This identification necessarily leads us to recognize in the territory of Shê-mi the valleys of Kafiristan, which border on Chitral to the west and south, and I have already above had occasion to show how well this location agrees with the description of Shê-mi given by Sung Yün, who passed through it on his way from Badakhshan to Udyana.26 The further details concerning Chieh or Chieh-shih given by the Tang Annals' notice also accord well. ` The climate is always warm ; one finds there rice, wheat, millet, and beans ; sheep and horses are reared.' This description is remarkably appropriate to the open and fertile part of the main valley which contains the group of large villages collectively known as Chitral, and which must at all times have formed the political centre of the whole territory now called Chitral or Kashkar. Here, at an elevation of about 5,000 feet, both climate and produce closely resemble those of the Kashmir valley.27 The notice places Chieh at 12,000 li from the Chinese capital, and mentions in addition the custom there prevailing of abandoning dying people in the mountains. It also records an embassy sent to the Imperial Court in A.D. 619, with presents of precious girdles and cups in glass and rock crystal.
It only remains to point out that the identification of Chieh-shih, which is established thus by topographical arguments, finds support also on the philological side ; for, as I have shown elsewhere, the Chinese forms of the name can easily be accounted for as attempts to transcribe the local name Kashkar, or an earlier form of it." The application of this term to the territory of Chitral is well attested by Muhammadan sources from a relatively early date, and it is still in current use throughout these regions alongside the name Chitral, which perhaps was properly applicable to the capital only.2° Chieh-shik, as an attempt to represent Kashkar by Chinese characters, has its exact parallels in the varying forms Chieh-cIz`a (Fa-hsien), Ch`i-slza (Chih-mêng), Chia-shik (Tang Annals), Chia-sha (Hsi-yü-chi), by which Chinese authors of successive periods have endeavoured to give a phonetic rendering of the old name of Kashgar, the Turkestan town and oasis.30
It is more than mere chance which has preserved this glimpse of old Chitral history precisely for that period. M. Chavannes has made it clear, in his masterly analysis of the Central-Asian records furnished by the Tang Annals,31 that the first half of the eighth century is a time of exceptional interest in the history of the great Central-Asian basins which extend east and west of the Pamirs. Chinese policy was then carrying on a protracted and vigorous struggle in the "Trim Basin and the adjoining regions against two great powers for the control of the vast Central-Asian territories it had inherited from the Western Turks. Imperial power in those regions had to defend itself at the same time against Arab aggression from the west, and against the constant inroads which the Tibetans, then a nation of considerable offensive strength, were directing both into the basin of the Tarim and into far-away Kan-su. In the course of this protracted struggle, which by the light of the Chinese historical records we can follow in its varying phases with tolerable accuracy, the endeavour of the Tibetans to join hands with the Arabs on the Oxus, and eventually to secure a fresh line of advance into Chinese Turkestan across the Pamirs, clearly defines itself. Formidable as the barriers raised by nature against aggression across the Hindukush and the Pamirs must seem to us, yet in reality the difficulties there encountered are far less than those which any large body of men would have to face in the endeavour to reach the Tarim Basin from the inhabited parts of Tibet, across the high and forbiddingly desolate plateaus
and ranges of the Kun-lun.
R6 See above, p. 9 ; also Ancient Khotan, i. pp. 14 sq.
2' Cf. Biddulph, Hindoo Koosh, p. 6o. Mughul Beg, whose surveys are utilized by Raverty, says in his account of Kâshkâr: ' Barley, wheat, and rice are produced in abundance'; see Raverty, Notes on Afghanistan, p. 153.
28 Cf. Ancient Khotan, i. p. 15 ; Biddulph, Hindoo Koosh, pp. 58 sq.
29 Raverty, Notes on Afghanistan, pp. 152 sqq., 16r.
'° Cf. Ancient Khotan, i. p. 48, where the references and
Chinese characters for these forms have been recorded. $1 See his Tures occid., pp. 290-99.