Sec. iv] HISTORICAL RELATIONS BETWEEN BARKUL AND HAMI 545
timber could be found anywhere to the south of this easternmost portion of the Tien-shan. What
precise locality is meant by Hei-kan-so I am unable to determine. But evidently some thickly
wooded portion of the northern slopes of the range within easy reach of the Barkul pass is
M. Chavannes has already pointed out that the inscription does not record the successful issue Siege-train
of the campaign, which was secured later by the taking of Kao-ch`ang in the eighth month of the constructed
on heights of
same year, but the preparation of the siege-train by which this result was effected.31 The shortness Hei-kan-so.
of the time which elapsed between Chiang Hsing-pên's arrival on the range and the date of the inscription recording his exploit (the 25th day of the sixth month corresponding to July 20, A. D. 670) is explained by the fact that, as noted by his biographer in the Tang shu, the inscription was engraved on a stélé from which an earlier record of Pan Ch`ao's great achievements had first been removed.32 The famous Han general had, as already mentioned, won his first distinction by defeating the Huns of Hu-yen on the shores of Lake Barkul. Owing to the wish of the Tang commander, more than five and a half centuries later, that the memory of his own engineering feat should be promptly perpetuated by an inscription on a stone, we have thus probably lost a record of still greater historical interest.
We may reasonably assume that advantage was taken of the very large forces which the Emperor T`ai-tsung had sent for the conquest of Turfân,33 to undertake on the same occasion an advance north of the range against the territory which in Han times was known as ` Posterior Chü-shih ' and the chief place of which was destined soon to become, under the name of Pei-t`ing, one of the ` four garrisons ' securing Tang supremacy over the ` Western kingdoms '. But the Tang Annals show that this territory submitted after Kao-ch`ang had been taken by the Chinese, and that it was before Kao-ch`ang that the siege-train constructed in the mountains of Shih-lo-man 34 by the army's engineer-in-chief had been effectively put to use.
The conquest of the Turfàn region and the subsequent establishment, during nearly a century Turkish
and a half, of Chinese control over the `Western kingdoms ' did not make any change in the occupation occupying
of the Barkul territory by Turkish tribes. The notice in the Tang Annals on the Sha-t`o • Barkul
shows that this tribe, belonging to the Ch`u yiieh irg branch of the Western Turks, ` lived to the territory.
south of Mount Chin-so 4 and east of the lake Pu-lei' (Lake Barkul) ; there was a great stony
desert there called Sha-(o ; thence they were called T`u-chüek (i. e. Turks) of Sha-t`o.35 From another notice in the same text we learn that in A. D. 653 the territory of the Ch`u-yüeh was divided
into two districts of n Chin-man and Sha-t`o. Chin-man is identical with the seat of the
Pei-t`ing protectorate, which can be located with certainty near the present Jimâsa, west of Guchen,
35 Hei-kan-so P )) (literally meaning ` the place of
black purple colour ') may, perhaps, be the head of the valley towards the Bökunat-dawan (Map No. 34. D. 2), where, judging from the slopes we saw farther north, conifer forest growth is likely to be particularly abundant.
31 See Chavannes, Dix inscriptions, p. 36. For the use made of these engines of war before Kao-ch`ang, cf. Chavannes, Turcs occid., p. 106.
32 See Chavannes, Dix inscriptions, p. 34.
33 Lines engraved on the side faces of the stêlê mention forces of hundred and fifty thousand men under each of two assistant commanders of the Chinese army ; cf. Chavannes, Dix inscriptions, pp. 32 sq. Even taking into account the fact that the numbers mentioned comprised big contingents of Turk and Tölös (Uighur) auxiliary cavalry (cf. Chavannes,
Turcs occid., p. 105), it seems very difficult to explain how such large numbers could possibly be fed along desert routes such as those from An-hsi to Hami and thence to Turfân.
But, no doubt, they comprised all those employed on the lines of communication—and we have other historical examples of wonderful Chinese performances in the matter of `transport and supplies ', as illustrated by the Êrh-shih general's march through the Lop Desert (see above, pp. 341 sq.) and Kao Hsien-chih's march over the Pamirs and Hindu-kush; cf. Serindia, i. pp. 53 sqq. ; Geogr. Journal, 1922, Febr., pp. 109 sqq.
34 For this name as a general designation of the easternmost Tien-shan, cf. Chavannes, Turcs occid., p. 18, note 2 ; also p. 305.
35 See Chavannes, Turcs occid., p. 97.