LOESS AND THE IOWAN DRIFT. 363 



undermined some of the loess, washed the shells out, and 

 mingled them finally with sand brought from a higher drift, 

 the cavity finally being filled with such sand and loosened 

 loess. 



An additional possibility is also suggested by the fact that 

 this region lay in the path of the Illinoisan glacier, and that 

 much of the frozen loess was probably displaced, and in its 

 basal parts mingled with the underlying drift. As the Illi- 

 noisan drift lies over this loess, there is no question that the 

 ice-sheet passed over it. 



The occurrence of fossils in the sub-loessial gravelly belt 

 might make the reference of this belt to loess appear more 

 plausible, but would not absolutely prove its correctness. Such 

 cases are exceedingly rare, and may be due wholly to local dis- 

 turbances. Occasionally gullies are cut through loess into the 

 underlying drift. In such cases fossils are sometimes washed 

 out of the loess, and mingle at the bottom of the gully with 

 sand and gravel washed from the drift. If such a gully should 

 again be filled by slumping, and perhaps further covered by 

 subsequent deposition of loess, a misleading mixture of fos- 

 siliferous loess and drift would result. Some such local cause 

 may account for the very rare cases in which fossiliferous mix- 

 tures of the kind have been observed. In any event the con- 

 ditions here discussed are very unusual, and as proof of the 

 intergradation of loess and drift, are certainly of little value. 

 It should also be noted that McGee placed more or less em- 

 phasis on the occurrence of loess-kindchen and iron-tubules in 

 his gravelly member of the loess series* and it is probable that 

 this also influenced him in referring this layer to loess, but it 

 is not uncommon to find lime-nodules and iron-tubules in both 

 gumbo and undoubted drift, and their presence offers no ab- 

 solute measure of the identity of the deposit. 



In his communication to the American Geologist, Leverett 

 positively states that "the part in eastern Iowa has been c'early 



*See 1. c, pp. 465, 492, etc. 



