388 The American Geologist. Jane, 1894 
the Oriskany valley. When it had retreated to the Mohawk, 
or at least far enough to open an outlet in that direction at a 
lower level than the eol at Hamilton, the direction of flow of 
the lake water was reversed : and as the ice abandoned the 
mouth of the valley, the lake was drained, and the flow of the 
water which fell in the valley resumed its ancient direction. 
The Oriskany, shrunken to its present size, began work upon 
the drift deposits in the valley bottom, and gradually con- 
structed from their re-arranged material the low-level flood- 
plain that now borders the stream. 
Returning once more to the kames, several points of interest 
occur to the observer. In the first place, why should they ap- 
pear chiefly on the sides of the valle} r , instead of being dis- 
tributed across its entire width? To this a number of an- 
swers may be given. It might be said, in the first place, that 
they do extend completely across the valley, but that they 
are concealed from view, as a rule, in the middle, by the flood- 
plain of the present stream. Their shape, however, shows 
that the mounds diminish in size toward the center of the 
valley ; so that, even if they do cover the entire valley bottom, 
they are largest at the sides. The same argument would ap- 
ply to the suggestion that they have been diminished through 
cutting away by the present stream. The appearance of oc- 
casional concavities in their outline shows that some cutting 
has been thus accomplished against them ; though the fact 
that the present stream is accumulating flood-plains of fine 
material along its course also testifies that it can not have 
been a very powerful eroding agent. But the fact that the 
actual sizes of the kames are greatest at the sides of the val- 
ley seems to show that the chief points of discharge of the 
subglacial streams were similarly situated. This may per- 
haps be best explained as follows: When the ice-sheet covered 
the country, it spread over hills and valleys alike; but when 
it melted away, being deepest ami thickest in the vallej^s, it 
would obviousl} r take longer to melt from them than from the 
neighboring hills. Hence the melting ice front would have 
an irregular outline, with tongues of ice stretching down the 
vallej^s. Many of the escaping streams of water, therefore, 
would come forth from the ice at the sides of those tongues, 
and, flowing along between the tongue and the adjacent hill, 
