Sec. v] THE DOMOKO-YAR AND THE REMAINS AT MAZÀR-TOGHRAK 205
it is easy to realize that the additional water-supply, which the springs of the Domoko-yâr brought to light, would not have led to the formation of the new colony of Malak-âlagan but for the incipient pressure of population which the same economical development has caused to be felt within the old oases.6 All these facts are plain enough while there is living evidence to enlighten us. But let us assume that the erection of a dam had surpassed local resources, and the present main oasis had been abandoned accordingly, for its ruins to be discovered by a future archaeologist, say, a thousand years hence. How little could he have hoped then for any definite proof as to the cause which explained the desertion!
Proceeding to the south of the dyke for about a mile along the left bank of the wide reed-covered depression of the Domoko stream, I reached the popular shrine of Mazâr-toghrak, marked by a fine grove of old Toghraks from which it derives its name. About i 50 yards to the west of it, and less than half a mile from the stream, was the spot, the provenance of ` old Khats', about which Mullah Khwâja had heard from Haidul Khwâja, an aged villager, but had never touched himself. It presented itself as a little plateau covered with scanty pottery, measuring about 2 1 o feet from north to south, and 135 feet across, where widest. Its west edge rose about nine feet above the adjoining sandy ground, manifestly wind-eroded, while the east side sloped down four or five feet to the level occupied by the grove of the Mazâr.
When on the morning of October 4 I began to clear the plateau, I soon realized that I was opening an ancient rubbish mound adjoining and partly overlying some completely ruined habitations. From its layers, nowhere more than about three feet in depth and covered only with a foot or so of drift sand, there came the same pungent smells of long-decayed animal refuse, with all its unsavoury associations, which I remembered so well from sites explored on my first journey. The deposits were made up chiefly of decomposed wood, animal bones, and decayed lumps of clay evidently from walls which had crumbled away. It did not take long before some finds of small fragments of wooden tablets inscribed with cursive Brâhmi characters came to my help in approximately dating them.
The largest, M.T. i. 4, measured about five inches in length and showed the wedge-shape familiar from my finds of tablets at the Niya Site, with the string-hole at the pointed end and two lines of faint Brâhmi writing on one side. Small torn pieces of paper documents followed, chiefly from the north-west portion of the mound, and about a dozen in number. They turned up generally less than six inches from the plastered flooring where this could still be distinguished. The paper fragments showed very flimsy texture, and bore writing only on one side. With the exception of two Chinese pieces all the records were in very cursive Brâhmi script, and presumably in the old language of Khotan. Among them was the torn part of a rolled-up document, measuring about a foot in length with a width of two inches. Disappointing as such scraps were instead of the promised great haul of ` Khats', yet their material and writing sufficed to indicate for the site approximately the same date as the Khâdalik ruins.
Among the miscellaneous finds there were pieces of rugs, felts, and coarse fabrics, looking like cotton ; a hemp shoe (M.T. i. 13), resembling in shape and make one found at Dandan-oilik ; a number of clay loom-weights, M.T. 005-007, and other implements used by weavers (M.T. 001, 003 ; i. 002). For all these as also for such finds as a key and part of a lock in wood, various