154 
reach of the surf. In short, both the situation and the forms 
of the potholes make it impossible to regard them as in any 
sense the product of marine erosion. It is equally clear 
that no ordinary or surface river ever flowed over these ledges. 
We are forced, therefore, to ascribe their origin to the action 
of glacial streams. But here the question arises as to whether 
they have probably been formed by subglacial rivers rushing 
along over the ledges, or by streams which, flowing over 
the surface of the ice-sheet, plunge through crevasses to 
the solid rocks below. Such a waterfall in the ice is called a 
moulin or glacial mill; and since the crevasses must often be 
of great depth—hundreds or even thousands of feet—these 
glacial mills are generally recognized as the most efficient of all 
agencies in the formation of potholes. Subglacial streams, 
it is believed, must usually, like ordinary rivers, follow the 
depressions of the rocky surface; while moulins—the chief 
source of the subglacial stream—may strike the ledges with 
resistless force at almost any point, and especially on the crests 
and southeastern or lee slopes of prominent ledges and ridges, 
where the conditions are most favorable for the formation of 
profound crevasses. А strong presumption is thus created in 
favor of the glacial mills as a cause of the potholes of Cohasset. 
One difficulty, however, still remains. As the ice-sheet moves 
continuously forward, carrying the crevasses and moulins with 
it, how is it that the potholes escape elongation in the direction of 
the movement? Of course a moulin cannot move forward 
indefinitely, but only until a new crevasse is formed behind 
the one through which the water first fell, a few feet or yards at 
the most ; and the moulin then returns to its starting point. But 
why do the potholes not show even this small amount of elonga- 
o 
tion ? 
Mr. Bouvé has discussed this point in the following words! : 
** [t has, indeed, been thought strange that, as the ice moved 
continuously on, the holes were not found generally elongated 
1 Proc. B. S. N. H., XXIV, 224. 
