EVIDENCE FROM MARGINAL BELT WEST OF NEW ENGLAND 315 



perched high above the existing streams. The rivers known to have been 

 affected include the Kanawha, Allegheny, Monongahela, Youghiogheny, 

 Kiskiminitas, Guyandot, Big Sandy, Kentucky, and the Ohio itself. 



The most numerous cases of diversion are found along the north- 

 flowing streams (see plate 14). 



Campbell explained the stream diversions by postulating many, tempo- 

 rar}^ ice-dams. Summarizing his view, he wrote : 



"The lower courses of the Allegheny, Monongahela, Kanawha, Guyandot, 

 Big Sandy, and Kentucky rivers are characterized by abandoned channels, 

 which generally range from 100 to 200 feet above the present streams. Gen- 

 erally these channels are deeply covered with silt, but sometimes the rock-floor 

 is only partially obscured by a thin layer of sand and gravel. The streams 

 which have forsaken these vallej's have sought new routes, along which they 

 have carved deep channels through the upland topography. Teay Valley, in 

 West Virginia, is perhaps the most noted example, but the old channels at 

 Carmichael and Masontown, on the Monongahela River, and opposite Parker, 

 on the Allegheny River, are also well known. 



"No reason has been assigned for the abandonment of these channels; they 

 can not be considered as 'ox-bows,' and they are all beyond the limit of glacial 

 ice. The present hypothesis seeks to explain them through the breaking up of 

 river ice and the formation of local ice-dams which were of sufficient height to 

 force the water over the lowest divide in the rim of the basin and which per- 

 sisted long enough for the stream to intrench itself in its new position." " 



CampbelFs explanation would be more acceptable if any existing river 

 were known to have dammed itself by ice for any of the long periods of 

 time demanded by his hypothesis. On the other hand, diversion through 

 excessive alluviation is well illustrated by existing streams; Shaw and 

 Munn prefer this cause for the abandonment of valleys in the Ohio 

 basin. Their statement for the type cases observed along the lower 

 Allegheny may be quoted : 



"The high terraces [remnants of the floors of ancient valleys that had steep 

 sides but broad flat bottoms] developed as a unit through the overloading of 

 the Allegheny in early glacial time and the later redissection of its deposits. 

 The overloaded condition of the Allegheny was probably due to several causes, 

 among which the following may be mentioned as being more or less effective : 

 First, an actual increase in load derived from (a) material fed more or less 

 directly by the glaciers ; ( & ) debris from the cutting of new gorges across old 

 divides; (c) material brought after the ice melted by streams in the glaciated 

 area as they cut new valleys. Second, a decrease in velocity and carrying 

 power, produced by (a) the attraction of the ice mass . . . (&) crustal de- 

 formation, due to the weight of the ice; (c) the crossing of divides or ice 

 barriers, each of which would check the velocity and cause deposits for a short 



" M. R. Campbell : Bull. Geol. Soc. Am., vol. 12, 1901, p. 462. See folios 69, 72, 82, 

 04 of the U. S, Geol. Survey for Campbell's fuller statement. 



