534 ACROSS THE PEI-SHAN TO BARKUL [Chap. XV
to the north-east, clear proof that ` aspiration ' was drawing the prevailing winds down from the cold heights of the ` Snowy Mountains '. Then a mile farther on came a steep drop into a reed-filled bay where a subterranean flow of water was gathering in marshy springs,. and finally we arrived at the straggling little village, almost hidden as we came close to it by the banks of the deep-cut trough in which it is situated. It was a great relief to meet at once with a hearty Turki welcome and to be assured that it was really Bai, a village administered from Hami. To have reached it after nearly four weeks of continuous desert travel without the loss of a single animal was no small satisfaction. Our poor Chinese guide' had almost to the last persisted in believing that the place we were marching down to was but an illusion created by deceptive spirits.
Though Bai is probably the easternmost settled place in Dzungaria everything seemed closely to reflect the conditions of life with which the Tarim basin had made me familiar. To that still distant ground I felt carried back during the pleasant day of halt that we spent here, with my tent sheltered in a small orchard where the fruit trees were still in full leaf. The water irrigating the fields and gardens is all kara-su, derived from springs rising a short distance higher up in the wide drainage bed. Its volume as measured at the village mill was only two cubic feet per second. No flood water from the mountains was said ever to reach the village lands, and from their position in the very bed this statement appeared to be correct. Of the half-dozen places with cultivation to be found north of the Karlik-tagh and belonging to Hami, Atürük, our next stage westwards, which comprised about a hundred households, was described as the largest, Bai itself counting about fifty. But the numerous camps of more or less nomadic Turki families grazing their flocks and herds in the valleys and on the lower slopes of the Karlik-tagh farther west probably represent a very considerable addition to the population of this small but distinctly interesting submontane tract. In spite of the semi-Chinese dress adopted by the men, these ` Taghliks ' or hill people appeared to me to have retained a great deal more of genuine Turki speech and ways than their fellow subjects under the Wang of Hâmi, in the oases south of the range.
About the route from the Kan-su side I did not succeed in eliciting any clear information. Apparently the rare caravans proceeding by this route to Barkul do not touch Bai, but keeping closer to the foo't of the mountains move direct towards Atürük, where grazing, food supplies, &c., are abundant. But how we failed to come upon their track on our descent towards Bai remained puzzling. That people from Bai graze their camels and donkeys during the summer in a portion of the range through which we had made our way was acknowledged. Omar-tagh appeared to be the general name given to it, and the existence of springs at two points known as Chaganburgase-bulak and Kutghoi-bulak, evidently Mongol designations, was mentioned. Of routes passing north-eastwards into the Mongolian ` Gobi ' the people of Bai could, or would, give no information. But of their existence there could be little doubt, and the ruined fort found close to the south-western end of the cultivated ground and near the springs already referred to was obviously meant to guard the approach from that side. From the report and the photographs brought by Afraz-gul, whom I sent for a preliminary inspection, it appeared manifestly Chinese and of no great age.4 Probably it was meant to shelter one of those posts by which the Chinese, during their struggles with the Dzungars under the Emperors K`ang-hsi and Chien-lung, protected their newly won foothold at Hâmi, essential for the conquest of Hsin-chiang.
4 Afraz-gul's sketch shows the fort as a square circumvallation, measuring about 320 feet on each side of the interior and entered through gates in the middle of the north and south faces. Rectangular bastions project at the four corners. The walls of stamped clay are 8 feet thick and of no great height. Parallel to the walls were traceable wall-foundations
of rows of barrack-like small quarters, all badly decayed.
Tâsh-bulak (Map No. 37. A. 3) was another such post, established by the Chinese and still garrisoned recently, for the purpose of guarding the southern issue of the route crossing the Karlik-tagh from the side of Bai and Atürük.