Sec. i] THROUGH THE DESERT RANGES OF THE PEI-SHAN 525
pool below the easternmost buttress of the range that we had crossed from the south, and thence, as the map shows (No. 42. B, C. 3), along the northern slopes of the same. At the spring of Chi-chich`üan, at the foot of the reddish hills already referred to, he had found excellent grazing.
From Mou-wo a perfectly open plateau stretched away to the north-west, sloping up gently in that direction and sending its dry drainage beds down towards the north-east. On the gravel
surface; bare but for scanty tufts of scrub, the caravan track showed up clearly and allowed us to cover with ease a march of close on 25 miles to the well of the Nan-ch`üan. Some six miles before
reaching it we crossed a broad and much-decayed hill chain with its crest rising only about 35o feet above the elevation of \'Iou-wo and almost completely buried in detritus. It was seen to sink away to the north-west into a wide depression, which appears to gather whatever drainage passes down the ground crossed on this and the next two marches. This chain and two other low spurs, marked by crumbling rocky ridges, over which our route led between Nan-ch`üan and Lo-t`o-ching (Map No. 40. D. 3), probably represent eastern offshoots of the fourth Pei-shan range of Futterer. According to his map and description this range is broken up farther west also into a series of parallel chains. To this assumed nexus points also the east-north-easterly trend that he records for the range as a whole.2
Our march of September 9th took us, after we had proceeded about 13 miles, across the crest of another gently rising and much broken hill chain, at an elevation of about 5,30o feet, similar to
that of the chain passed on the way to Nan-ch`üan. At the mouth of the winding Nullah in which
the route descended from it to the north-west we passed low dark red cliffs which looked to me like porphyry. The well where we halted, called by our guides Hung-tou-span-ching, lay on the bare
gravel slope of a wide open valley offering but the scantiest scrub for the camels and no grazing
whatever for the ponies. Low broken ridges limited the horizon to the west, but northward a distant vista opened over a wide trough-like depression. Beyond it far away rose the rugged crest line
of a range which after the flatness of the ground passed on the last two marches looked quite
impressive. Our guides recognized it as the Ma-tsun-shan, which had been mentioned to us at Mao-mei as containing valleys regularly grazed by the flocks of 50 or 6o Mongol families from the
side of the Etsin-gol. The Russian Trans-frontier Map roughly marks it to the north of the route towards Hâmi. Of the large river along which this reported route was supposed to approach the range we could see nothing.
On September loth a short march took us first over an utterly bare plain of gravel, where only strips of detritus io to 15 feet high marked the position of low ridges completely decomposed.
Then another flat spur was crossed at about the same elevation as that of the two last mentioned. The wide basin-like valley beyond seemed far too barren to offer an opportunity for a halt. But after crossing it westwards the track brought us to a little hollow containing a rather brackish well, called by our guides Kuo-ti-ching,3 and a small patch of reeds growing round sandy hillocks. Here we were glad to camp and let our animals have some badly needed grazing.
The next day's march lay first for about io miles across a wide valley containing numerous low terraces of a very fine grained almost black rock of sedimentary origin. Then a gentle ascent
past rounded flat-topped ridges brought us on a wide plateau to a large reed-filled basin watered by the springs of Lo-t`o-ch`iian. These, about a dozen in number, gather near the western edge of a depression close on two miles wide. They are evidently fed by subterranean drainage from the chain of hills visible to the west and south-west of Lo-t`o-ch`üan. This chain rising to heights