Conclusion 49 



recession of Saint Anthony falls proper is a continuation of the former on 

 the Mississippi above the confluence. 



During the entire recession two events are presumed necessarily to have 

 influenced the volume of the rivers : One was the melting away of the 

 glacier from the head of the Mississippi (see figure 1), presumably caus- 

 ing a marked reduction in volume of the river from its Glacial to its 

 modern stage. The other was the breaking of the ice-barrier and open- 

 ing of drainage to the northward, by which lake Agassiz was drawn off 

 and the river Warren reduced to the Minnesota river stage. 



Regarding the reduction of the Mississippi river, a double system of 

 terraces and scarps is evident above the walls of the Saint Anthony gorge 

 in several places.^* These belong undoubtedly to the river at stages pre- 

 ceding the making of the gorge. They are interrupted by the widening 

 of the gorge, and thus are not continuously traceable, but yet appear to 

 represent two distinct stages of the river. The lower of the two lies dis- 

 tinctly within the upper, so that a person may readily interpret them as 

 representing respectively the modern stage and the Glacial stage of the 

 Mississippi. The higher or Glacial Mississippi terrace extends from 

 Nicollet island to and continuous with the terrace and scarp of the river 

 Warren, as shown in plate 1. The lower terrace coincides with the 

 present river channel above the falls and is evident for 6 miles, or to the 

 mouth of Minnehaha creek. I have not distinguished it beyond the Sol- 

 dier ravine. The change in the Mississippi river consequent to the melt- 

 ing of the glacier from its head may be referred to the stage of the river 

 to which the making of the Soldier gorge belonged — that is, when the 

 falls were li/^ miles above Fort Snelling. 



The reduction of the river Warren may also be correlated with a stage 

 of the recession of Saint Anthony falls and the making of the gorge. 

 The rock gorge of the Mississippi at Fort Snelling is 75 feet deep below 

 the level of the river (figure 1, plate 2), and the valley thence to Saint 

 Paul is the same. The falls did not excavate to that depth, as is shown 

 by the evidence seen in terraces and also by comparison with the gorge 

 above Lake Street bridge at lock and dam number 2, where the river was 

 found to run on the sandrock bottom of the gorge. The bottom of the 

 gorge at Fort Snelling and thence to Saint Paul is to be considered as a 

 refilled valley. The time of the maximum depth may well be considered 

 as that of the last days of the river Warren. The time of maximum 

 depth of the gorge should also correlate with the time of greatest height 



" A still higher and very distinct terrace runs parallel to the east side of the gorge, 

 but It belongs to the time of glacial occupation preceding the forming of the Mississippi 

 river proper, and Is not considered here. 



