l6o GEORGE D. HUBBARD 



the stream in attaining its former course. Farther back from 

 the moraine, even to the stream's sources, the drift is so deep 

 and the channel is so shallow that the absence of bed-rock in the 

 bottom of the valley proves nothing as to the position of the 

 earlier stream and its valley. South of the moraine the till is 

 much thinner, and a valley with such depths as those of the 

 Embarras would often, if not almost constantly, reach bed-rock, 

 were it not within an earlier valley. Hence there was, first, a 

 broad, well matured preglacial valley similar to those at present 

 in nearby carboniferous rock in unglaciated parts of Indiana, 

 Kentucky, and southern Illinois. Second, this valley was nearly 

 filled with drift of Illinoian age. Third, after the retreat of the 

 ice of the Illinoian stage and before the Wisconsin invasion, a 

 valley was excavated in the above-mentioned filling. This 

 probably extended some distance north of the present terminal 

 moraine, possibly as far as the present sources of the Embarras. 

 Fourth, this interglacial valley was partly filled with overwash 

 material (entirely filled, north of the moraine, with drift) of Wis- 

 consin age. And, fifth, the present youthful postglacial channel 

 has been cut in the filling of the interglacial gorge, but the process 

 has not gone far enough to reach bed-rock where the stream is 



fairly within the earlier valleys. 



George D. Hubbard. 



Cornell University, 

 Ithaca, N. Y. 



