176 Over the Riin-tst-la ,; 
except far away in the south, where it looked like late 
evening, so subdued was the light. It grew darker and 
darker, and at last it dawned on me quite suddenly that 
this was an eclipse of the sun, a phenomenon I had never 
before witnessed. 
The eclipse was almost complete, the sun showing as a 
red disc surrounded by a narrow silver rim. Seen through 
a film of cloud the dark shadow of the obstruction was 
visible passing over from above downwards, and seen 
through snow glasses it appeared quite black. The eclipse 
began about 9a.m. and lasted for two hours or so, our 
position then being roughly 28° 50” N., 99° 15” E. The 
Tibetans had taken no notice of the growing dimness, but 
when they saw me staring up at the sun through blue 
glasses they began talking a good deal amongst themselves, 
though they were far from exhibiting signs of consternation 
and scarcely evinced surprise. 
Dropping down to the bottom of a deep ravine we 
began the ascent, the torrent here taking a series of im- 
mense leaps down a colossal granite stairway, hemmed in 
by cyclopean walls of the same rock. At first we scrambled 
up vast screes of rough blocks, through dense shrub vege- 
tation similar to that in the arid valley below, but in the 
shade of these cliffs growing far more luxuriantly, with 
festoons of Clematis and twining Polygonum hanging from 
every bush, and a rank herbaceous undergrowth all around. 
Then came scattered maples and willows, oaks, poplars, 
birches, elms, and other trees, with pines clinging des- 
perately to the precipices above. Just beyond the last 
cataract lay two or three immense blocks of granite fallen 
from the crumbling cliffs, one of which I estimated to 
contain at least 20,000 cubic feet. Riding between them 
was a dangerous game, for in some places the pony 
jammed, and I almost had a leg crushed. 
Beyond this point the valley began to broaden out, and 
quiet reaches followed, where there were groves of poplar 
and walnut trees. As already pointed out the mountain 
torrents in the arid regions reverse the normal sequence of 
valley structure. Instead of gorges at the source of a river 
opening out into a valley which grows wider and wider till 
at length a flood plain is formed, we find a wide valley in 
