PHYSIOGRAPHIC OBSERVATIONS. 1 35 
It is thus seen that, whatever these Alai \'alley escarpments are, and howe\-er 
they were produced, their fomiation took place during the time of the deposition of 
the new moraines obser\-ed. It is also clear, since they support the bottoms of the 
old broad-trough valle\s and are cut by the new, narrower troughs, that their forma- 
tion must have taken place between the establishment of the old and of the new 
trough floors. This places the large, old, worn-down moraines as contemporary 
with the former broad troughs and the new, fresh, narrower moraines as contempo- 
rar)' with the narrower or present trough bottoms. 
The Trans-Alai range is largely car\-ed into cirques. We had an excellent 
opportunity to stud)' those low down on the northern flank of the range. They 
may be divided into two classes — the very large and the small cirques. These 
cirques empty either in groups into a trunk trough or open directly into the Alai 
\'alley. The large cirques emptying directly into the Alai \'alley come down on a 
nonnal slope to the plain, breaking through the escarpments in which their sections 
form double cur\'es similar to the sections of the twice-troughed vallejs. In and 
in front of them lie piles of moraine with fresh topography. 
Fig. 98. — Section across Kizil-Art Valley at Bor Daba, looking north. 
The bottoms of small cirques opening direct!}- into the Alai Valley lie high up 
on the escarpments, and the moraines contained by them are worn down to smooth, 
low masses. There is no double cur\-e in their cross-section. The whole depres- 
sion with its moraine is, in cases, much dissected. 
The absence of the fresh moraines in the small cirques, and the absence of the 
secondary- depression through the escarpments, is good evidence that these small 
cirques were carved largely during the epoch or epochs predating the escarpments. 
From this we may reason that onl\- the large cirques of those in question were, to 
any great extent, glacially active during the epoch following the formation of the 
escarpment. It seems probable that this difference was because their upper slopes 
begin several thousand feet abo\'e those of the small cirques, and thus accumulated 
ice when the line of peipetual snow was above the smaller ones. A careful study 
of cirques at the critical levels might determine the perpetual snow-line of the later 
glacial epoch. 
Of existing glaciers we had, from a short distance, a good view of three in the 
respective valleys back of the long moraines described. Each appeared to lie on a 
moraine. Directly in front of the ice this moraine floor was cut b)- a small rounded 
