DRIFT SHEETS OF NORTHWESTERN ILLINOIS 281 



narrowness of the gorge, the rapids of the stream, and the occurrence 

 of the Illinoian gumbotil horizon essentially to the valley wall on 

 both sides (see Fig. 2) and high above the bottom of the gorge. If the 

 gumbotil was developed on a plain under conditions of poor drainage, 

 as set forth by Dr. G. F. Kay and J. N. Pearce,' the gorge could 

 not have existed during Sangamon times. The diversion of the 

 waters to this course seems to have taken place later and to be refer- 

 able to the invasion of the Green River lobe. If so, the Green River 

 ice must have reached the west side of the present Rock River 

 Valley in northeastern Rock Island County. The loess is so thick 

 here as to obscure all of the underlying deposits, so that there is no 

 opportunity to trace the margin, but relatively fresh drift is exposed 

 in patches on the east side of Rock River Valley. The cutting of 

 the Cordova gorge must have been accompHshed before the ice 

 receded, which conclusion is consistent with the finding of small 

 remnants of Late Wisconsin Valley train in the gorge. 



' "The Origin of Gumbotil;" Journal of Geology, XXVIII (1920), 89-125. 



