ACROSS THE PEI-SHAN TO BARKUL
SECTION I.--THROUGH THE DESERT RANGES OF THE PEI-SHAN
Programme THE journey for which I had chosen Mao-mei as our rendezvous and starting-point was
for journey intended to take us right across the great desert area occupied by the Pei-shan ranges where its
and Turf width is greatest, in the direction from south-east to north-west. Following routes which the Russian
and Turfân. g g
Trans-frontier Map vaguely marked as derived from native information, I proposed to make my way to the eastern flank of the Karlik-tâgh, itself the easternmost extension of the Tien-shan. Thence I intended to skirt the northern slopes of this great chain to beyond Ku-ch`êng-tzû or Guchen, from which point we could cross it to the Turfân depression, the ground chosen for our next archaeological labours. The execution of this programme would allow us in the first place to survey a large and practically unexplored portion of the Pei-shan, and next to see something of south-eastern Dzungaria, which by its geographical and historical connexion with the oases to the south of the Tien-shan was of special interest to me.
Attempts to The extent of unsurveyed desert to be crossed to the nearest inhabited place north-east of
secu nee the Karlik-tâgh was very great, and knowing the difficulties to be expected in connexion with
through water and grazing, I had already in May, on my first passage through Mao-mei and along the
Pei-shan. Etsin-gol, endeavoured to secure reliable information about the reported routes and guides
acquainted with them. The result of these inquiries had been extremely meagre. The existence of routes occasionally followed by camel caravans towards Hâmi and to Bai, my particular objective, was known to some Chinese traders at Mao-mei. But the attempts to secure Mongols acquainted with them as guides had been unsuccessful. On my return I had to rest content with engaging two Chinese, labourers by profession, whom the young district magistrate, willing to help as before, had managed to produce. They stated that they had accompanied camel caravans proceeding by the direct routes to and from Barkul and H ami, and as they wished to return to those places for work, they were prepared to act as guides. Though the account they gave of their itineraries sounded rather vague, I was glad enough to accept their services ; for I knew that previous map-work could assist us only at one point, the cross-roads of Ming-shui which they mentioned, the position of this being fixed in relation to the routes farther west visited by Russian travellers and by Professor Futterer.
Part of In view of the length of the journey before us and of the total absence of resources on the way,
detached, careful arrangements had to be made for the supply of food for men and ponies. Fortunately these
were facilitated by the good harvest which the adequate flood of the Kan-chou river had assured to Mao-mei that summer. In order to lighten our impedimenta as much as possible I decided to send off Li Ssû-yeh, accompanied by Naik Shamsuddin and two Turki followers, with spare baggage to Su-chou. Thence they were to proceed to An-hsi, and picking up there the cases of antiquities, &c., deposited under the care of the faithful Ibrahim Bèg, to move on by the Chinese high road' to Turfân, where we were to meet at the end of October.
Start from In order to enlarge the area mapped, the rest of us were to move, wherever possible, in two
Mao-mei. parties and by separate routes. An opportunity for thus proceeding offered at the outset, as our