DESERTS. 279 
short and the base-level aggraded back into the valley again, refilling it with waste. 
During this process, and when it had refilled to a height of about 250 feet some 
three-quarters of the way upstream, the second-epoch glacier advanced to 45 
miles below the present ice front. This ultimate point is near the oasis of Mad- 
rushkent. ‘There, in the face of a 300-foot deep canyon section, may be seen a 
thickness of some 20 feet of finely stratified light-gray clays, contrasting in lightness 
and texture with 20 feet of overlying and 200 feet of underlying coarse gravels. 
But the important feature is its distorted stratification, evidently having resulted 
from a pushing of the ice front, which also beveled the distorted layers with a sur- 
face declining upstream or against the river’s grade, the very thing to be expected 
under the frontal margin of an advancing valley glacier. At intervals for about 
20 miles above this point there are moraines scattered over terrace G standing 
half-buried in its alluvium, and in a tributary canyon section (see plate 62) oppo- 


Post glacial\ Fes=2== 
alluvium 
(cobbles) 

Thrusted alluvium 
Glacial_=> hfs : 
boulders (fine light gray clays) 
Thrust ———————_» West,down stream 
Fig. 457.—Section of Glacier-thrusted Alluvium in the Zerafshan Valley, 45 Miles 
below the Glacier. 
site Packshiff some of this moraine is seen standing on an irregular surface of allu- 
vium scraped over by ice and partly buried by later waste. So the second glacial 
epoch came to a close as the third-cycle gorge continued filling. While aggrading, 
the valley had widened and, ere the next uplift came, established the present 
terrace G. 
FOURTH EROSION CYCLE. 
Cycle 4, with the last uplift, has resulted in the present canyon, a channel 
incised from the last meander held by the river at stage G and thus crossing often 
the old-filled valley of the third cycle. And in its present torrential fall of nearly 
6,000 feet in 150 miles it must be rapidly cutting down. Indeed, the deep rumble 
and grinding of cobbles heard beneath the river’s roar is ample indication of 
corrasion. 
The Zerafshan glacier, with its ancient moraine, its relation to other glaciers 
and uplift, the fine grindings it has supplied to loess steppes, and its influence 
on civilization, becomes of great interest. Only one epoch of abandoned moraine 
could be distinguished, and that remarkably far-reaching and of such antiquity 
that it must be classed as belonging to the first or second of the glacial periods. I 
have attributed it to the second, because it still rises from terrace G in good pres- 
ervation. Nowhere has the first epoch moraine been seen with its topography 
preserved. Recently the glacier has advanced into a part of the valley that had 
been ice-free for so long that its sides had struck an even slope to the flood-plain 
