268 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
drawal of the ice from the divide, then, there is a coalescence of the 
many original lakelets into a smaller number of larger lakes ; some lake- 
lets are wholly drained off. Lakes increase in size, and decrease in 
number. At the same time, they come more nearly into an agreement 
of level, the higher levels being abandoned for the lower. 
Besides this merging into larger lakes, as the ice retreats from a 
divide, characteristic of early stages, there soon begins a succession of 
lowerings of level, brought about by the uncovering of lower cols in the 
rim of the basin. Each time a lower col is uncovered, a new outlet is 
formed, and the lake-level falls accordingly. So while the lakes are 
gaining steadily in area, during the withdrawal of the ice boundary, they 
are always liable to a loss in area through the lowering of the water-level 
to fit a new outlet. Thus in the diagram, by the retreat of the ice- 
front from B..Btoc.. 0, the lake gains in area; but since during the 
retreat a lower col has been uncovered, at level 1, the water-level of the 
lake has fallen, and there has been a consequent loss in area. In this 
case, the gain through withdrawal has been greater than the loss by lower- 
ing of level. With continued withdrawal, then, there are both gains and 
losses in area, and successive lowerings of water-level to fit new outlets. 
Changes of this sort will go on in the ice-datnmed lake until at last 
the ice abandons the basin and allows the natural drainage to take 
its place. 
Factors which control the Stability of Ice-Front Lakes. 
The slower the rate of withdrawal of the ice-front, the longer will the 
lakes maintain their level and their form ; for it is through withdrawal 
that new outlets areopened and lowerings of level are brought about. 
The distance from an active col outlet to the next lower one, which will 
replace it in draining off the water, also controls in part the permanence 
of form and level of the lake. Again, the strength of material over 
which the overflow takes place — whether it be till or bed-rock — will 
affect the permanence of level of the lake, and therefore its form ; for if 
an outlet is actively cutting down its bed, it is lowering the level of the 
lake which it drains. 
The permanence of lake-levels is of great consequence to us, for on 
it depends the degree of development of all those features which bear 
record of the lake’s history. The longer a water-level is maintained, 
the greater and the more perfect will be the development of lake-shore 
forms and deltas, and the more data will be accessible for reconstructing 
the water-planes which they mark. 
