338 
will probably be found to be the ancient course of the Mississippi river. 
That it is the position of a preglacial valley is indicated by a deep well 
at the St. Paul Harvester Works, situated in the present topographical de- 
pression, which penetrated rock at 285 feet beneath the surface or 628 feet 
above the sea, which is 55 feet beneath the present low-water level of the 
Mississippi river at St. Paul. The lake Phalen depression is separated 
from the head of the Mississippi canon valley by a moraine which is evi- 
dently based on a comparatively low surface, for it does not rise nearly as 
high as the drift to the east or west. As seen from the opposite side of 
the valley, its escarpment or bluff at the head of the old canon valley shows 
such topography as is usually produced by the erosion of drift. In short, 
all the evidence favors this lake Phalen depression as the position of the 
pre-glacial continuation of the Mississippi cafon valley.” (46: 263.) 
From the southeastern corner of St. Paul to Leclair, Hershey believes 
with other geologists that the valley is pre-glacial. In the vicinity of Du- 
buque, however, he thinks that the valley is proportionately too small for 
the stream which it carries, that the preglacial stream flowing past Du- 
buque could not have been larger than the present Rock river, or possibly 
no larger than the Pecatonica. The valley is canon shaped and narrow and 
the rock floor is about 300 feet below a deep filling of drift. The divide 
is suggested to be somewhere between La Crosse and Prairie du Chien, par- 
ticularly where “military ridge’ is traversed by the present river. 
(46 :266.) 
Hershey believes that the stream north of this ‘supposed divide flowed 
toward central Minnesota instead of away from it, but that the reversal 
came early, before the Ice Age, probably at the end of the Ozarkian, by an 
uplift in the north, or, as an alternative view, it may have “resulted from 
the disturbance of other drainage systems by the accumulating northern 
ice. For instance, it is quite possible that the Kansan ice-sheet had ad- 
vanced across the cutlet of the supposed northwardly flowing ancestor of 
the upper Mississippi river, obstructing its flowage, and after the produc- 
tion of a great extra-glacial lake, turning the drainage of the entire region 
over the lowest point on the divide which intervened between it and the 
headwaters of the southwardly flowing central Mississippi river, long before 
it glaciated the country south of the ‘driftless area.” (46 :267.) 
