472 THE RUINS OF AK-SIPIL AND RAWAK [Chap. XIV
within recent times. West of the Hanguya tract we had to cross a broad belt of sand known as Arkalik, which represents an inroad of the desert southwards. There were no regular dunes, and traces of canals and plentiful débris of modern-looking pottery showed that this waste had passed out of cultivation at no very distant period. Isolated plots of new fields in the direction of Yurung-kash furnished welcome proof that here, too, efforts at reconquest had started. At the Hanguya-Ustang we crossed the boundary line dividing the districts of Khotan and Keriya, and riding on towards the town of Yurung-kash I soon observed the manifest signs of old and unbroken cultivation in the high level of the fields above the roads and in the large size of the trees lining the latter.
On April 6 I halted in Yurung-kash, where fresh supplies and labourers had to be secured
and many repairs to be effected in our equipment. Increasing heat by day and recurring dust-storms warned me that the season was close at hand when work in the desert would become impossible. Instead of taking the rest we all by this time felt much in need of, I hastened to set out for the ancient sites which still remained to be explored in the desert north-east of Khotan. So after discharging Ibrahim Akhttn, the worthy Darogha of the Keriya Ya-mên, with a liberal and amply earned reward in glittering gold roubles, I set the caravan in march again on the morning of April 7.
The ruined site known to treasure-seekers as Ak-sipil (` the White Wall '), and situated
among high sand-dunes at a distance of about fifteen miles from the right bank of the Yurung-kâsh opposite Khotan, was my first objective. On the march, however, and close to the edge of the cultivated area, I took occasion to examine the interesting small site generally known as Tam-Öghil, from an adjoining hamlet, where ancient ` culture-strata ', yielding some leaf-gold, besides old coins, terra-cottas, &c., are worked under conditions exactly similar to those described at Yôtkan, but on a far more limited scale. Turning to the east of an isolated and conspicuous sandhill which bears the tomb of Sultan Masan Basri, some two miles in a direct line to the north-east of Yurung-kâsh, I was taken to the loess banks which form the scene of these excavations. They rise steeply above the level of the fields of lucerne adjoining them on the north, and extend in an irregular line for about 23o yards from west to east.
Just as at Yôtkan they were said to have been formed solely by the soil having been
gradually cut off for the sake of washing the gold-containing layer below. This proved to be from 3 to 6 ft. in thickness at different points, while the stratum of fertile earth overlying it varied in height from i o to 18 ft. The riverine loess soil of which this top layer manifestly consists showed here and there distinct traces of stratification. Considering the short distance, less than three miles, which separates this site from the present right bank of the Yurung-kash, and the existence in its vicinity of an older bed marked by the jade pits of Kumat 4, it appeared to me possible that these slight layers, i to i 2 in. thick, might be due to exceptional floods from the river.
From the local cultivators I ascertained that the excavations, the site of which they called
Mazar-oili-Sai (` the sai of the Mazar lands '), had been started some twenty years before my visit when the formation of a small ` Yar' just as at Yôtkan—had revealed the presence of gold in the culture-stratum below the loess soil. The cutting of the banks was said to have begun at a distance of about 150 yards to the north of the present face of the banks, and to have been gradually carried further and further into the rising ground south. The work has been carried on practically only by Tam-öghil villagers, as water for ` washing ' the soil is