DESERTS. 285 
perhaps the unconformity between Tertiary and Quaternary time. This uptilted 
series contrasts with that above in being of fine silts and sandstones, with only 
an occasional bed of conglomerate, and is much more consolidated. But its 
detailed structure is that of alluvium and appears exactly like those silts deflated 
on the surface farther out and whose thickness up to 4o feet is exposed in channels 
cut by larger streams. The beveled strata rise more and more to view with 
occasional masses rising above their general surface of erosion as monadnocks 
buried in the later piedmont, and, in time, some ridges rise out into open air, thus 
taking part in the topography of to-day. 


Fig. 463.—Terraces of Markan Su. 
The canyon is now perhaps 500 feet in depth and cut half in the older series. 
Its depth increases to 1,000 feet and more while the older series rises, outcropping 
from the now much-dissected piedmont more often, until at length we find it as 
spurs projecting from border masses of ranges entirely composed of it (the older 
series), and its tilt has steadily increased. These mountains, thus composed of 
desert piedmont strata built as these now forming, were in Quaternary time the 










Fig. 464.—Deflated Silts and Residual Sand-dunes (Tarim). 
source of the red silts and conglomerates composing the more recently dissected 
piedmont and are together with them the source of materials laid down on the 
oasis playas and piedmonts of to-day. Continuing the journey into the mountains 
there were seen remnants of a high-uplifted and much-dissected topography of 
mature form and doubtless corresponding to the dissected piedmont derived from 
them. This topography, none of which may be termed a peneplain, but which 
nevertheless conforms to that found on the border ranges of Pamir, is doubtless an 
equivalent of the mature stage found by Davis in the Tian Shan and Bailey Willis 
in Central China. 
Such in general are the great features of Western Tarim. Yet one character- 
istic of importance should be considered. We have seen how a wide margin of 
