to Maral-bashi and thence to strike across the desert to the Khotan river by the nearest route which might prove practicable. As far as Maral-bashi I proposed to follow the foot of the barren
hill chain which here forms the southernmost rampart of the Tien-shan. Two reasons combined to make me choose this route. The first was that this range almost throughout its length from
the hills north of Artush to south of the Kelpin oasis (Maps Nos. 5. A—D. 1; 8. A. I ; 7. A, B. 4) had
as yet remained practically unsurveyed, as appears from the portion of Dr. Hassenstein's map of the Tarim basin that includes it.4 The other reason was archaeological and had its origin in
reports first heard at Maral-bashi in 1908 and again later on the Kashgar side.5 These seemed to indicate that an old route, now but vaguely remembered in local lore, had during some earlier period skirted the foot of that chain and been used for traffic instead of the present Chinese ` high road ', or more correctly cart track, leading much farther south along the actual course of the Kashgar river.
Having previously dispatched all baggage that was not immediately needed by trader's caravan to Khotan, I took the precaution, as I expected water to be scarce, of reducing my party
still further by sending Surveyor Muhammad Yagiib along the high road to await me at Maral-
bashi. Accompanied by Afraz-gul Khan and Naik Shamsuddin I set out myself to the northeast. The first day's short march brought us through fertile ground, which I had already seen
in 1900, to Sedir in the canton of 13ésh-karim. Passing again the well-known shrine of Bû-Mairyam
and sighting beyond Eski the Stùpa ruin of Mauri-tim which had been surveyed in the same year,6 we then made our way to the large oasis of Astin-Artush (Map 5. A. 1). Its famous Ziarat
of Sultan Boghra Khan has been visited and described by the Forsyth Mission.' We passed it as darkness fell, on our way to our night's quarters at a large farm at Wakwak, not far from the eastern end of the cultivated zone of Artush.
On October 11th a long march brought us to Kalta-yailak, a narrow and far-stretching belt of hamlets, which forms the terminal oasis of the river of Artush. Almost the whole of our route
led over desert ground near the foot of the outermost hill range. The big bends in which the
river winds here eastwards indicate the almost level nature of this ground. There is no marked glacis along the foot of this range, and as a consequence former river-beds and marsh-beds still
liable to inundation from the south approach close to it. These were subsequently found to be
characteristic features of the whole of the ground traversed by us on the way to Maral-bashi. It presents a striking contrast with the wide glacis slopes of piedmont gravel that everywhere
edge the foot of the K`un-lun range on the opposite side of the Tarim basin. This contrast is clearly
due to an essential difference in the morphology of the two mighty ranges which form the north and south rims of that basin, and therefore may here receive passing mention. The surveys carried
out by Lai Singh in 1908 and 1915 among the outer hill chains of that portion of the Tien-shan which lies between the longitudes of Kelpin (Map No. 7. B. 3) and Kashgar may be of use to others more competent than myself to follow up the point.
I had previously received information about some old remains at the very foot of the outermost hills, and consequently left the track leading to Kalta-yailak beyond the river crossing of Kichik-langar.8 After passing over a desolate plain covered with scanty scrub and much salt
4 See ` Karte des Tarim-Beckens ' in Hedin, Reisen in Z.-A., Petermann's Mitteilungen, Ergänzungsheft No. 131.
5 Cf. Serindia, iii. p. 1307, 1310.
6 Cf. Ancient Khotan, i. pp. 79 sq.
7 See Yarkand Mission Report, p. 17; Figs. 41, 42.
8 The bed, where we passed it again after marching about 7 miles beyond Kichik-langar, was about 3o feet
wide and held water to a depth of close on 2 feet ; the measurement taken indicated a volume of not less than 8o cubic feet per second. During the period of the summer floods the river was said to fill the bed to a height of some 7 feet more on the average. These indications suggest the possibility of far more extensive cultivation than is to be seen at Kalta-yailak at present.