TO KAN-CHOU AND THE CENTRAL NAN-SHAN
SECTION I.-A DESERT ROUTE TOWARDS KAN-CHOU
ON June 8th we were glad to be able to start from our camp below Dzusulun-tsakha on our journey south towards Kan-chou and the cool heights of the Nan-shan. The hiring of transport to take us as far as the Chinese high road from Su-chou to Kan-chou had been a lengthy and troublesome matter, in spite of the friendly attitude of the ` Beili ' of the Torguts, a weak if well-meaning person ; for no one would agree to work camels in the heat of the season. The Mongols were unwilling to supply even the minimum number of ponies and donkeys required to carry our limited baggage and supplies, except on condition that we should march at night only, and this in addition to exorbitant rates of hire. Such an arrangement, proposed rather truculently as a conditio sine qua non, would have rendered a journey by the reported route leading first south-east through sandy desert to Ghornan-gol and thence straight south to Kan-chou practically useless for survey work. I was moreover anxious to examine the ground east of the Etsin-gol opposite the point up to which we had previously succeeded in tracing the line of the Han Limes. So finally a compromise was arrived at by which we were to move up the Etsin-gol by night marches as far as the point just named and then proceed over previously unsurveyed ground to Kao-t`ai. As this ground was said to be hilly and cooler than the Etsin-gol valley our Mongols agreed to take us over it during daylight, marching in the early mornings and evenings.
The first part of this programme was accomplished between the 8th and 15th of June by a succession of night marches rendered trying by the constant straying of pack animals, loss of loads, &c. Nor shall I easily forget the discomfort from the intense heat and frequent sand-storms that attended our weary halts in day-time (Fig. 237). The last of these marches had brought us past the ruined posts of Arun-tokhai and Tara-lingin previously described' to a narrow strip of vegetation, known as Atik-tsagan, on the right bank of the river (Map No. 45. A. 4). From there it was possible to see the last trees of the Mao-mei oasis on the same bank and also the tower T. XLVIII. b, which marks the point where the line of Limes abuts on the left bank of the river (Map No. 42. D. 4).
A succession of five ruined watch-towers was found to stretch from near this spot north-eastwards to the vicinity of the small fort of Ulan-dürüljin. They occupy rising ground on the bare Sai of gravel which commands a complete view of the valley, and their position suggested that they formed a line of advanced posts intended to guard the approach to the Limes from the right bank. On visiting the southernmost of these towers, T. XLVIII. g, from Camp 159 I found its remains badly decayed ; they were built of bricks measuring 14" x 8" x 6" with layers of reeds separating every three courses. The size of the bricks and the reed layers are constructive details corresponding to those observed in the Limes towers near Mao-mei and support the above suggestion. Such scanty broken pottery as was found near the decayed tower was of greyish colour, but offered no definite indication of date.