THE PREGLACIAL DRAINAGE OF OHIO. 7S 



Kanawha was backed up into a lake which rose until it over- 

 flowed the col at (B). Joining Licking again, the two followed 

 their last channel as far as North Bend and probably out past 

 Cleves ; but there is a possibility that before the col at (B) was 

 removed the ice had advanced far enough to reach the hill below 

 North Bend and obstruct that outlet. In this case the new lake 

 would have included Licking as well as Kanawha, and had to rise 

 to the level of the col at (D) before it could have begun to 

 drain off. If, however, the col at (B) was worn down in time 

 for the water above it to escape past Cleves, then, when the ice 

 had advanced across the valley below Cleves, a fifth lake covered 

 the upper Ohio valley before the col at (D) was eroded and 

 the present drainage to the mouth of the Great Miami establishd. 

 It is possible there was still a sixth lake, though if so it was of 

 less extent and shorter duration than any of the others, and 

 was due to a projection or loop of the glacier pushing out of 

 Miami valley as a dam to the new Ohio — which name is now 

 applicable to the river for the first time — until its waters broke 

 through a ravine back of Petersburg, Kentucky. The depres- 

 sion thus formed is usually spoken of as "an abandoned chan- 

 nel of the Ohio," but it was occupied only while torrents from 

 melting ice were far above existing flood plains. It furnishes 

 about the only evidence, by the way, that the glacier ever 

 reached the Kentucky hills. 



The theory advanced here in regard to the succession of 

 glacial lakes is based entirely on the assumption that the col at 

 Madison (E) was broken down sufficiently to drain the first one 

 formed, and upon the further assumption that the ice reached 

 each necessary point for the formation of a lake, in the order 

 here given. There seems to be no doubt regarding the first 

 and most extensive one ; the existence of the others depends 

 upon the strength of the col at (E) and the relative periods of 

 time at which the subsidiary streams were blocked. It is not 

 necessary to presume a constant forward motion of the glacier ; 

 its advance may have been frequently interrupted, or there may 

 even have been an occasional recession without in the least 

 invalidating the argument. The effect would be the same in 

 the end, whether there was a continuous progressive motion, or 



