576 THE TURFAN TERRITORY [Chap. XVII
capital of the ` Anterior tribe of Chü-shih ', held between 433 and 45o by I-1ê In 46o
An-chou himself was attacked in Kao-ch`ang and killed by the Juan-juan, who installed in his place a descendant of Kan Shuang, Han Po-chou ;; , fa A.23 There is evidence that this puppet king and his son reigned till the year 481. The memory of An-chou is celebrated in the Chinese inscription of 469 that Professor Grünwedel acquired from the ruins of a Buddhist shrine of Idikutshahri, dedicated that year to Maitreya ; this inscription has been edited and discussed by Professor Franke in the above-quoted paper. After 481 Turfân passed through a protracted period of disorder, in which the neighbouring Uigur tribe of the Tölös or T`ieh-1ê made its influence strongly felt. But there can be no doubt that in Turfân, as throughout the regions along the Eastern T`ienshan, the paramount power all through the fifth century was wielded by the Juan-juan or Avars, until, about the middle of the sixth century, their empire succumbed to the Tu-chüeh or Turks.24
An interesting notice of the Pei-shih, which M. Chavannes has extracted,25 tells us that from A. D. 507 onwards the throne of Kao-ch`ang was occupied by the Ch`ü, a family of Chinese origin, whose old home was in the vicinity of Lan-chou fu. The founder of the house was Ch`ü
Chia ] , who was followed in succession by his son Ch`ü Chien . and his grandson
Ch`ü Po-ya - ti Ili.'" The mention made that the last prince's grandmother was a daughter of the Kagan of the Tu-chüeh or Turks is only one of a series of indications showing the close relation that subsisted between these rulers of Turfân and their Turkish neighbours to the north.
Yet when Chinese imperial power under the Sui dynasty began again to make itself felt beyond the Kan-su marches, Ch`ü Po-ya, with the Turkish chief of I-wu or Hâmi, was the first to offer tribute, A. D. 608.26 In the following year the Turfân king came in person to do homage at the imperial court and received a Chinese princess in marriage. On his return, A. D. 612, he promulgated a decree ordering his people to adopt Chinese fashions in dress, and in due course received imperial thanks and titles as a reward for this renunciation of barbarian customs. Yet, significantly enough, we are told that Ch`ü Po-ya nevertheless did not dare to break off relations with the T`ieh-lê } .' jj or Tölös, to whom he had become subject after their victory over Ch`u-lo, Kagan of the Western Turks, 27 and who claimed the taxes levied by him on all traders passing through his territory.28 This notice is of interest, because it illustrates the natural dependence of Turfân upon its nomadic and consequently more virile neighbours established on the northern side of the mountains, and the ease with which these could always levy blackmail on the trade passing through the oases on the south. It also foreshadows the conditions that probably prevailed in the fertile settled district of Turfân during later times, when after the close of Tang domination it had passed under the rule of the Uigurs, the most famous tribe of the Tölös.
The interesting and detailed description of the territory of Kao-ch`ang which is furnished by
in the Pei shih composed in the seventh century has been fully translated and discussed by Professor Franke in his above-quoted paper.29 It will therefore suffice to mention here only certain points that have a distinct antiquarian bearing. In that portion which relates to the 4th-5th century of our era, Kao-ch`ang is said to contain eight towns, all of them including Chinese among their inhabitants. Mention is made of the warm climate and the fertility of the soil, which allow cereals to ripen several times in the year. Irrigation of the fields is specially noted ; also sericulture and abundance of fruit and wine. The people are said generally to worship ` the spirit of Heaven '
23 See Franke, loc. cit., pp. 21 sq.
24 Cf. Chavannes, Turcs occid., p. 225.
25 See ibid., p. roz, note 2. For the correct date of Chü Chia's accession, cf. Franke, loc. cit., p. 25, note 5.
25a [But see M. Maspero's reconstruction of the dates,
B.E.F.E.O., x955.—Dr. L. Giles.]
28 Cf. Chavannes, Turcs occid., p. 569, note 8.
27 See ibid., p. 89, note 3.
28 See ibid., p. 503, note.
29 See Franke, Inschrift aus Idikutsahri, pp. 27 sq.