Sec. iv] HISTORICAL RELATIONS BETWEEN BARKUL AND HAM' 541•
extremity of the Tien-shan and particularly in the big valley of Barkul, before they could secure a safe bridge-head at Hâmi for the shortest and easiest route across the Pei-shan desert from the side of Tun-huang. Thus the opening of the route which to the present day has remained the main line of communication between China and Central Asia became possible only after the Hun tribe holding the Barkul valley had been signally defeated and the Hâmi oasis wrested from their domination. Of the name Hu-yen of Tl7 M. Chavannes has shown that it was borne by one of the great clans of the Hsiung-nu or Huns, which is already mentioned by Ssit-ma Chien as amongst the noblest of the nation.' If we find just this clan established on the easternmost Tien-shan, we may perhaps see herein a sign that the great natural advantages which the valleys of Barkul and Tur-köl offered for nomadic life were then duly appreciated.
That the Chinese victories of A. D. 73-4 did not finally oust this powerful Hun tribe from the region of Barkul is abundantly proved by the records contained in the Later Han Annals of subsequent struggles for l-iâmi. As early as in A. D. 77 the Chinese withdrew from the oasis, which again fell under Hun domination. Not until a great defeat had been inflicted upon the Huns elsewhere was Hâmi reoccupied by Chinese troops in A.D. 90.8 The fruits of the great work accomplished by Pan Chao, who in the same year had assured the submission of the whole of the ` Western Countries ', were completely lost by A. D. 107, when the conquered territories were once again abandoned to the Huns. A renewed attempt in A. D. 119 to hold Hâmi ended with the destruction of the Chinese force sent there, and the Hsiung-nu, from their convenient base on the easternmost Tien-shah, were again free to ravage the westernmost marches of Kan-su.0
A vivid picture of the situation that resulted is conveyed by the memorial addressed in A. D. I 23 by Chang Tang, Governor of Tun-huang, to the Throne, and reproduced by the Hou Han shu.1°
Considering that among the barbarians of the North the Hu yen king is constantly on the move here and there between the lake Pu-lei (Lake Barkul) and the lake Ts`in 4 ; that he imposes his law upon the Western Countries and joins them to ravage and to plunder ', he urges as the best plan the assembling at the Kunlun barrier of a force of over two thousand men drawn from the command of Chiu-ch`üan. ` Then one will attack the Hu-yen king and separate him from what to him is like his root.' This report is of special interest because it definitely indicates the valleys of Barkul and Tur-köl as the chief haunt of the Hun chief whose activity was particularly felt on the Tun-huang border ; for there can, I believe, be little doubt that Dr. Hermann is right in identifying the ` lake Ts`in' with the Tur-köl, the only lake in the area, other than the lake of Barkul, to which reference can reasonably be intended in such a connexion." Considering the comparatively small force proposed for this expedition, it is clear that its objective could not have been at a very great distance from Tun-huang. The plan recommended as the best by Chang Tang was not adopted by the Emperor. But instead, Pan Yung, the son of Pan Chao, was in A. D. 123 sent to establish a Chinese military colony at Luchun in the Turfân basin. Thence he succeeded in A. D. 126 in defeating the ` Hu-yen king of the Hsiung-nu ' and in securing for a time for ` the six kingdoms of Chzi-slzih 4 j ' freedom from Hun inroads.12
There is reason to believe that the territories designated by this term included the Barkul basin. I shall have to refer to others among them in connexion with the ground that we traversed farther west. We may therefore conveniently consider here the brief notice that the account of
7 See Chavannes, Dix inscriptions, p. 18.
8 Cf. Chavannes, T°oung-pao, 1907, p. 158.
0 See ibid., 1907, pp. 160 sq. ; also 1906, p. 218.
10 See ibid., 1907, p. 162 ; also Dix inscriptions, pp. 20 sq.
11 Cf. Hermann, Seidenstrassen, p. 75. The note quoted by M. Chavannes, Dix inscriptions, p. 21, from the corn
mentary on the Hou Han shu connects the lake Ts'in with the country of Ta Ts'in, i.e. the Roman Orient. This attempt to identify the lake with the ` Western ocean ' only shows that the name had become unknown at the time when the commentary was composed.
12 See Chavannes, Dix inscriptions, p. 22.