Sec. iii] THE RAWAK STÛPA 483
The route from Ak-sipil first led, for a direct distance of about three miles, over high dunes of March to coarse greyish looking sand, which evidently had been deposited from an earlier river-course. The Rawak• latter itself could be recognized in the broad pebble-covered Sai on which we emerged, and which we proceeded to follow in the general direction towards north-north-east. Further to the south it was said to connect with the Sai near Tam-öghil and Kumat, referred to above as an ancient river-bed'. Though water was alleged never to reach it now, tamarisk growth and other scrub marked its course all along and far away to the north, where I subsequently struck it again near the site of J umbe-kum. After covering, in great heat, and glare, a distance of some six miles, we reached a well with some tamarisks, which Jiya people going for wood use as a resting-place. Thence striking to the north-west, we soon were again among heavy riverine dunes. After about two miles we passed a small open depression, where some old pottery appeared on the bare loess ground. Here I was shown by Muhammad Sharif the remains of a large pot of coarse red clay, brought to light from beneath what may have been the mud floor of a completely eroded dwelling. Of timber or other structural débris there was no trace. Muhammad Sharif declared the haul of Chinese copper coins made here to have amounted in value to the equivalent of about 30o Khotan Tangas (say somewhat over Rs. ioo). From among the 87 coins which
I purchased out of this find, and which have all proved to be wu-chu pieces, are taken the three specimens reproduced on Plate LX X X I X, 20-22. Most of them show few marks of wear, and thus are not likely to have been long in circulation before the deposit was made.
At Rawak, which was reached about a mile beyond, an unexpected and most gratifying First view
discovery awaited me. My honest old guide Turdi had spoken only of ` an old house ' to of Rawak
be seen there half buried in the sand ; but in reality the first glimpse showed a large Stûpa, Stûpa.
with its high base and enclosing quadrangle, by far the most imposing structure I had seen among the extant ruins of the Khotan region. Large dunes of sand rising in places to about 25 ft. covered the quadrangle and parts of the massive square base of the Stupa on the south-west and north-east faces. But towards the south the drift-sand was lower, and there great portions of the Stûpa base, as well as of the walls marking the quadrangular enclosure of the Stûpa court, could be readily made out. Fig. 6o and Plate X I I I. a show the ruins as seen on the morning of April i r, from the east and from near the south corner, respectively. Fragments of coloured stucco, evidently from the heads and upper parts of colossal statues, were lying in the sand on the outside of the enclosing wall near the southern corner, showing with some exposed portions of reliefs on the wall behind that ` treasure-seekers ' had been at work here quite recently. I subsequently learned that a party of six men from J iya had decamped on the morning before my arrival. Fortunately they had cleared away only a few feet of sand, and the damage done by them was not extensive.
I realized at once that there was scope here for extensive excavations, and accordingly lost Proximity
no time in sending back an urgent requisition to the Beg of Yurung-kash for a reinforcement of subsoil
of labourers. Fortunately the position of the ruin, within a day's march of the oasis, enabled me to secure in quick succession contingents of willing workers from the nearest villages of the
J iya tract. A favourable factor of still greater practical importance was the relative ease with which the question of water-supply for such a number of men was solved. Though the sand-dunes surrounding us looked more formidable and sterile than at any ancient site previously explored, it was possible to dig a well in a depression within two miles of the Stûpa, and there the labourers' camp was conveniently established. A look at the map shows that the distance from the Rawak site