= 
— 1 
stationary, while the ice moves on, through the backward ero- 
sion and melting of its up-stream side; and, second, that when 
a pothole is formed at the bottom of a moulin it is not the 
direct impact of the water upon the face of the ledge that does 
the work, nor do the stones carried down by the water wear 
the ledges appreciably by their direct fall, but the pothole is due 
to their subsequent movement and especially their rotation by 
the water. This rotation implies an antecedent depression or 
hollow to hold the stones, and thus the conditions are seen to 
be essentially the same as for ordinary river potholes. Since 
the rotation of stones in a pre-existing hollow is an essential 
condition of the glacial as of other potholes, and the moulin 
simply supplies the power, it would seem to make little or no 
difference whether the water plunges into the up-stream side, 
the middle, or the down-stream side of the hollow. The pot- 
hole will be made where and only where the hollow is; and 
during the progress of a moulin across the hollow there 
would not, apparently, be any marked tendency to elongate it. 
In the case of a linear group of potholes on the lee slope of a 
ledge, it 1s reasonable to suppose that the upper one, which, on 
Cooper's Island, is always the smallest and most indefinite, marks 
the shifting position of the moulin, and that the others were 
formed by the subglacial flow of water from the bottom of the 
moulin. 
Reverting once more to the question as to what phase of the 
continental elaciation——the beginning, the maximum, or the 
waning—was most favorable for the formation and preservation 
of glacial potholes, it may be noted: First, that during the de- 
velopment of the ice-sheet precipitation was mainly in the form 
g; the true 
of snow, and greatly exceeded the waste by melting 
glacier ice was covered by a great thickness of névé and snow ; 
the water escaped largely by seepage ; and the superficial streams 
were small. Second, that while the growing sheet of ice, névé, 
and snow must have been virtually stationary for a long time, 
it must also have attained quite a high velocity (for an ice-sheet) 
