29(5 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY. 
channel in the bottom of the valley will in time be given a depth 
and width that will better suit the needs of ice discharge than did 
the initial basin or trough of the uplifted surface. The upper sloj^es 
of the glacial stream will thus be steepened, while its lower course 
will be given a gentler descent. Owing to the diminution of the 
glacier toward its lower end, the channel occupied by it will dimin¬ 
ish in depth and breadth downwards from the point of maximum 
volume; this being analogous to the decrease in the size of the 
channel of a withering river below the point of its maximum volume. 
A time will come when all the energy of the glacier on its gentler 
slope will be fully taxed in moving forward the waste that has been 
brought down from the steeper slopes; then the glacier becomes 
only a transporting agent, not an eroding agent, in its lower course. 
This condition will be first reached near the lower end, and slowly 
propagated headwards. Every part of the glacier in which the 
balance between ability to do work and work to be done is thus 
struck may be said to be “ graded ”; and in all such parts, the sur¬ 
face of the glacier will have a smoothly descending slope. Maturity 
will be reached when, as in the analogous case of a river, the nice 
adjustment between ability and work is extended to all parts of a 
glacial system. In the process of developing this adjustment, a 
large trunk glacier might entrench the main valley more rapidly 
than one of the smaller branches could entrench its side valley ; 
then for a time the branch would join the trunk in an ice-rapid of 
many seracs. But when the trunk glacier had deepened its valley 
so far that further deepening became slow, the branch glacier would 
have opportunity to erode its side valley to an appropriate depth, 
and thus to develop an accordant junction of trunk and branch ice 
surfaces, although the channels of the larger and the smaller streams 
might still be of very unequal depth, and the channel beds might 
stand at discordant levels. If the glaciers should disappear at this 
stage of the cycle, their channels would be called valleys, and the 
discordance of the channel beds might naturally excite surprise. 
The few observers who, previous to 1898, commented upon a dis¬ 
cordance of this kind, explained it as a result of excessive erosion of 
the main valley by the trunk glacier; while the hanging lateral 
valleys were implicitly, if not explicitly, regarded as hardly changed 
from their preglacial form. 
When the trunk and branch glaciers have developed well-defined, 
maturely graded valleys, the continuous snow mantle that covered 
the initial uplands of early youth is exchanged for a discontinuous 
