Interglacial Chronometer. — Wi7icheIL 77 
9ume that it was driven there by the events and forces of the first 
glacial epoch, and that it had occupied it during the time that in- 
tervened between the two glacial epochs. But that is an excavated 
river gorge, cut in rock in a manner similar to the present gorge, 
except that it is wider. Our problem is to measure the time re- 
quired for such excavation, for we are obliged to infer that the 
manner of its excavation was by the recession of an interglacial 
falls of St. Anthony. 
In this problem there are some elements identical with those 
which were employed in the calculation of the time required for 
the formation of tJie post-glacial gorge, and there are some which 
are somewhat different. 
Those which are the same are: 
1 . The general slope of the country. 
2. The contact of the two formations which conspire to cause 
the water-fall. 
3. Uniform and horizontal position of the same rocks. 
4. Definite limit for the upper end of the eroded gorge. 
The elements which are different, or uncertain, are the follow- 
ing: 
1. The size of the river. 
2. The lower end of the excavated gorge, i. e. whether the 
falls began at that point above Fort Snelling where the Mississippi 
at that time met with the Minnesota river, or at a point at St. 
Paul opposite Dayton's bluff, where the lower right-angled turn 
takes place. 
As to the size of the river during interglacial time we have 
no definite data, except that we may infer that it was not so large 
but that it could flow in the gorge it excavated. We have no re- 
liable data by which to estimate the annual precipitation over the 
region which the river must have drained. It was a period in 
which some of the mountain ranges of the western country did not 
exist. The Rock}' mountains did not extract and return to the 
Pacific ocean the moisture from the western winds quite so readily 
as they do now, and hence the contrast in annual rainfall between 
the plains and the eastern portions of North America was not so 
marked. Again there must have been active volcanoes in the 
western portion of the United States and perhaps some in Colo- 
rado. These must have disturl^ed the atmospheric conditions, 
causing at least in their near vicinit}', copious rains, and these 
