:j 1 !S The American Geologist, November, 1898 
n drainage basin of less than a fourth of n square mile, or 
one 17-") feet wide from four square miles of land surface. 
The table shows that the constricted valleys have bottoms 
varying from five to thirty times as wide as the present streams. 
An examination of a score or more has shown the average 
ratio to be not less than 10 : 1 . 
Widening of valleys by the undermining and general degra- 
dation of the walls is relatively a much slower process than is 
the first cutting of the gorge to the water level. Granting 
that the Pleistocene streams, with their supposed greater size 
and efficiency as compared with those which have existed in 
the region in postglacial times, would have been able to cut 
an outlet for the lakelets back of them in a period equal in 
length to that which lias been occupied in the erosion of the 
newer drift, the phenomena of the rock gorges would still seem 
to indicate at least ten times as long a period of erosion. 
Hut this difference is further emphasized when we come to 
compare the erosion of till in the region of the newer drift 
and the erosion of solid limestone rock in that of the earlier 
drift. The great majority of lakelets in the former region 
owe their continued existence to the want of time to enable 
their outlets to cut down through a barrier of till. The 
streams of northwestern Illinois have excavated their gorge- 
almost entirely in Galena limestone, a tolerably resistant ma- 
terial, as is shown by its frequent occurrence along streams in 
perpendicular bluffs. A comparison of the amount of erosion 
effected on the newer drift in the vicinity of lake Michigan 
and in the earlier drift region of northwestern Illinois has 
convinced me that there is not much difference in the amount 
of excavation, relative to the size of the streams, in till in one 
region and solid rock in the other. Now. it seems evident 
that the erosion of a given amount of Galena limestone would 
take at least ten times as long as for an equal amount of till. 
It has been determined to the satisfaction of most giacial- 
ists that the ice-sheet last retreated from the regions whose 
valleys have been selected for comparison, namely, around tin- 
head of lake .Michigan, the Kettle moraine in Wisconsin, and 
the vicinity of Niagara falls, about 6.000 to 10.000 years ago. 
with a probable average of 7.000 years. As the length of 
