Ag^e of the St. Croix Dalles — JJpham. 351 
the elbow of the St. Croix and the mouth of the Sunrise river. 
On the east, between that low tract and the St. Croix valley, a belt 
of rolling and hilly glacial drift or till, underlain in part by the 
bed-rocks at a greater altitude than the sand and gravel area west- 
ward, divides it from this valley. * * * * 
About a sixth part of the St. Croix basin, lying east and south 
of Taylor's Falls, appears to have been drained during the Tertiary 
era by a stream coinciding nearly with the Apple river and the 
lower thirty miles of the St. Croix river. The large basin and river 
first described may be called the preglacial St. Croix, and the lower 
small stream may be distinguished as the enlarged preglacial Apple 
river. 
These Tertiary drainage areas, which by the vicissitudes of the 
Ice age became united into one stream, the present St. Croix, I 
think to have been divided, up to the time of the ice accumulation 
in the Glacial period, by a watershed of the very old trappean and 
Cambrian rocks, extending from northeast to southwest across 
the sites of the town of St. Croix Falls and Taylor's Falls. 
3. Berkey ascribes the erosion of the river gorge at 
the Dalles to Late Glacial and Postglacial time, and thinks 
that the larger part of the erosion was accomplished at 
the immediate close of the Glacial period, during the time 
the river served as the overflow channel for the West 
Superior glacial lakes. 
Nearly the same view is presented also by Mr. Cham- 
berlin, who, reasoning from the bights of the sand and 
gravel terraces, modified drift deposits, in this valley at 
Taylor's Falls and northward, concludes that the cutting 
of the deep, but short, gorge in the Keweenawan has all 
been accomplished by the St. Croix since the retreat of the 
last ice-sheet. 
Again I may most clearly and definitely give my 
reasons for a dififerent view by quoting from the same 
paper, Pleistocene Ice and River Erosion in the St. Croix 
Valley of Minnesota and Wisconsin, read before the Geolo- 
gical Society of America five years ago. 
In the Twenty-third Annual Report of the Minnesota Geological 
Survey, for the year 1894, I have stated (on pages 188-190) the evi- 
dence that the recession of the ice-sheet during the Buchanan inter- 
glacial stage, which succeeded its Kansan stage of maximum area 
west of the Mississippi, .extended northward beyond the site of 
Barnesville, Minnesota, on the southern part of the great valley 
