77o ON THE ANCIENT ROUTE ALONG THE KONCHE-DARYA [Chap. XXI
to fire, which probably had destroyed quarters built against them. Reddened clay and ashes were found elsewhere also within the enclosure, and shallow pits dug in the debris, perhaps by ` treasure-seekers ', disclosed fragments of burnt timber. Large refuse heaps were found along the foot of the western and northern faces of the tower, resting on layers of ashes and on soil reddened by fire. They had obviously been formed through occupation of the interior after the tower had been subjected to a conflagration, the outer walls, no doubt, affording welcome shelter from the bitter winds of the desert. That this occupation belonged to a period much later than that of the original construction and defence of the station was made clear by fragments of a Tang coin, with the legend K`ai yiian, which was discovered near the surface of the refuse in the north-eastern corner of the enclosure. Most of the rubbish-heaps consisted of stable refuse, chips of wood, and reed-straw. Among the miscellaneous small articles found mixed up with these and described in the List, may be mentioned the iron fitting Y. i. 02 (Pl. CX), suggesting a sword sling ; the fragments of a bronze pendant, Y. 1. 012 (Pl. CXI), and bronze buckle, Y. i. 014 ; an iron arrow-head, Y. i. 015 (Pl. CXI), of unusual shape ; part of a wooden fire-stick, Y. i. 03 ; the fragment of a bamboo arrow-shaft, Y. I. o6 ; remains of a string sandal, Y. i. 04, of same type as those found at L.A. and at Limes stations ; miscellaneous fragments of silk and woollen fabrics &c. These relics, together with the coin, show that traffic had moved along this route down to Tang times, if not even later.
The examination of the structural features of this ruin, together with my subsequent observations at the other watch-stations of the route, has led me to the conclusion that the small fort dates from Former Han times. Until Chinese political control had been firmly established along the foot of the Tien-shan, in the second quarter of the first century B. C., by the appointment of a Protector General,2 the newly opened road passing through Lou-lan to the foot of the Tien-shan and the northern oases evidently needed protection against Hun raids. These could easily descend upon it from the side of the Kara-shahr valley, ever an open gate for nomads holding the vast grazing uplands of Yulduz at its head, and equally from the side of Turfân across the western Kuruk-tagh.3 We also know from the Han Annals that in ioi B. c. Chinese military colonies had already for a time been established at Lun-I`ai, the present oasis of Bugur, and in the conterminous territory of Ch`ii-li, which must be located on the Inchike and Yârkand rivers to the south.4 The ancient road marked by these watch-stations was certainly the most convenient line of communication to the former territory, and was also of importance to the colony in the latter ; for from the side of Lou-lan the easiest access to it lay through Wei-li, its immediate neighbour, i. e. the cultivated tract on the upper Konche-daryâ.
On March 23rd we started from this ancient post in a north-westerly direction in order to trace the succession of ruined towers which Dr. Hedin had been shown by his guides on his march from Korla, and which his description of the route briefly mentions. I could no longer feel much doubt as to their antiquity, even though their first discoverer, in view of their ` perishable material ' and fair preservation, was not prepared to ascribe to them an age greater than a few centuries.5 But I was not certain of actually finding them all ; for we lacked a guide familiar with the district, and very careful as I knew Dr. Hedin's compass traverse to have been, Dr. Hassenstein's map embodying its results could not be expected to replace local guidance, as it was only on the comparatively small scale of i : i,000,000 and did not mark the position of the towers. Fortunately
2 See Chavannes, T`oung-pao, 19.07, pp. 153 sq. ; above,
3 Cf. Serindia, iii. p. 1180. Regarding the power inter-
mittently exercised by the Buns in the Turfân basin down to the middle of the first century B. c. and again later, see De
Groot, Hunnen, pp. 205 sqq., and above, ii. pp. S7o sqq.
4 Cf. Serindia, iii. pp. 1236 sq., with regard to the notice of the Chien Han-shu translated by Chavannes, T`oung-pao, 1907, p. 153, note 2 ; Wylie, J. Anthrop. Inst., x. p. 22.
5 See Hedin, Reisen in Z.-A., p. 76.