STREAM CAPTURE 



13 





/ 



V// 







\\ 



\ 



tv ^ 



3Kf 



<.^^ 





\\\\^' 



working backwards or sideways through the divide succeeds in reoccupjnng it. 

 In such case it would be more appropriate to speak of the former stream as dying 

 and being succeeded by tlie latter, rather than to speak of one stream capturing 

 another. 



We ought to keep clearly in mind in considering the phenomena of the case the 

 fact that the visible rivers and streams which are actively engaged in the work of 

 surface erosion represent but a remnant or residuum of the rainfall after a part has 

 been evaporated and a large part has passed beneath the surface of the ground to 

 find its way in subterranean channels. 



In Michigan, where there is a rainfall of about 3 feet, not more than a foot can 

 be accounted for in the surface runoff. Hence it is that an inci'ease in rainfall 

 causes the visible runoff to in- 

 crease much more rapidly, while 

 this visible runoff may disappear 

 entirely when the rainfall be- 

 comes only low. According as 

 the stream bed strata are more 

 or less extensively porous will 

 the visible runoff be more or 

 less conspicuous. Thus I have 

 often observed streams appear 

 in considerable volume where 

 they had cut down to or nearly 

 to impervious bed rock, which 

 above and again below that 

 point were lost in the sands and 

 gravels of their own courses, and 

 utterly inefficient in erosive ac- 

 tion. 



The first time that I saw the 

 connection between subsurface 

 drainage and 'stream or valley 

 capture was in a stream simi- 

 larly situated to the Donglass- 

 Houghton, and as the same phe- 

 nomena are more likely to be 

 visited and studied there by the 

 geologist since it is classical for 

 other reasons, I will use it as an 

 illustration. Figure 1 shows in 

 a sketchy way, with the help of 



Bulletin 23 of the United States Geological Survey, the contours of a part of the 

 stream, each row of hachures corresponding to an interval of 20 feet. 



The Douglass-Houghton comes down the trap range in a shallow, relatively broad 

 valley which has cut in but 10 to 20 feet. Just before reaching the edge of the trap 

 range, which at this point has overridden the "eastern sandstone," which else- 

 where is deposited unconformably on it,* it plunges by a picturesque waterfall into 







— III, - \J; ■■ .J!-. =l\ 



^■^^Wl.V ^11'"'% 



f'^'M\f 





^ "5= « -^ 



^^^^^^#1% 



■^ W^^ 









--3 Hi"*= " * v; 



/. 



Figure 1. — Douglass-Houghton Ravine, Keeweenaw Point. 



Showing how it has tapped by springs at ^ a shallower 

 ravine to the south. The line A-B represents the contact 

 line of the impervious traps with the porous eastern sand- 

 stone. 



*See vol. vi, part ii, Geological Survey of Michigan, by L. L. Hubbard. In press. 



