458 KARA-DONG AND THE SEARCH FOR HSOAN-TSANG'S PI-MO [Chap. XIII
acknowledge ever having visited the ' kbne-shahr' near it. Our wanderings of the previous two days had left me no illusions as to the value of our two worthy guides. Yet as better were not to be got and time was getting more than ever precious, I decided to set out with them on the morning of March 25. Old Turdi had never visited these sites, and from the ' treasure-seeker's ' point of view, did not rate them highly. They were too near, he thought, to the cultivated area to have retained much of either ' treasure ' or antiques. Yet he was intelligent enough to realize my interest in them, and I could rely on his instinctive experience of the desert coming to the help of whatever local knowledge the Domoko guides might prove to possess. In order to be prepared for the excavation of any structural remains that might possibly be met, I took along an adequate number of labourers. The heat of the daytime was already becoming troublesome in the desert. But by the precaution of having all the water-tanks filled before the start, I provided against immediate risk resulting from prolonged wanderings. The three days' Odyssey which followed, and which could be traced in detail only on a large scale map, fully bore out my misgivings. Yet there were instructive experiences and interesting antiquarian observations to compensate for the fatigue.
The first two miles in a north-westerly direction brought us to the limit of the newly-irrigated land, when to my surprise I came upon unmistakable marks of earlier cultivation beyond. Old fields overgrown with tamarisk and thorny scrub could be clearly distinguished by the little embankments dividing them, as well as by the lines of dry ' Ariks ' that once carried water to them. There was but little of drift-sand, and that mainly where rudimentary tamarisk thickets were forming to catch and retain it—the first beginnings, it seemed to me, of the tamarisk-covered sand-cones I have had so often to refer to. The Domoko labourers explained that these were the fields of ' old Ponak ' village, which had been abandoned ' in their
grandfathers' time ', i.e. forty or fifty years before. Passing along the road still frequented
by the people visiting the cemeteries of the deserted villages, I arrived, some three miles further north-west, at a point known as Kobza-giram-shah 1.
Here we had reached the southern edge of ' old Domoko ', an area covered with the remains of a deserted village group. The crumbling ruins of mud-built dwellings, constructed and arranged exactly as in the now inhabited villages of this tract and forming detached groups, seemed to extend, together with the interspersed orchards, cemeteries and fields, for about three miles from east to west. Going towards the north-west we kept between them for nearly three miles. The mud-walls, strengthened by the insertion of vertical bundles of Kumush, still rose often 4 to 5 ft. above the ground, and the massive fireplaces were intact even to a greater height, as seen in the photograph (Fig. 55) of one of these ruined houses. The deserted homesteads had been stripped of all materials that could be of use, such as beams, wooden doorposts, &c. As scarcely any sand had accumulated about the crumbling ruins, their rapid and complete disappearance seemed inevitable as soon as erosion set in. Of the latter no trace, however, appeared as yet over this area.
It may be due to this preservation of the surface layer of loess that so little of pottery débris could be seen on the ground. Erosion would probably bring to light plenty of it from different layers. Of other characteristic Tati remains, such as smelting refuse, pieces of charcoal and broken stones from the river-beds, there was a great deal to be seen on the ground. Of the fruit-trees in the orchards only low stumps remained, these trees having evidently died off very soon after irrigation ceased and having then been cut down for timber or fuel. On the