404 The American Geologist. .Jane, 1894 
beneath a till sheet not unlike that which forms the great 
ground moraine. 
(c) Peculiarities of Distribution. — It has been stated* that 
the distribution of drumlins does not favor this hypothesis 
but, so far as I can ascertain, no condusive reasons have been 
put forward. Certain reasons have been suggested in an 
earlier part of this paper why the author draws exactly the 
opposite conclusion. These reasons need not be more fully 
referred to here; but it may be remarked that, if the distribu- 
tion of drumlins is opposed to the theory under discussion, a 
statement of the reasons is called for. 
5. — Restatement of Destructional Hypothesis. 
There seems to the author to be a sufficient body of fact, 
and a sufficient number of indications favoring the destruc- 
tional hypothesis, to call for a reconsideration of it. Taken 
in connection with the form of rock hills these indications 
form an argument which must at least be answered. It is 
certain that glacial erosion en n />ro<t nee it m mlinoid J'ornis \ is 
there any other kind of ice work of which as much can be 
said? 
Grant a deposit of drift, a moraine let us say of variable 
thickness and composition, overridden by the ice during a 
re-advance; in what manner would its removal be accom- 
plished? and would there be any difference between the forms 
produced from a drift sheet and those resulting from the ero- 
sion of an irregular but uniform rock mass? 
The first question which arises is, would ice erode equally 
in all parts of such a till mass? or would not some point be 
eroded more rapidly than others, either by reason of softness, 
or original form, or abundance of erosive material in the 
grasp of the ice? Once an irregularity of work is begun, 
would it not continue? and would not the resulting form be 
the same as that which is produced in a rock of uniform tex- 
ture? In the latter case the forms produced are gently irreg- 
ular lenticular slopes of least resistance, rounded rock hills 
whose size and details of outline depend upon the swing 
of the erosion curve. The same form is present in drumlins. 
and it is a fair question to ask if it is not probable that the 
same cause has operated. 
*R. D. Salisbury, Annual Report, New Jersey Geol. Survey, 1891, p. 73. 
