172 The American Geologist. Marcii, 1893 
4. Beds of subaeriallj' deposited glacial gravel and sand, with 
till above them. These oan seldom be shown to have sufficient 
extent to prove an interglacial epoch, l)ut are corroborative when 
occurring with other evidences. 
5. Differential weathering, which is dependent on length of 
time and warmth of climate. If a former land surface of glacial 
drift deeply weathered has a great northward extent under a later 
sheet of till, it is acceptable pi'oof of an interglacial epoch; but 
the preservation of such a soil layer would be fragmentary and 
local. On many tracts the erosion by the later ice invasion 
wholly removed the interglacial soil and stratified deposits, with 
their fossils and forest bed. 
6. Different amounts of subaerial erosion on areas of the older 
and newer drift. On some tracts of the former it appears to be 
ten times more than of the latter. 
7. Valleys excavated between successive depositions of till. 
These are especially important if eroded in rock, and most so 
when far inside the limits of the newer drift. 
8. Successive different directions of the ice movement on the 
same ground, as shown bj' striae and transportation of drift. 
Within any single epoch of glaciation gradual changes in direction 
of the glacial currents were produced by variations in the extent 
and thickness of the ice and especially by its unequal and irregu- 
lar melting during its stages of retreat; but where the changes 
were abrupt the}- indicate distinct epochs. 
9. Varying altitudes and slopes of the land. If a tract like 
the basin of the Mississippi river was at one time in the Ice age 
low and afterward much higher, it would imply a long interven- 
ing time; and if the record of this change as shown b}' the glacial 
deposits seems sudden, this would probably be due to an inter- 
glacial epoch. 
10. Vigor and sluggishness of ice action. During the time of 
maximum extent of the glaciation in the Mississippi basin the 
action was sluggish, but afterward was vigorous when the mo- 
raines of retrocession were accumulated. 
Some of these criteria may be sufficient singly to prove distinct 
glacial epochs; but when several or nearly all of them furnish 
concurrent testimony of such epochs, they seem equivalent to a 
mathematical demonstration. 
In the discussion of this paper, Prof. G. Frederick Wright 
