Sec. ii] SITE OF PEI-TING AND POSTERIOR COURT OF CHU-SHIH 555
crossing the pass of Pa-no-pea, was evidently identical with a route described in an itinerary of the rang Annals, translated by M. Chavannes, as leading from the old capital of Turfân Chiao-ho x p, the present Yâr-khoto, to Pei-ling It g. This place, the seat of an important Chinese
Protectorate of Tang times, corresponding to Chin-man 4. of the Han period,1 had been
correctly located by a Chinese antiquarian scholar, Hsü Sung, the author of the Hsi yii shui lao chi, at a ruined site to the north of the town of Jimasa, situated one march to the south-west of Guchen. It therefore seemed easy to combine a visit to the remains of Pei-ling with the subsequent journey due south by the Pa-no-pea pass.
But Li Ta-lao-yeh had not failed to warn me that the Pa-no-pea pass was liable to be closed for a time by heavy snow-fall in the autumn, and moreover the state of my leg still made riding rather trying and walking for any but short distances very difficult. It was therefore wise to avoid the risk entailed by a long postponement of the crossing. As the route is practicable only for mules, ponies, or donkeys, and in places, as I found later, difficult even for these when laden, I decided to send our camels with all dispensable baggage to Turfân across the Ku-ch`üan pass (Map No. 31. B. I, 2). This easterly route, leading via San-ko-chrüan towards Pichan, is the nearest in that direction by which laden camels can be taken. Lâl Singh was put in charge of the convoy and thus secured an opportunity for surveying that portion of the range which the persistent dust-haze, following the blizzard of October 12th-13th, had rendered completely invisible during the days of our journey to Guchen.
On October Igth I started with Afrâz-gul from Guchen by the high road leading to Urumchi. A march of thirteen miles, through a grassy steppe crossed by a number of streams and dotted in places with village lands, brought us to the eastern edge of a large tract of continuous cultivation belonging to the sub-prefecture of Fu yfian.2 From the small and modest town which contains its head-quarters and is generally known by its non-Chinese name of Jimasa, I visited on the following day the ruined site where the ancient Pei-ling has been rightly located. Before I record the observations made at the site, which, as far as I know, has so far not been described by any European archaeologist, I may here briefly indicate the reasons in favour of this important identification, as extracted by M. Chavannes from Hsü Sung's geographical treatise published in 1823.3
The identification primarily rests upon a Chinese inscription of the 'rang period actually found at the site, which definitely proved that it was once occupied by the sub-prefecture of Chin-man / `► 4.4 A passage of the Chiu Tang shu, composed in the second quarter of the tenth century,
states that `Chin-man in Later Han times was the seat of the Posterior king of Chü-shih , afj ! ]'.
When the kingdom of Kao-cheang (Kara-khaja), i.e. Turfân, passed under Chinese domination in A. D. 640, the district head-quarters of ring Chou A Al were established there. Subsequently this district was changed in A. D. 702 into the Protectorate of Pei-tying 1E Jam, constituting one of the Four Garrisons among which the Chinese dominions in Central Asia were administratively divided. The above passage of the Chu Trang shu also states that the ancient royal seat of Posterior Chü-shih comprised five towns, ` whence its popular designation was the territory of the five towns
A '. Hence Hsü Sung, and apparently other Chinese antiquarians before him, very justly
concluded that the well-known Turki name I3ésh-balik, meaning ` the five towns ', borne during
1 See Chavannes, Turcs occid., p. II.
2 Fu-yuan=it is the Chinese official name in present use. The former designation, as given in the Hsi yü shui
tao chi, was Pao-hui 1* * ; see Chavannes, Turcs occid., p. I I.
3 Cf. Chavannes, Turcs occid., pp. II, 272.
4 This inscription may be the same as that referred to in the abstract from the Hsi yu lu, the work of Chingiz Khan's minister Yeh-lü Ch`u Tsai, as being at Bésh-balik ; cf. Bretschneider, Med. Researches, i. p. 15.