his family in the retention of his possessions.l7 On the division of Chinghiz Khan's huge empire between his sons, Bésh-balik with Turfan passed under the suzerainty of Chagatai's branch. It was then that information about the Uigurs first reached Europe. To Friar John de Rubruck, who in 1253-5 visited the Great Khan Mangu near Kara-korum on a mission, we owe a clear indication of the mixture of creeds prevailing among the Uigurs. Though he mentions them as the first among the idolaters, i. e. Buddhists, eastwards in Asia, he also notes Nestorians and Muhammadans as being mixed with them.18 He likewise justly recognizes the important part played by them in the literary use of Turkish speech, a part abundantly demonstrated by the manuscript remains in Uigur yielded by the ruins of Turfan.1°
Whether Mongol supremacy, with its religious tolerance and the easy and constant intercourse it assured between China and Central Asia, had something to do with retarding the spread of Muhammadanism in Uigur territories cannot be stated definitely. But certain it is that Buddhism, and Taoism also, survived longer there than elsewhere in Eastern Turkestan among populations of Turkish speech. The narrative of Chang Chun, the Taoist sage sent for by Chingiz, mentions visits that he received from Buddhist and Taoist priests on his passage in 1221 through Bésh-batik and at some town towards Manas. But he significantly adds that west of that town there were `neither Buddhists nor Taoists'. `The Hui-ho (Uigurs) only worship the west (i. e. turn towards Mecca). 20
The chiefs of Turfan and Bésh-balik and their emissaries whom the Ming Annals mention from the last quarter of the fourteenth century onwards all bear Muhammadan names.21 Yet in 1408 we are told of a Buddhist from Turfan, with his disciples, reaching the Chinese capita1.22 A notice in the Ming Annals, dating from the first half of the fifteenth century, particularly mentions, concerning Huo-chou k Al, by which name Kao-ch`ang (Kara-khôja) was known since Mongol times, that ` there are more Buddhist temples than dwelling-houses of the people'. We can scarcely be wrong in assuming that in this statement are included ` the ruins of an ancient city, the remains of the capital of ancient Kao-ch`ang ', i. e. the ruins of the site now known as Idikut-shahri or Dâkiânûs-shahri, which the same notice, immediately after the passage quoted, describes as situated to the east.23 That Buddhism in 142o was still the prevalent cult in the Turfan tract is conclusively shown by the record of Shah Rukh's embassy, which states : ` They found that in that country most of the inhabitants were polytheists (i. e. Buddhists), and had large idol-houses, in the halls whereof they kept a tall idol.' 24 At Kumul (H ami), too, the same record notes a fine Buddhist temple rising by the side of a mosque.
The fact thus established that complete conversion to Islam took place in the Turfan area so much later than in the Tarim basin calls for special notice here, because, from an archaeological standpoint, it has had a very important influence upon the survival in the former of remains of antiquity and upon the condition in which they are found. It allowed relics of pre-Muhammadan civilization, including objects of cult, literature and art, to exist in this territory, comparatively well cared for, to within four or five centuries of our own time and that on ground which has been continuously occupied. The same fact explains why a large proportion of those remains belongs
17 See ibid., i. pp. 249 sq., 26o sq.
18 Cf. ibid., i. pp. 262 sq. ; Rockhill, The Journey of William of Rubruck, p. 141.
19 ` Apud Iugures est fons et radix idiomatis Turci et Comanici ' ; cf. Rockhill, loc. cit., p. 152.
Friar J. Plano Carpini, travelling in 1245-6 to Karakorum, had already duly noted that the script of the Mongols was adopted from the Uigurs. Had knowledge that this Uigur
script was itself derived from the Estrangelo of the Syrian Christians something to do with his erroneous attribution of the Uigurs in general to the Nestorian sect ? Cf. Rockhill, loc. cit., pp. 147, 150.
20 Cf. Bretschneider, Med. Researches, i. pp. 65, 67 sq.
21 See ibid., ii. pp. 193 sqq., 235 sqq.
22 Cf. ibid., ii. p. 144. 23 See ibid., ii. p. 187.
24 Cf. Yule-Cordier, Cathay, i. pp. 272 sq.