Sec. iv] HISTORICAL RELATIONS BETWEEN BARKUL AND HAMI 543
from the desert to the north-west of the Jade Gate barrier.18 I have shown in Serindia that this route between the Jade Gate and Posterior Chü-shih, first opened A. D. 2, must necessarily have crossed the Tien-shan by the easy saddle over which passes the present Chinese cart-road from Hàmi to Guchen, between the stations of Ch`i-ku-ching and Ta-shih-t`o (Map No. 31. c. I ; D. 2).19 Eastern Chü-mi, like the rest of the small ` kingdoms ' dependent on Posterior Chü-shih, must have lain on the northern side of the Tien-shin. Hence we can safely locate it in the valleys and plateaus to the west of the Barkul lake which are reached across that saddle and which we crossed on our way from Barkul to Guchen, as seen in Maps Nos. 34. A. I ; 31. A—D. I.
I shall have occasion farther on to give a brief description of this region ; but I may at once observe that its physical character entirely agrees with what the Hou Han shu tells us of the Eastern Chid-mi.20 The territory is there said to include over three thousand households and some two thousand fighting men. Its people are described as nomads living in huts and tents and leading a pastoral life, agriculture being practised only to a small extent. The Tien-shan sinks to a much lower elevation to the west of Barkul, before it rises again to a crest line carrying permanent snow in that portion of the range which divides Guchen from Turfân. Consequently there is less moisture to be found in the valleys west of Barkul until the forest-clad slopes east of Mu-li-ho (Map No. 31. A. i) are reached. Yet grazing grounds are to be found in most of this area, and also patches of cultivation, which gradually increase in size and importance as the tract of Guchen is approached. Since Western Chü-mi is not mentioned in the Later Han Annals we may conclude with their commentator, quoted by M. Chavannes, that this territory had then probably been absorbed by Eastern Chü-mi.21 This again would account for the latter being credited with a population considerably in excess of that attributed to I-chin or the Barkul basin.
From this survey of the territories along the northern slopes of the Tien-shan with which Barkul appears, at any rate for a time, to have been linked, we may now return to the remaining historical data concerning it preserved in records of the Later Han period. The freedom from Hun inroads secured by Pan Yung was destined to be but a short one. As early as A.D. 131 we read that the Chinese were under the necessity of placing a military colony at I-wu or Hàmi in order to prevent the Hsiung-nu from making it a base for their depredations.22 In A.D. 135 Posterior Chü-shih was again attacked by the Hu-yen king of the Northern Hsiung-nu, and a Chinese force sent against him from Tun-huang failed to achieve success.23
A better result attended the effort made two years later by Pei Ts`ên . , prefect of Tun-
huang, of whose victory the only surviving memorial is an inscription, originally set up in a temple by the Barkul lake and now preserved in a temple outside the north-west corner of Barkul town.24 It records the success gained by Pei Tseên, who, with a force of three thousand men, is said to have destroyed the king Hu-yen and his people in a battle fought in September A. D. 137. But this victory, too, which curiously enough the historical texts of the Later Han pass over in silence, had no lasting result. In A. D. 151 the Hu-yen king, as the Hou Han shu tells us, was ravaging I-wu or Hàmi with three thousand horsemen. A Chinese detachment sent to the east of the Peu-lei or Barkul lake was completely annihilated by him. A Chinese force sent from Tun-huang in the same
Description of ` Eastern Chü-mi '.
Historical data for Barkul.
Inscription of P`eiTs`ên, A. D. 137.
18 Cf. Chavannes, T'oung-pao, 1905, pp. 556 sq.
19 See Serindia, ii. pp. 705 sq.
2° Cf. Chavannes, Toting-pao, 1907, p. 21o. 21 See ibid., 1905, p. 557, note I.
222 Cf. Chavannes, Toung-pao, 1907, p. 167.
23 See Chavannes, Dix inscriptions, p. 23 ; T`oung-pao,
1907, pp. 213 sq.
24 For this inscription, first discovered in
Chavannes, Dix inscriptions, pp. 17 sqq. It was previously edited and translated by M. Devéria in Grenard, Mission D. de Rhins, iii. pp. 136 sq.
The temple which now shelters this, the oldest epigraphic record of Chinese Turkestan, is appropriately enough dedicated to the Chinese war-god Kuan-ti. It offered me peaceful shelter in October, 1914.