92 FROM KASHGAR TO THE KHOTAN RIVER [Chap. III
on our desert crossing to the Keriya river end, as to the slowness with which vegetation follows such deltaic changes.5
Journey up Leaving Yalghuz-kum, where to my surprise I had found a small ` terelgha' recently opened
Khotan by a few shepherds of Abaci, we followed the route up the Khotan river-bed that I had previously
taken in 1908. Our fresh survey showed that the bed of the river, up to two miles wide in places, had changed here but little. The distance to the Mazar-tagh was covered in four long marches. These were made rather trying by the bitter cold, as much as 34 degrees Fahr. below freezing-point combined with a cutting wind and a grey dust-laden sky. On nearing Dârun-satma, where we camped on November 16th, the bold river end of the Mazâr-tâgh, though rising only some 220 feet above the wide sandy bed, was already well in view. Three small hills, seemingly detached, which were sighted rising above the big dunes to the south-west, evidently form part of the Mazartagh ridge where it continues into the Taklainakân. One of them was said by Kâsim to be visited by people of Islâmâbâd and Tawakkél for the sake of flints embedded in the rock.
Objects November 17th was devoted to a fresh examination of the ruins on the top of the Mazâr-tagh,
from refuse which had been previously explored in 1908 and are described and illustrated in Serindia .6 I had
of Mazar- not been surprised to learn from Kâsim Akhûn when he rejoined me at Kash gar, that since m
tagh ruins. P g my
former visit ` treasure-seekers ' from his native village of Islâmâbâd had resumed their burrowing around the ruined fort. They had not found the precious things they hoped for, but only small remains of implements, &c., such as I had recovered in plenty from the debris within the small fort and still more from the extensive refuse layers outside it. The specimens of their finds which Kâsim Akhûn had brought me, and which the List below specifies,' left me in no doubt that they were in general of same origin and date as those yielded by my clearings of 1908. They comprised inter alia plenty of wooden slips with Tibetan writing ; small pieces of paper with Tibetan,
BrâhmI, Uigur writing; fragments of a wooden bow and arrow-shafts (M. Tagh. 03, 046-8); pottery stamps (026, 031, 049) ; horn and wooden combs (039-40, 041-2) and key (044) ; miscellaneous wooden implements (021-4, 027, 032) ; string shoes (04-6), &c., as shown by the specimens in Pl. V, VI. But in addition to these relics, which obviously originated, like my far more numerous and exactly similar finds of 1908, from the rubbish left behind by those who garrisoned the fort in Tang times, there were a few small objects, like the votive stucco plaques, showing a seated Buddha (07-10, 028) and miniature clay Stûpa (oso), that pointed to some ruined shrine. And it was just the remains of such a place of Buddhist worship that on my former visit I had failed to trace—to my disappointment at the time, I may add. For the very name Mazar-tagh and the existence on the eastern edge of the hill of what were supposed to be tombs of Muhammadan saints had made me look out for such remains, in view of the continuity of older local worship so constantly proved at such sites.
Refuse In accordance with arrangements made from Marâl-bâshi, Muhammad Bég, the intelligent
layers under local head of Islâmabâd, with a small number of diggers from the same place, had patiently
NW. wall of
fort. awaited my arrival at Mazar-tagh. Going with them over the ruins I soon found that the miscel-
laneous small relics first mentioned had come to light from refuse layers, much less extensive than those I had cleared in 1908, but exactly of the same origin, which Islâmâbâd men had come upon under the north-western walls of the fort and its outer court. Small rock fragments and pebbles carried up by the prevailing winds had covered the slope and previously hidden this rubbish from view.
5 Cf. Desert Cathay, ii. pp. 406 sqq. quently received by me from Badruddin Khan, the Khotan
e See Serindia, iii. pp. 1285 sqq.; Figs. 329-31, 335; Ak-sakâl, who had acquired them from treasure-seekers'.
Pl. 59. [The hoped-for inventory of the Tibetan documents has
7 A few pieces of exactly similar character were subse- not yet been received.]