488 PROCEEDINGS OF THE ALBANY MEETING 



taneously. We have stated a general rule. When either stream is without floods 

 •the channel may not appear. 



One of the first examples to attract the writer's attention was in the high terrace 

 between Swan Lake creek and the Missouri river, Walworth county, South Dakota. 

 Its formation dated back to the Glacial period.^ 



A good recent example is shown near Fort Snelling, at the junction of the Min- 

 nesota and the Mississippi. f 



Other ancient examples are found crossing the high terraces above the junctions 

 of Loup fork and Shell creek with the Platte, in eastern Nebraska. t 



Also the junction of the Kansas and ^lissouri rivers, at Kansas City. The former 

 stream now occupies the ancient flood-relief channel. 



Of recent examples, an interesting one is found at the junction of the Big Fork 

 or Bow String with Rainy river, in northern Minnesota ; and in Missouri, the junc- 

 tion of the Grand and Missouri rivers, near Brunswick; the Osage and the Mis- 

 souri, above Dodd island ; the Loutre and the Missouri, opposite Herman ; the 

 Missouri and Mississip)>i, above Mobile island. Other cases nre less clear, as the 

 Chariton and the Missouri, the Nishnabotna, Tarkio, and Platte with the Missouri. 

 The later are less clear because the Missouri in its windings has cut into their 

 channels. Other exam})]es can doubtless be found without ditficulty. 



As a result of the foregoing we have the following corollaries : 



1. If eitlier stream eventually cuts lower than the other the higher is apt to take 

 the flood-relief channel for its mouth. This change will take place probably in 

 time of flood. In most cases the tributary will be the one taking it. For example, 

 the Kansas river before mentioned. On the other hand, the main stream may 

 sometimes occupy it by having its current thrown through it by the shifting of its 

 bends. This will rarely occur, but seems to have been the case at Vermilion, 

 South Dakota, when the main current of the Missouri came into the Vermilion west 

 of Kidders island about 1860, and again about 1874. 



2. It remains to consider whether flood relief channels may be so persistent as 

 to keep open during the erosion of deep valleys, and thereby produce a high island 

 detached from the promontory above the junction of two streams. We may 

 readily admit that the chances are against it, for the meanderings of the streams 

 about it are likely to destroy such an island ; but if there be a very durable rock or 

 rapid currents in the streams, such a result seems conceivable. 



Cote Sans Dessein, which is a rocky ridge rising 100 feet above low water in the 

 midst of the valley of the Missouri near its junction with the Osage, may be an 

 example of this sort. It doubtless is a part of the ancient divide between the 

 Missouri and Osage rivers. Its position is favorable to such an explanation ; but 

 perhaps it is more likely to be the result of one stream cutting through more 

 recently into the valley of the other, as has been suggested by Mr C. F. Marbut.^ 

 It is evident that the Missouri has encroached upon that side of the valley, as he 

 very clearly shows. 



Velocity checked by Overflow, though Slope is increased 



A third point is that a flooded stream has its velocity checked l:)y overflowing its 



* See U. S. Geol. Survey Bulletin 144, p. 42, pi. xvi. Lower view by a typographical error. 

 fPart ii of the Report of the Minnesota River, 1874, p. G. 

 t See.Science, March 11, 1892, p. 148. 

 *See Amer. Geologist, vol. xxi, p. 86. 



