508 TO KAN-CHOU AND THE CENTRAL NAN-SHAN [Chap. XIV
Search for The search made there and farther south on that day failed to reveal any sign of the continua-
continua- tion of the Limes wall, such as is likely to have existed on the right bank of the river opposite Lion of
Limes. T. XLVIII. b, its last traced point. The line of the Han wall might possibly have made a big bend
so as to protect the northern end of the oasis, but there was no time for a prolonged investigation of the ground higher up the river as a halt was objected to by the Mongol guides and owners of our transport. The look-out we kept on our marches farther south also failed to reveal any indications of the Limes on the barren hills and plateaus that we passed. Thus the task of tracing its continuation to the east of the Etsin-gol had to be left to some future explorer. Considering the nature of the ground to be searched, he will need all the help that a cool season and adequate camel transport can offer.
Desert Another five marches, covering a total of over a hundred miles, brought us by Midsummer
hills crossed Day to the town of Kao-t`ai by the Kan-chou river. Of the route followed, hitherto unsurveyed,
to Kao t`ai. there is little for the non-geologist traveller to record, apart from the topographical details shown in Maps Nos. 42. D. 4 ; 45. A. 4 ; 46. A. I, 2. It took us over four gradually rising hill ranges ; the ascent of the first was over a bare glacis of gravel, and the next three were separated from each other by wide valleys where scattered patches of hardy scrub and tamarisk growth offered grazing for camels. Except in the southernmost range crossed between Camps 162 and 163, the rocks which compose these hill-chains rarely emerge from the masses of detritus that smother their sides (Fig. 259). Until the traveller reaches the northern slope of the range just mentioned, overlooking the Kan-chou river, water is to be found only in the single well of Ta-shan-kou (Map No. 45. A. 4) and the two scanty and distinctly brackish springs of Ytieh-kung-ch`üan and Zigda-kaya (Maps Nos. 42. D. 4 ; 46. A. 1). A few stunted Eleagnus (Jigda), clinging to the cliffs of granite which flank the latter spring (Fig. 260), were the only trees met with.
It was clear that these wastes of decomposed rock, sand and gravel can receive but the scantiest rain or snow-fall. Such atmospheric moisture as may pass northward from the Nan-shan is probably almost all caught and precipitated on the higher range to the south, known to the Mongols as Köka-ula, which clearly shows the effects of water erosion on its barren slopes. This range, as Map No. 46 shows, is undoubtedly a continuation of the high and rugged chain that overlooks Kan-chou from the north and is itself a portion of the Ala-shan system. Whatever change this ground may have experienced through ` desiccation ' during historical times, it is certain that the route leading across it could never have served, like that along the Etsin-gol, for the advance of a large invading host. But raids by small parties of nomads were easy enough here, and it was to give warning of these that the watch-towers were set up which we found crowning the last rocky knolls before passing through the crumbling line of the mediaeval ` Great Wall '. Immediately beyond we were greeted by the smiling green fields of the village of Lo-pa, which stretch down to the right bank of the Kan-chou river (Map No. 43. D. 2). It flowed at the time in a single deep channel about 8o yards wide, which we crossed by a ferry-boat ; and after a mile's ride I found myself back at the eastern gate of the lively little town of Kao-t`ai, which I well remembered visiting in September, 1907.
In view of the fatigues undergone by men and beasts on the long hot journey, I was here obliged to make a two days' halt in the cool shady quarters which were once again hospitably offered me in a large temple outside the town gate (Fig. 264). My time was occupied in arranging for cart transport by which I might reach Kan-chou quickly by the high road, and also in a brief survey reconnaissance to the south. I ascertained from this that the barren foot of an outlying spur of the Richthofen range approaches here within five miles of the Kan-chou river, thus greatly reducing the width of the cultivable portion of the `thalweg' that forms the great natural highway