________________
227
OCTOBER, 1917] THIRD JOURNEY OF EXPLORATION IN CENTRAL ASIA
ment of Kirghiz grazing in the open valleys to the south-west of the Great Kara-kul. Having obtained there fresh transport from imposing old Kokan Beg, the Ming-bashi of the northern Pamirs, and having started my anthropometric work, I moved down the Tanimaz Valley to its junction with that of the Murghab or Bartang River. Here at the picturesque hamlet of Saunab, the Tashkurghan of the Kirghiz, I reached the first Iranian-speaking settlement of hill Tajiks or Ghalchas, all fine-looking men. Their ethnic type of pure Homo Alpinus, their old-world customs, preserved by alpine isolation, and the survival of much that seems ancient in domestic architecture, decorative motifs, etc., interested me greatly and amply justified a day's halt, which allowed me to secure anthropological measurements and arrange for the load-carrying men we needed.
The only route open to us for reaching the southern Pamirs led up by the Bartang River, and progress in its narrow gorges proved exceptionally trying owing to the results of the great earthquake of 18 February 1911, which had transformed the surface of this mountain region in a striking fashion. Already on the lower Tanimaz we had come upon huge masses of rock débris which had been thrown down from the slopes of the flanking spurs and now spread for miles across the open valley bottom. Here in the defiles of the Bartang the huge landslides attending that memorable earthquake had choked up in many places the whole river passage and practically destroyed what tracks there ever existed along or above it. The big river once rivalling in volume the main feeder of the Oxus, the Ab-i-Panja, had here ceased altogether to flow. Strings of deep alpine tarns, with colours of exquisite beauty, had replaced it here and there and helped to increase the difficulties of progress. It took three days' hard scrambling along steep spurs, almost impassable for load-carrying men, and over vast slopes of rock débris spread out in wildest confusion, to get beyond the point near the mouth of the Shedau side valley where the fall of a whole mountain has completely blocked the river, and converted the so-called "Sarez Pamir" into a fine alpine lake over 15 miles long now and still spreading up the valley."
Enormous masses of rock and detritus had been shaken down from the range on the north and had been pushed by the impetus of the landslip up the steep spur flanking the Shedau debouchure. They had thus formed a huge barrage, which even now seemed to rise more than 1200 feet above the level of the new Sarez Lake, and is likely to dam it up for years, if not for centuries. It cost another day's stiff, and in places risky, scramble before we succeeded in getting the baggage safely across the few miles of precipitous rock slopes and dangerous débris-shoots above the Yerkht inlet. Fortunately the men collected from the uppermost hamlets of the Roshan Valley proved all excellent cragsmen and quite expert in building rafaks, or galleries of brushwood and stones, along otherwise impassable precipices.
Opportunely succoured by Kirghiz ponies, which had been sent from the Alichur Pamir to meet us, we crossed the Langar Pass, close on 15,000 feet above the sea, by August 20. It gave us easy access to the Yeshil-köl Lake, where I found myself on ground of varied geographical interest. I can mention only two points here and those in all briefness. On the one hand, with the experience gained at the newly formed big lake fresh before me, it
In an important paper (Comptes rendus de l' Académie des Sciences, clx. pp. 810 sqq., Paris, 1915), reference to which I owe to Mr. E. Heawood's kindness, Prince B. Galitzine has shown strong reasons for the belief that the Sarez landslide was not the consequence but the cause of the earthquake of 18 February 1911, which was registered at many distant seismological stations. This earthquake is declared to present an exceptionally interesting case where the epicentre can be proved to coincide with the hypocentre itself.
