I reached Singer by the route already surveyed in 1907 and, after securing.there Abdurrahim's youngest brother as guide, I proceeded westwards to examine localities where traces Of earlier occupation were reported. Passing thus from P'o-ch'êng-tzu to Shindi I was able to map interesting and as yet unsurveyed ground in the mountains; their rugged ranges and deeply eroded valleys were in striking contrast with the worn-down uplands met in the Kuruktagh further east. The Khangol peaks passed on this route seemed to exceed 10,000 feet and probably represent the greatest elevation of the Kuruk-tàgh. 7°
I then made my way south-eastwards over barren gravel plateaus to the salt spring of Yardang-bulak at the south foot of the Kuruk-tagh and by the second week of March entered the waterless desert to the south. Besides exploring certain ancient burial grounds I completed the survey of the Kuruk-darya, the dried-up river-bed which once carried the water.of the
Konche-darya to the Lou-lan sites and the ancient delta to the south. 71 •
The day after my return to Yardang-bulak I was rejoined by Afraz-gul whose safe arrival at this appointed desert meeting place I had been.eagerly awaiting.
Afrûz-gut's surveys Some anxiety about the safety of the overdue little party was justified
'along dried-up Lop sea.
by the truly forbidding nature of the ground he had to traverse and the length of the strain put on our brave camels. Afraz-gul had carried through the difficult programme laid down by me with remarkable completeness and intelligence, his success on this survey alone fully justifying the award to him two years later of the Macgregor Silver Medal by the Intelligence Department of the Indian General Staff.
Guided by a third brother of Abdurrahim he first gained Altmish-bulak by the most direct track leading due south of Deghar: 72 Thence he surveyed certain ancient remains in the extreme north-east of the once-watered Lou-lan area for the examination of which I had been unable to spare time a year earlier. Replenishing his supply of ice from the salt springs, he struck out to the south-east for the point where my explorations of the preceding year had shown the ancient Chinese route from Lou-lan to Tun-huang to have entered the salt-encrusted bed of the Lop sea. 73 From there he traced its shore-line to the south-west, making plenty of interesting observations on inlets and terminal flood-beds once carrying water from the Kuruk-darya. Finally he reached, at Chainut-köl, the northern edge of the area in which the spring floods of the dying Tarim spread themselves out to undergo rapid evaporation in lagoons and marshes. 74 He arrived, as I had intended, just before the usual inundation could interfere with his progress to ground affording some scanty grazing for his hard-tried camels.
After a few days' rest he turned northwards into the wind-eroded desert and striking Crossing from terminal the line of the southernmost branch of the ° Dry River ' traced more Tarim marshes to 'Dry remains of the ancient settlement discovered along it a year before.
River.' Finally after crossing my route of December, 1906, in an area of for-
midable dunes, he gained the main riverine belt of the Kuruk-daryâ along the foot of the outermost Kuruk-tagh. 73 From this exceptionally difficult exploration which had kept Afrâz-gul and his three plucky companions from contact with any human being for a month and a half, he brought back, besides interesting archœological finds, an accurate plane-table survey and careful records of topographical details such as 1 could not have hoped for from any of my surveying assistants employed on this or my previous expeditions.
From Yardang-bulak we moved westwards to the point known as Ying-p`an where the ancient bed of the Kuruk-darya, is crossed by the Turfàn-Lop track.
Lrxpinrationsnear There a short halt was made in order to explore interesting remains at
and near a fortified station situated at the debouchure of the dried-up
stream of Shindi and occupied during the early period when it guarded the ancient Chinese high road from Lou-lan. The same opportunity was used also for surveying the °belt of drift-