-ho about two miles from the southernmost point of the Chin-tea oasis. The bed, over a quarter of a mile wide and excavated to a depth of about six feet, was corn-
1 See Serindia, iii. pp. 1126 sqq.