DELAWARE COUNTY. 303 
the townships of Thompson, Radnor, and the northern part of Scioto, 
has the features of the flat tract in north-western Ohio known as the 
Black Swamp. The banks of the Scioto are low (ten or fifteen feet), and 
consist of Drift, the rock rarely being known in its bed. The Drift ap- 
pears fresher and the surface is smoother than in the rest of the county. 
A short distance above Millville the banks begin to be rocky, the exca- 
vation beginning in the Waterlime, over which it has been running 
since it left the western part of Hardin county, but without making the 
slightest excavation, rarely revealing it in its bed by rapids. Within a 
mile from Millville the amount of erosion in the underlying rock in- 
creases to a remarkable extent, and at Sulphur Spring Station, about two 
miles below Millville, the erosion in the rock amounts to sixty or seventy 
feet. From there southward the rest of the Scioto valley is between 
high, rocky banks. This exemption from erosion in the upper waters of 
the Scioto can not be due to the harder nature of the rock there, because 
the Waterlime is much more rapidly worn out under such agencies than 
the Lower Corniferous, en which it enters at Sulphur Spring Station. 
The composition of the Drift about the headwaters of the Scioto is the 
game as about the lower portions of its course. It is in both cases a 
hard-pan deposit, made up of a mixture of gravel-stones, bowlders, and 
clay, rarely showing stratification or assortment—such a deposit as is, 
without much difference of opinion, attributed to the direct agency of 
glacier ice. The conclusion is inevitable that the lower portion of the 
Scioto has been at work digging its channel in the reck much longer 
than the upper portion. The slope is in both cases toward the south, at 
least that portion of it in Delaware county; and that agency, whatever 
it was, which served to make this change in the valley of the Scioto 
‘from no excavation to deep rock erosion, could not have been quiet, 
standing waters over one portion of the valley and not over the other, 
since such waters would have retired last from the lower part of the val- 
ley, and we should there expect less instead of more erosion. The only 
possible way to explain this phenomenon, in the light of plausible theo- 
ries, is to refer it to the operation of the last glacial epoch, or to the op- 
eration of a glacial epoch which projected the ice field only so far south 
as to cover the upper part of the Scioto valley, leaving the lower portion 
of the valley, which probably pre-existed, to serve as a drainage channel 
from the ice itself. Subsequently, when the ice withdrew, the upper 
tributaries were located in such places as the contour of the surface 
allowed or demanded. 
There are other evidences that the townships of Radnor, Thompson, and 
the northern part of Scioto were for a time under glacial ice, while the 
b | 
