While excavations kept me busy at the site of Kharakhoto, R. B. Lal Singh had carried our survey right down to the terminal basin of the Etsin-gol. There the delta ends in two lake-beds at no great distance from each other, but separated by a gravel plateau. This bifurcation had a special interest for me as it corresponds exactly to the one I had observed where the Su-lo-ho finds its end in the desert west of Tun-huang. The eastern of the two lakes, which for some time past had ceased to receive flood water, was very brackish, while the other, which is reached by the present main branch of the river, held fresh water, though it too is without drainage.
By the middle of June the rapidly increasing summer heat had obliged us to stop work on this trying ground and to turn towards Kan-chou. It was reached by a desert route leading due south of Mao-mei, while our hard-worked camels were sent off for their much-needed summer holiday to the Kongurche hills, north-eastwards on the border of independent Mongol territory. They subsequently rejoined us when, by the last days of August, I returned there from the Nan-shan still crippled by the effects of the riding accident related in the last chapter.
Then, on September 2, 1914, we started from Mao-mei on the journey which was to carry us right across the great desert area occupied by the ranges of the Pei-shan where its width is greatest, in the direction from south-east to northwest. The routes we followed for close on 50o miles had never been surveyed. I knew that only at one point, the crossroads of Ming-shui, could we expect to touch ground the position of which was known relative to routes previously visited by Russian travellers. Wherever it could safely be
IIIF Color HighRes
