Geography in the United States. — Davis. i8i 
ficult, because the beds on which glaciers he cannot be well 
examined. The other method is to deduce the appropriate 
consequences of both the affirmative and the negative suppo- 
sitions, and then to confront these consequences with the facts 
found in regions once glaciated, and see which set of conse- 
quences is best supported. This deductive method is very 
simple. Its application involves no principle that was not per- 
fectly well known fifty years ago, though it doe.s involve a 
facility in theorizing that does not seem to have- been familiar 
or habitual with geographers until recent times. On the sup- 
position that glaciers do not erode, the valley systems of once 
glaciated mountains ought not to exhibit any significant pecul- 
iarity of form, but should correspond to the normal stream- 
worn valley systems of non-glaciated mountains. On the sup- 
position that glaciers do erode, the valley systems of once glaci- 
ated mountains should exhibit the highly specialized feature of 
a discordant junction of branch and trunk ; for the channels 
eroded by a small branch glacier and by a large trunk glacier 
must stand at discordant levels at their junction, just as the 
channels of a small stream and a large river do, though the 
measure of discordance is much greater in the channels of the 
clumsy, slow-moving ice-streams than in the channels of the 
nimble, quick-moving -vater-streams. There can be no. ques- 
tion that these well specialized consequences, deduced from the 
postulate that glaciers can erode their channels, are much more 
accordant with the actual features of valley systems in once 
glaciated mountains than are the consequences deduced from 
the opposite postulate ; but my reason for introducing this 
problem here is not to call attention to the value of "hanging 
valleys" in evidence of glacial erosion, as first clearly set forth 
by Gannett in 1898 in his account of Lake Chelan, but rather 
to point out how slow geographers have been to employ the 
deductive method in solving this long-vexed problem. The 
moral of this is that geographers as well as geologists, phys- 
icists, astronomers ought to have good training in scientific 
methods of investigation, in which all their faculties are em- 
ployed in striving to reach the goal of full understanding, in- 
stead of depending so largely on the single faculty of observa- 
tion. 
