40 SEASONAL DEPOSITION IN AQUEO-GLACIAL SEDIMENTS. 
The swinging of a stream across a delta will give rise to irregular, discon- 
tinuous, alternating coarse and fine material. Such banding must be dis- 
tinguished from that which originates in front of a delta in deeper water. 
There may be, in certain locations, transitional zones between the delta-type 
of banding and seasonal-type. In case the glacial stream enters deep water on 
leaving the glacier, the delta-type of banding may not be found. (See p. 27). 
EK. In the retreat of ice fronts it is recognized that those in water deep 
enough for the calving of bergs, retreat much more rapidly than those on land. 
(Tarr, 1912, p. 19-20). Ice fronts in salt water retreat more rapidly than those 
in fresh. If a large ice sheet should have parts of its front in water deep enough 
for the formation of icebergs, and parts on land, the parts in water should occupy 
reéntrants in the front. In such cases streams from the land ice might enter 
the main basin of deposition at points miles below the part of the front in water, 
and mingle their sediments with the sediment from streams debouching from 
the part of the front in water, thus confusing the consequent stratification in 
the seasonal components of the banding. In such cases it would be impossible 
to identify the different seasonal events by means of minute characters of the 
banding. It should be possible, however, to distinguish winter and summer 
effects, but very little more. 
A similar mixing of sediments would occur in the case of several branch 
streams from different glaciers entering a main-trunk stream which empties 
into a lake or other body of water. The Rhone Valley from its head in the 
Rhone glacier to the Lake of Geneva may be taken as an example of such a case. 
At least twenty streams from glacial ice enter the Rhone above the Lake of 
Geneva. The study of the layers in the Lake of Geneva could be of very little 
value as a record of the weather for any season. Winter and summer deposi- 
tion, however, should be registered. 
F. In the subglacial channels, stream capture must go on, in much 
the same manner as with rivers in general. When such a capture takes place 
it would be relatively sudden. In this manner the sudden increase in water 
of the capturing stream would cause a thickening of the seasonal layers in the 
area of deposition of that stream. 
G. In shallow water wave action might stir up the sediment on the bot- 
tom for a short interval. When the sediment settled again there would be 
evidences of violation of normal deposition. Wave ripples should be found 
in sediments subject to wave action. 
Bergs falling from an ice front would cause waves, as is the case today 
with all the glaciers reaching water of any depth. 
