National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

Sec. ii]   CHINESE RECORDS OF THE LOP DESERT ROUTE   559

shows, there could in historical times have existed no other route practicable for a large party than the existing caravan track from Tun-huang to Mirân or Abdal. This conclusion agrees both with Fa-hsien's description of the conditions experienced by travellers and with the length that he gives for the journey. H is seventeen days of travel correspond exactly to the seventeen marches in which I covered the ground from Abdal to the Tun-huang oasis, with two days of halt added. In the same way, his estimate of about 1,500 li approaches the distance of about 38o miles that I reckoned for the aggregate of the marches as closely as could possibly be expected, seeing that an equation of 4 or 5 li to the mile is the average which can safely be deduced from the records of Chinese pilgrims to India wherever it is possible to check their estimates of distance between definitely known localities.

Sung Yün, the next Buddhist pilgrim, as far as we know, to visit the Lop tract (A.D. 519), reached it, as we have seen above, not from Tun-huang but across the Koko-nor region and Tsaidam.24 The interesting report of Pei Chu on the Western Countries, compiled in A. D. 607 from information collected at Kan-chou, mentions, indeed, Shan-shan or Lop as the first territory on the southern of the three roads westwards.26 But it does not appear to indicate the route by which this was reached. We can, however, feel quite sure that it was the desert route from Mirân to Tun-huang which was followed in the winter of A.D. 645 by Hsiian-tsang, the greatest of the Chinese Buddhist pilgrims, on his return from India to China. It is to be regretted that his own record in the Hsi yii-chi stops short with his arrival in the territory of Na-fu-po, ` which is the same as the old Lou-lan country ', and, as we have seen above, identical with Lop.2G

Hsiian-tsang evidently considered the ground covered by the remaining portion of his homeward journey as lying within the borders of the Chinese empire, which since he first set out in A. D. 63o for the Western Regions ' had, under the emperor Tai-tsung, vigorously commenced its fresh expansion westward, and therefore as outside the scope of his record. But from his Life, written by his disciples Hui-li and Yen-tsung, we know that Hsiian-tsang actually accomplished this final part of his travels by crossing the desert from Lop to Tun-huang or Sha-chou.27 The letter from the Emperor Tai-tsung which Hsiian-tsang received at Khotan in reply to his application for permission to return, and which is reproduced in the Life, distinctly states that the magistrates of Tun-huang had been instructed to conduct him ` through the desert of shifting sands'.

That the route through the desert connecting Tun-huang with the Lop tract continued to be used during Tang times is highly probable. But the itinerary of the Tang shu from Tun-huang to Khotan, of which M. Chavannes kindly supplied me with a translation,23 starts from the Yang barrier, and may therefore, in view of the explanations given further on,28a be assumed to refer to the route leading to Mirân along the Altin-tagh. I regret that I have not access to the information probably to be gleaned from the newly recovered Chinese geographical texts dealing with the Tun-huang region which are to be found among the manuscripts from the Thousand Buddhas' Caves in my collection, and similarly also in that of M. Pelliot.29

As has been pointed out above, the very existence of a Tibetan fortified post at the site of

Chü's record of southern route.

Hsüantsang's desert crossing.

Desert route to Tunhuang during T'ang times.

2' See above, p. 323.

26 Cf. Richthofen, China, i. p. 530, note, giving extracts from Neumann, Asialische Studien, 1837, pp. 187 sqq.

26 See above, p. 321 ; Julien, ill/moires, ii. p. 247 ; Watters, Yuan Chwang, ii. p. 304.

" Cf. Julien, Vie, pp. 288, 290; Beal, Life, pp. 210, 212.

It would be interesting to ascertain whether the ' conveyances ' for which Hsiian-tsang at Na-fu-po exchanged the horses and camels supplied so far by the King of Khotan, and

which he used for the journey to Tun-huang, were really carts, as the French translation seems to indicate.

28 See Appendix A, II, (Extract from Tang shu, chap. xLIII b).

28a See below, chap. xvi. sec. iv.

29 Cf. for one of these texts Dr. L. Giles's paper,Tun-huang Lu : Notes on the district of Tun-huang, J.R.A.S., 1914, pp. 703 sqq. ; Pelliot, J. Asiat., 1916, janvier-février, pp. I I I sqq.