296 A. G. LEONARD DRAINAGE CHANGES IN NORTH DAKOTA 



ice-sheet : "There, then, on that limit a river must have been formed to 

 carry away the melting water from the glacier, and this limit was the 

 Missouri River, and that was the river formed thereby." 2 



Todd, in 1884, stated his belief that the Missouri River formerly flowed 

 east or northeast, either to the present James or to the Mouse River. 3 

 In a later paper he elaborates this view and gives reasons for his theory 

 that the river .flowed northeast to the Mouse Valley. 4 Todd believes that 

 there is evidence that the Heart and Cannon Ball rivers once flowed on 

 east to the James River, and that the valley of Snake Creek is the pre- 

 glacial valley of the Missouri River. The valley of Snake Creek is no 

 larger than the valleys of other tributaries entering the Missouri above 

 this point, and the notch in the east front of the divide north of Fort 

 Stevenson, which is shown on some maps, does not exist in reality, and 

 there is no evidence of any preglacial valley here. The valley of Long 

 Lake Creek could hardly have been the eastward extension of the Cannon 

 Ball River Valley, since its lower course is not opposite the mouth of the 

 latter stream, but joins the Missouri four miles to the north. There is 

 no reason for believing that the Knife River ever joined the Missouri 

 near Fort Stevenson, and the lower valley of this stream has every ap- 

 pearance of great age, having a broad flood-plain and gentle slopes. It is 

 clearly a preglacial valley. The Heart River is thus the only important 

 tributary of the Missouri which might have continued eastward to the 

 James River if the valley of Apple Creek, which has its mouth just oppo- 

 site the Heart, is an indication of this. But Apple Creek is readily ac- 

 counted for as a preglacial tributary of the Missouri and one of the chief 

 outlets for the glacial waters when the ice-margin occupied the position 

 marked by the Altamont moraine. Its valley is largely filled with glacial 

 outwash from the moraine. 



There is abundant evidence that the Missouri Valley below the mouth 

 of Snake Creek is preglacial, and that the river was not forced by the ice- 

 sheet to take its present southerly course through North Dakota. This 

 evidence is based on the presence of glacial boulders on the valley bottom 

 and at many points on a terrace representing a former flood-plain of the 

 Missouri. Boulders have been encountered in two wells in Bismarck at 

 a depth of 125 feet below the surface or 80 feet below river level. These 

 wells are near the edge of the terrace bordering the Missouri Valley at 

 Bismarck, and since the boulders rest on the bedrock they indicate that 

 the valley was excavated to this depth prior to the Glacial period. In 



2 Annual Report, Chief of Engineers, U. S. Army, for 1S68, pp. 307-314. 



3 Proc. Am. Assoc. Adv. Sci., vol. 33, 1884, pp, 381-392. 

 * Science, vol. 39, 1914, pp. 265-274. 



