DAVIS: GLACIAL EROSION. 
307 
consumption of pre-existent mountains by the great extension of 
corrie-glacier floors, each similar to J' H', thus seems mechanically 
possible; but it is nevertheless climatically very improbable, and it 
seems to me deficient in not attributing enough work to normal pre- 
glacial erosion. 
Overdeepened Valleys and Oversteepened Walls. — As in the 
case of the normal cycle of denudation in which the life history of 
river systems is involved, so in the glacial cycle, all manner of com¬ 
plications may arise, causing great departures from the ideal case. 
The assumed initial land form may be a surface previously more or 
less dissected by river erosion, on which glaciers must then proceed 
to develop a drainage system appropriate to their own peculiar 
needs, as has been partly considered above in connection with glacial 
distributaries. It will be instructive to make out a «;ood series of 
examples illustrating different combinations of river and glacial 
action, and including young, mature and old river valleys, modified 
by young or mature glaciation. For example, the existing valley 
of the Rhue in the Central Plateau of France shows a submature 
river valley with incised meanders, moderately affected by young 
and relatively light glaciation; the valley of the Ticino in the south¬ 
ern Alps is a well-matured preglacial river-valley system, modified 
by strong submature glaciation. The fiords of Norway result from 
the submature and intense glaciation of a river-valley system whose 
stage of preglacial development is not yet well determined. 
Interruptions of regular progress in the glacial cycle must, as in 
the river cycle, be occasioned by elevation, depression, or deforma¬ 
tion of the land mass ; but no examples of complications of this kind 
can be adduced. Variations of climate may replace creeping glaciers 
in young, mature or old stages of development, by flowing rivers; 
and the early stages of such rivers are of much importance among 
existing geographic forms. Lakes, delaying the river flow, occupy 
the depressions of the glaciated surface, as has been known since 
Ramsay first pointed out the correlation of lacustrine and glaciated 
regions in 1861 ; but the analogy between lakes in the beds of 
melted glaciers and pools in the beds of dried-up rivers has perhaps 
not been sufficiently insisted upon. Waterfalls connect the streams 
that occupy the discordant beds of glacial channels, as has lately 
been clearly set forth. Landslides frequently occur after the sup¬ 
porting glacier withdraws from the oversteepened banks of its huge 
channel; fallen masses of this origin have been repeatedly mistaken 
for moraines in Alpine valleys, as has been lately shown by Bruck- 
