IOWA ACADEMY OF SCIENCES 
217 
On the hill-tops along the middle stage of the streams we occasionally 
find places where the sandy borders give place to bare, fiat-topped 
stretches where the bed-rock lies everywhere practically at the surface. 
An example of this may be found to the south of the rock cut of the 
railroad company at Fayette; another occurs on the hill-tops back of the 
large spring one-half mile north of Fayette; another, one-half mile to 
the northeast of the small cemetery on the West Union & Fayette road; 
and still another, one-half mile to the west of the cemetery at Dunham’s 
Grove, in Center township. 
There is no disintegrated limestone at the surface of these miniature 
plateaus, only the hard level surface scantily covered with soil. Occa- 
sionally small boulders may be found but they are by no means abundant. 
Neither can glacial scorings be found, though the absence of rock w^astc 
would seem to be evidence of glacial action. It is unlikely that the 
Kansan drift sheet could have planed off such surfaces, because we coulu 
not suppose that the fresh, unweathered Devonian limestone dated bacii 
to the same time as the highly oxidized and even decayed Kansan drift. 
We are then left to believe that such action was caused by the Iowan ice 
sheet, even though Iowan till and boulders are almost entirely absent. 
This second phase of the valleys is interesting as indicating that the 
glacier along its margin was deeply indented by the streams of the reg- 
ion. The length of the notches back into the Iowan drift varies in six 
good instances, found between West Union and Fayette, from a half mile 
in the case of the little streams three to four miles north of Fayette to 
two miles in the case of the Coulee, a stream southeast of Donnan Junc- 
tion, and at least six miles in the Volga Valley, or from a short distance 
above Fayette all the way to Maynard. To that extent were the streams 
able to keep their channels intact in spite of the Iowan ice sheet. The 
length of the second phase of each valley varies in proportion to the 
size of the stream. 
The third phase of these valleys begins where the sandy border lands 
give place to the higher and much more uneven Loess-Kansan. The val- 
leys here are much deeper than in the second stage, but this is due more 
to the loess accumulations on the hill-tops than to deeper trenching of 
the streams into the limestone. McGee has well described the paradoxi- 
cal appearance of the streams rising in the low-lying Iowan plains and 
flowing into the higher loess-mantled region, and this fact is everywhere 
well shown in this section. 
It seems very strange and unaccountable that in the second and third 
stages of these valleys as above described, there should be no gravel 
trains or terraces, but such is the fact. The only exception observed is 
along the Volga from Maynard northw’ard for some two and a half miles. 
There we find the valley pretty continuously lined with gravel deposits, 
but from here on for some five miles in a limestone channel through the 
Iowan drift, down to Fayette in the Loess-Kansan stage of the valley, no 
gravel trains or terraces were observed. The sandy outwash from the 
Iowan is so general just at the margin of the Loess-Kansan hills, and 
