Sec. ii] TURFAN FROM LATER HAN TO YANG TIMES 581
The final scene of the struggle was close at hand. The Uigurs' attempt at relief in the spring of 790 failed, and the people of Pei-t`ing, tired of Uigur exactions, submitted to the Tibetans, together with. the Sha-t`o tribe, a branch of the Turkish Ch`u-yüeh, who appear, as early as the first Tang advance to Hâmi and Turfân,56 in semi-nomadic occupation of the Guchen region. The Chinese administrator of Pei-t`ing, Yang Hsi-ku, with his force of two thousand men, was obliged to retire to Hsi-chou or Turfân. Towards the close of 790 a fresh effort was made by the Uigurs to retake Pei-t`ing, but led to their signal defeat. Yang Hsi-kou, who had shared the attempt, was preparing to take refuge in Turfân with the few hundred men he had saved. But the Uigurs treacherously detained him and ultimately put him to death to save themselves from further complications. ` After this An-hsi (Kuchâ) was completely isolated and no one knew what became of it. But the district of Hsi-chou (Turfân) continued to hold out bravely in order to remain faithful to the rang.' 57
SECTION III.—TURFAN UNDER THE UIGURS
The complete predominance which the Tibetans appear to have gained in Eastern Turkestan during the early part of the ninth century accounts for the absence of further references to Turfân in the Chinese records for this period. But soon after the middle of that century, Tibetan supremacy in that region and in westernmost Kan-su was broken by the Uigurs, whom Kirghiz attacks and internal dissensions had forced to move from their former seats in Mongolia to the south and south-west .1 The Sung Annals, whose record of the events leading to the foundation of this new Uigur dominion after A. D. 847 is in substantial agreement with the Tang shu's notice on the Hui-ho or Uigurs: distinctly mention Hsi-chou or Turfân as comprised in it, together with Kan-chou and Sha-chou or Tun-huang.2
It was in the western portion of their new territories that the Uigurs were destined during a prolonged period to play a part of historical importance, and one of considerable interest to the student of Central-Asian civilization, literature and ethnology. The power of the Uigur Kagans holding Kan-chou and other parts of the Kan-su marches succumbed by I031 to the Tangut or Hsi-hsia.3 But the Uigurs in the west created a powerful kingdom, which for centuries extended far along the eastern Tien-shan and, even when divided into several principalities, continued as regards the race and traditions of the rulers to maintain its distinctive character well beyond the Mongol period. By the protection it affôrded to the oases under its control this Uigur rule exercised a far-reaching influence upon the cultural destinies of Turkestan. On the one hand it helped to preserve in those oases the cults and literary and artistic traditions derived during the
Pei-t`ing lost to Tibetans.
Tibetan supremacy broken by Uigurs.
Effects of Uigur
Kapisa, probably passed through Turfân on his way to An-hsi (Kuchà) and Su-16 (Kashgar). But the laconic record of his travels gives no details.
On his return journey he appears to have travelled straight from Yen-ch`i or Kara-shahr to Pei-t`ing, probably via Toksun and Urumchi, without stopping in Turfân.
Wu-k`ung distinctly mentions that, for the journey to the capital, the Imperial envoy and other Chinese officers to whom he attached himself had to take the route of the Uigurs, because the way across the ` river of sand ', i. e. the Hami—Kua-chou road (see Serindia, iii. pp. 1144 sqq.), was closed, no doubt by the Tibetans who held the Kan-su marches, if not also Hami. •
56 Cf. Chavannes, Turcs occid., pp. 96 sqq. Several
Sha-t`o chiefs are mentioned as holding appointment as governors of Chin-man or Pei-t`ing in the first half of the eighth century ; cf. ibid., pp. 98 sq.
57 See Chavannes in Anc. Khotan, i. p. 537, quoting the Tzû chih Bung chien's record for 79o.
1 Regarding the history of the Uigurs and the part played by them in Eastern Turkestan and the Kan-su marches after the Tibetan occupation of these territories,,cf. the extracts from the Tang and Liao Annals in Bretschneider, Med. Researches, i. pp. 241 sqq. ; also the lucid analysis of M. Grenard, J. As., 1900, janv.-févr., pp. 59 sqq.
2 Cf. Bretschneider, Med. Researches, i. pp. 243 sq.
3 See Bushell, The Hsi-hsia Dynasty of Tangut, p. 4.