i8 9 i.] 
229 
[Upham. 
and slopes upon the bed-rocks of plains and hills, or is amassed 
in round, oval, and elongated drumlins, or is irregularly heaped 
in the hillocks and ridges of terminal moraines. It is evident 
that these till-enclosed lake basins were formed by inequalities in 
the deposition of drift by the ice-sheet, without modification by 
water action. This class of drift lakes is probably twice as nu- 
merous as that discussed in the present paper. 
Glaciated regions have also two other classes of ponds and 
lakes, including those that lie in rock basins. Many tarns and 
small lakes in mountain districts fill hollows that have been 
scooped out of the bed-rock by glacial erosion. Lakes of this 
origin are less frequent or rare upon moderately hilly and even 
plain country, where inequality in the hardness of contiguous rock 
formations, or difference in their exposure dependent on their dip 
and strike, has permitted the ice-sheet to wear more deeply on 
some tracts than elsewhere all around them, producing lake 
basins. But if the drift covering the glaciated surface of the 
bed-rocks were removed, doubtless multitudes of shallow rock 
basins would be seen. This class of glacially eroded tarns and 
lakes contributes much to the beauty of many alpine landscapes, 
where the ice-sheet was efficient to abrade the rocks, rather than 
to deposit drift ; but it is scantily represented on the lowlands. 
It probably comprises only lakes of small or moderate size, up to 
limits of a few miles in length in mountain valleys, but to many 
miles on lower drift-bearing areas. 
The larger and very often deep lakes in rock basins of glaci- 
ated mountain districts, as the lakes of Geneva, Neuchatel, and 
Constance, the Lago Maggiore, and others in the Alps, and of 
continental expanses which have been covered by land-ice, as the 
great lakes of North America tributary to the St. Lawrence, the 
Nelson, and the Mackenzie, seem to me to constitute a fourth 
class, different in origin from the last, not being attributable to 
excavation by the ice, but to deformation of the earth’s crust 
through flexure and faulting, whereby portions of preglacial 
valleys of river erosion have become transformed into lakes. The 
more frequent occurrence of these lakes on glaciated than on un- 
glaciated areas seems to have been due to movements of elevation 
which with other conditions caused the ice accumulation, and to 
movements of subsidence which appear often or generally to have 
resulted from the pressure of the ice weight. 
