728 EXPLORATIONS IN THE KURUK-TAGH [Chap. XX
plateau above it. It was not until we had left the pass some six miles behind us that the first scanty grazing was reached near the shepherd station known as Kavûta-bâshi-öghil, at the mouth of a valley descending straight from Hsi-ta-shan. Here the valley of Kavûta, from which it takes its name, makes a sharp turn to the south-west and widens greatly. Across its open bottom an impressive view opened towards a jumble of rugged peaks to the west, all rising in island-like isolation above broad detritus slopes. Their fantastic shapes and needle-like pinnacles bore witness to excessive water erosion, whereas in the valley that we were following water was now nowhere to be looked for. Towards the end of the day's march the valley bottom contracted to a defile hemmed in by high and very precipitous rock walls. Apparently the drainage from the Hsi-ta-shan had here cut through a somewhat lower chain, culminating farther east in the peaks of Mohur-shan and Dunda-shan.
Next morning with an icy wind blowing from the west, a very unusual direction in these parts, we continued to descend the valley and found that after a couple of miles it debouched on a wide gravel Sai. This alluvial fan, covered with a good deal of scrub, was seen lower down to unite with one equally wide descending from the north-east. Here we met the well-marked track leading from Singer to Shindi. The drainage that we were following had cut a broad gap through a low chain striking east to west ; through this we passed and sighted the tiny oasis of Shindi about three miles lower down, with the broad mass of the Shindi-tâgh rising behind it to the south. As appears from the map (No. 25. D. 2), Shindi lies at the point where the flood-bed from the Kavûta valley unites with another draining a series of valleys to the north-west in the direction of Lai Singh's Elisen-daw5.n. The latter bed contains a small and lively stream fed by several springs rising between two and three miles above the point of junction. The stream receives some additional water from a spring rising in a luxuriant grove of wild poplars to the north-east of that point, and then enters a narrow defile to the south through which it makes its way towards the Ying-p`an site. The volume of water, as measured by me two miles below the junction of the two flood-beds, amounted to over 14 cubic feet per second.
The means for irrigation that this offered appeared to have been used at one time by Mongols in the intermittent fashion which is customary with those nomads. They were said to have called the place Khulastu.4 But such fields as they had laid out had been long ago overrun again by the scrub and tree-growth of the jungle. It was only four years before my visit that Abdurrahmân, the eldest of the Singer brothers, whom we met here, had begun to reclaim irrigable land close to the junction. The area actually under cultivation was estimated by him at about twenty Chinese acres (mu), while the extent of available land within and above the gorge of Shindi was at least eight times as much, the water-supply being amply sufficient for it. Abdurrahmân accounted for the slow progress of his reclamation by the inadequacy of labour, the few men from Tikenlik whose help as métayers he had been able to secure being shifty folk. He himself as a hunter did not much relish settled life at a spot so isolated as Shindi, and it was only a year before that he had brought his family to it and built there a homestead, partly of rush walls after the Loplik manner.
I lost no time after my arrival in visiting, under his guidance, the remains of which I had previously heard. They were found to consist in the first place of the scanty ruins of two small structures ; these had once occupied the crest of a steep hillock which rises close to the east of the junction of the flood-beds to a height of about i 20 feet above the stream. It is about 8o yards long at its foot. A small knoll at the southern end of the crest shows traces of having once been occupied
4 This was interpreted to me by Abdurrahman to mean `gorge'. Similarly the name Shindi was assumed by him to represent the Chinese Shin-ti, literally ` deep [lying] land ',