GLACIAL DAMS, LAKES, AND WATERFALLS. 375 



ated with reference to the glacial boundary that its mouth 

 was for a rime obstructed by the ice ; but, when the ice-front 

 had withdrawn two or three miles, drainage would again be 

 opened into the Scioto, and at a level which is a hundred 

 feet or more lower than the surface of the moraine between 

 Paint and Brush Creeks, so that the natural necessity for a 

 glacial outlet down Brush Creek would exist only so long as 

 the ice closed up the mouth of Paint Creek. From this de- 

 scription of the situation it becomes evident that, so soon as 

 the ice should have retreated from the Kentucky hills south 

 of Cincinnati, so as to raise the blockade and reopen the chan- 

 nel of the Ohio, it would doubtless also have retreated from 

 the mouth of Paint Creek ; and the line of glacial drainage 

 through that into the Scioto, and thence down the reopened 

 Ohio, would have been re-established. 



A theory which so naturally accounts for so complicated 

 a set of facts as these is well-nigh proved by this single in- 

 stance. The only competing hypothesis possible is that of a 

 general subsidence of the country producing the same water- 

 level above Cincinnati which the ice-dam is supposed to have 

 done. But such a theory lacks the positive evidence addu- 

 cible for a glacial dam at Cincinnati ; and, besides, it is not 

 probable that a general depression of the country, such as 

 would produce still water at the head of Brush Creek, would 

 be of such short duration as is implied by the facts connected 

 wnii this deposit. The gradual lowering of a barrier holding 

 the water at that height would have caused numerous benches 

 on the interior of the deposit toward the ice, whereas the 

 terrace on that side is even more abrupt than on the other. 



A second class of facts supporting the theory of the Cin- 

 cinnati ice-dam is drawn from the high-level terraces found in 

 the trough of the upper Ohio and its main tributaries. One of 

 the most significant of these occurs at Bellevue, on the north 

 side of the Ohio River, five miles below Pittsburg, where a 

 terrace is found about a mile long, half a mile wide, and fifty 

 or sixty feet in depth, preserved upon a shelf of rock facing 

 the river perpendicularly, and between two hundred and 



