Sec. iii] ACROSS AN ERODED ANCIENT DELTA 361
Rows of dead Toghraks and tamarisk trunks were repeatedly met with, rising over hard Pottery
gypsum-like banks close to ridges of low dunes. Their general direction seemed to be from west,to fragments
east, but there were signs of winding to south-east and north-east that river-beds were likely to show on such ground. Finds of worked stones of the type already described and of coarse pottery, manifestly of neolithic origin, continued to be plentiful. Among the pottery there was the fragment of a wide-mouthed vase of ill-levigated clay, decorated with three incised bands of the herring-bone pattern (C. 122. 001. a, Plate Iv). From about the sixth mile onwards, I thought I could recognize fragments of pottery of distinctly better make, showing a uniform black surface over a red, hard-fired core 1'a They might well have prepared me for an approach to ground occupied down to a somewhat later period. Yet it seemed a welcome surprise when, about three miles further on, I came upon what at once suggested the appearance of a small ` Tati ' of historical times.
For nearly half a mile the bare eroded soil was strewn with pieces of slag and potsherds, red Bronze ring or black on the surface, which by their finer grain and kiln-made look at once reminded me of the with in-
pottery débris met with about the Niya and other early sites (cf. for specimens C. 122. 004. a, Han coin.
005. a). This impression was soon proved to be right by the large and fairly well-preserved bronze signet ring which Tokhta Akhûn picked up here (91 miles from C. 122) under my eyes (C. 122. 0021). Its flat oval bezel shows in intaglio two long-necked gryphons, one above the other. In shape and design the ring tallied closely with similar finds of the first centuries A. D. which I well remembered having obtained at the Niya Site.'b A fragmentary square-holed Chinese copper coin, uninscribed, but unmistakably of a type associated with the Han period, was found close by, and furnished conclusive evidence that the pottery débris marked a site which must have been permanently occupied during the historical period.
Just before reaching it we had passed through a line of high sand-cones held together by dead occupation tamarisk growth such as are typical of the banks of ancient river-beds, and which I often saw along an-
higher up on the Kuruk-dar â during my surveys of I I The dry river-bed that these tamarisk-
b P Y g Y Y 9 5• Y bed.
indicated was perfectly recognizable, with a width of about 15o yards and rows of Toghraks. on its banks, where we crossed it on February It), 1914, some four miles lower down," to east-southeast. In its vicinity at different points we then picked up glass beads, the fragment of a well-finished bronze ornament, and three Chinese coins belonging to Han types. These finds furnish additional and conclusive archaeological evidence that this old river-course passed here through a belt of ground occupied by settlements of some kind during the early centuries of our era. It is worth noticing that this belt falls within the conjecturally indicated basin to which Dr. Hedin assigns the Lop-nor lake of that very period, and which .he assumed to have been covered with water northward up to the ruined station of Lou-lan.10
As we moved on beyond this ` Tati '-like ground, the cutting wind dropped slightly and for about half an hour light snow fell which limited the outlook. It lay only half an inch deep, and after the next morning's sunshine disappeared altogether except under the corniced edges of Yardang banks facing north-west. Even thus it helped us to economize ice for a couple of days. When we had gone twelve miles, we passed through a long row of dead Toghraks, rising to ten feet or more and clearly marking an ancient water-channel. About a mile and a half beyond, dusk and the fatigue of a march rendered trying both to men and camels obliged us to pitch camp amidst a thin belt of dead tamarisk-cones.
which near the shores of the ancient dried-up Lop sea I traced in my explorations of 1914 ; cf. Third Journey, in Geogr. J., 1916, pp. 126 sqq. ; also above, p. 341.
"a For specimens cf. C. 122. 002. a, 003. a, 007. a—oo9. a. 16 Cf. Ancient Kho/an, i. p. 415, N. 0015, 0016 (Pl. xLlx)
for shape ; ibid. p. 414, N. oo6 (Pl. L) for design of intaglio. For similar rings found at the Niya and L.A. Sites, see below, Pl. xxlx (L.A. 00107).
16 Cf. Hedin, Central Asia, ii. Pl. 40, 59 ; pp. 238, 326, 360, 627, etc.