begin. During this process of slow accumulation of the ice- 
sheet, the summer melting upon its surface would produce 
multitudes of rills, rivulets, and brooks, which might unite 
into a large stream, and this, pouring through a crevasse and 
melting out a cylindric moulin, might fall a considerable depth 
to the bed rock, perhaps one or two. hundred feet or more upon 
an area so moderately uneven as Cohasset, while yet the ice- 
motion, though sufficient to permit the formation of the crevasse, 
might not have gained a definite current to carry the crevasse, 
moulin, and water-fall away from the spot where they were 
first formed. We may thus explain the continuation of a glacial 
water-fall in one place while it was excavating one of these 
** Giant's Kettles " or potholes. After the ice-sheet acquired a 
current because of the greater thickness and pressure of its mass, 
such deep cylindrie excavations in the bed-rock could not be 
made; and during the recession and final dissolution of the ice- 
sheet, it seems probable that its receding border had steeper 
gradients and consequently even more rapid motion than in the 
culmination of the glacial epoch.” 
In criticism of this view it may be pointed out, first, that 
while the movement of the ice-sheet. was still practically nil, it 
would probably have accommodated itself to the irregularities of 
its rocky floor without cracking; second, that the face of the 
country, before it had been swept by the long-continued advance 
of the ice-sheet, was probably buried beneath a considerable 
thickness of soil and half-decomposed rock, the product of 
chemical and mechanical decay during long preglacial ages ; 
and third, that it makes no adequate allowance for glacial erosion 
of the hard rocks during the long periods of the maximum 
development and waning of the ice-sheet, which would inevitably 
have obliterated the potholes. In short, Mr. Upham’s view 
'arries as its logical consequence the reduction of glacial erosion 
not only to aminimum, but almost to nothing. 
While seeking for additional light upon this problem it has 
oceurred to me, first, that a moulin may remain approximately 
— me 
