This brings us finally to the question what part the route followed by us from Mao-mei and Chin-tea to Ming-shui, and thence bifurcating towards Hâmi and Bai, might have played in the history of China's relations with the region along the easternmost extremity of the Tien-shan. Whether any references to its use can be traced in Chinese historical records is a question which must be left to others to determine. But judging from the physical conditions now prevailing along it we may, I believe, safely assume that it could never have competed in importance, for the purpose of general traffic and military movements, with the route crossing the Pei-shan desert between An-hsi and Hâmi. This is sufficiently proved by the fact that the number of marches to be covered between places possessed of cultivation is nearly twice as great on the route we followed as on the latter.
At the same time it is certain that the track we followed could not have offered any serious obstacles to the movements of small bodies of hardy nomads bent on raids upon the Kan-su marches, nor even to tribal treks on a minor scale in either direction. As long as security' was assured by an established power controlling the territories both south and north of the Pei-shan, a certain amount of trade carried solely on camels might always have made use of this and the parallel routes farther west in preference to the ` high road '. Grazing along the latter must necessarily have been far more restricted, not merely on account of the increasing aridity of the ground as it approaches the absolute wastes that stretch north of the terminal Su-lo-ho valley, but also as a result of the inroads that continuous traffic would inevitably make upon the scanty patches of vegetation to be found at the rare wells or springs.
SECTION III.-PAST THE KARLIK-TAGH AND BARKUL
From Bai I started on September 28th on the journey that in a space of four weeks carried me westwards along the northern foot of the Tien-shan as far as the site of the ancient Pei-ting near Guchen, and thence south across the snow-covered portion of the range down to the Turfân depression. This was to serve as our base during the winter, and we must reach it rapidly, having regard to the time that the work there awaiting us would probably require. The direct route which we were obliged to follow led for the most part over ground the topography of which was comparatively well known, and hence offered little opportunity for fresh exploration except in its concluding portion. Yet a special historical and geographical interest made me anxious to see as much as circumstances would permit of the ground to the north of the Tien-shan ; for physical features, which in some respects curiously recalled those of another true land of passage, the region along the northern foot of the Nan-shan, had caused it to play an important part in great historical migrations, such as those which carried westwards the Great Yüeh-chih or Indo-Scythians, the Huns and Turks.
To the west of the pass connecting Hâmi with Barkul our route necessarily followed the well-known high road connecting Hâmi with all the chief places along the southern part of Dzungaria. It has served as an important artery of traffic whenever China's trade and political control extended into Central Asia, and it has been frequently taken by modern travellers. To the east of the pass, likewise, the physical character of the ground we traversed along the foot of the Karlik-tâgh had been previously examined with care by a very competent student, Mr. Carruthers.' It will therefore be possible for me to deal briefly with our journey as far as Guchen and to restrict more detailed observations to such points as have a direct bearing on the historical past of this region.
Our first march from Bai to Atürük, the largest village north of the Karlik-tâgh, served aptly to illustrate the change to the less arid climatic conditions that favour the tracts along the Tien-shan
1 See Carruthers, Unknown Mongolia, ii. pp. 521 sqq.