Eskers of (he Kansaii Epoch. — Heyshey. 239 
tiuie, if warped at all, tilted toward tiie northeast, instead ()!' 
toward the southwest, as the phenomena of the gravel knolls 
would require. 
Even if we were to consider it not full}^ proved that the 
material of the knolls has been carried to a greater hight than 
its original position, we have another way of determining that 
these streams eroded material from a lower level than their 
outlets. The ice occupjdng and completel}'^ filling the lower 
portion of the Pecatonica basin compelled the drainage to 
pass over the divide which separates it from the basins of 
Plum and Apple rivers. We cannot be mistaken in the direc- 
tion of flow in the glacial streams, for it is obvious that this 
direction coincided with the trend of the esker ridges and 
belts, and it certainly was not against the direction of ice- 
movement or into tlie body of the glacier. The westward 
flow is further indicated by the- manner in which the second- 
ary belts join the main belts, in the prevailing westerly' dip of 
the strata, and in the man}'' cases where the origin of the 
coarse material can be traced to the east. Now the lowest 
point, at present known, on the above mentioned divide, is at 
least 125 feet higher than a point in the Pecatonica valle}' 
where a stream eroded material by the action of a strong cur- 
rent. Consequently, unless our knowledge of hydrostatics is 
entirely at fault, only two hypotheses can be advanced in ex- 
planation of the phenomena, namely, that the district was 
tilted strongl}"" toward the west and southwest, or that the 
streams resulting from the melting of the ice "flowed up hill." 
As all the evidence negatives the idea of a tilting of the re- 
gion, the latter hypothesis must be adopted, which,, of course, 
implies that the stream channels were covered by ice — were, 
in fact, tunnels — and that the powerful force displayed by 
the currents which flowed in them was due to hydrostatic 
pressure. (This theory of the formation of eskers in other 
regions has long been before the geologic public, so that I am 
not taking new ground in applying it t{» the esker phenomena 
of northwestern Illinois.) 
It is supposed that the surface of the ice-sheet, under the 
influence of a milder climate than that which formerl}'' obtained 
in this region, melted freely to a distance perhaps fifty miles 
back from the margin. The streams thus formed flowed on the 
