56 FROM YASIN TO KASHGAR [Chap.Il
testimony also of the aridity of the climate in a curious notice of Sung Yiin about Han-p`an-t`o or Sarikol. He specially mentions that its inhabitants had to catch watercourses, i. e. to irrigate, in order to do their sowing, and dilates upon their incredulity when told that in China cultivators relied upon rainfall for this purpose.1°
From the little grazing oasis of Rang we turned up the wide alluvial fan of the Taghash valley in order to visit the old ` Kurghan ' at its mouth, of which I had first heard on my passage along the left river bank in 190611 This ` site ' was found to consist merely of the ruins of some rubble-built huts of uncertain date. But the detour rewarded me in a grand view to the north across the main Sarikol valley and Tagharma as far as the huge snowy dome of Murtagh-ata and its glacier-girt buttresses. From Jurgâl-gumbaz, where our camp stood that evening, the ` Faryâd-ariki ' continued to be traceable, in a clear and almost unbroken line, as we marched on September II down to Tâsh-kurghân. Rashid Beg was also able to point out to me, on the left bank, the head of an ancient canal which once took off from the river opposite Jurgâl-gumbaz and irrigated all the alluvial plateaux on that side right down to Ak-tam. What small patches of ground have now been brought under cultivation again at that point or near the mouth of the Pit and Vanaizraf ` Jilgas ' are all irrigated from side-streams.
Farther down, after a ride of about eleven miles from Jurgâl-gumbaz, we reached the head of the wide bare plain that local tradition knows, under the significant name of Bcizâr-dashi, as the site of a large ancient settlement. The ` Faryâd-ariki' keeps above it along the foot of a well-marked alluvial terrace ; but on moving across the 13525,r-dasht I noticed the traces of what evidently had been a large branch canal, as well as of several distributaries. I observed that the plain is here covered with a layer, three to four inches thick, of small stones, below which is found soft fertile soil, evidently riverine loess. This distinctly recalled the character of the ground surrounding the Mirân site far away near Lop-n(502 which I had noted in 1907. The inference that here too the condition of the surface is due to deflation was soon confirmed by my examination of the structural remains that still survive at Bâzâr-dasht, of which I had already heard in 1900.13
The badly breached fragments of a large and once massive walled enclosure in stamped clay, which I traced at a point almost opposite Ak-tam, as marked in Map No. 3. c. I, betrayed only too plainly the effect of long-continued wind erosion. Of the circumvallation a length of about 190 yards could be made out on the north-west face, in broken sections of varying thickness, and one of about 6o yards to the south-west. Nowhere do the extant walls rise to more than three or four feet above the ground, and in many places they have been worn down so as to be almost flush with it. This advanced state of erosion points clearly to the considerable antiquity of the site ; but no definite chronological indication could be found. As regards the local winds whose effect is so strikingly noticed on this bare plain, I may mention that according to the information received at Tâsh-kurghân they blow generally from the north and north-east and are often prolonged and of considerable force during the spring and winter. The rapid progress of their denuding effect on bare ground may be gathered from the fact that outside the gate of the small defensible post built by the Russians at Tâsh-kurghân I noticed a distinct lowering of the ground level through this cause since its construction about 1903.
A short distance below the north end of the Bâzâr-dasht the bed of the river widens greatly, and it is from here that the canal takes off which at present carries water across the scrub-covered
10 See Chavannes, Voyage de Song Yun, p. 23.
11 Cf. Serindia, i. p. 76. The local name there noted as Chan is more correctly spelt Rang.
12 Cf. ibid., i. p. S37. 13 See Ancient Khotan, i. p. 39.