1 84 G. H. SQUIER 



for disputing the competency of running water, landslides or 

 creep to give rise to such deposits, it cannot be denied that they 

 lie within the horizon where such agents are operative. It is, 

 therefore, very desirable to trace the deposits into higher levels.. 

 But the middle portions in all the valleys are so deeply covered 

 by the loess — which has a strong tendency to fill up depressions 

 and obliterate minor irregularities of surface — that nothing can 

 be seen save in the rare cases in which gullies are deep enough 

 to slightly expose the structure, and even these, although afford- 

 ing valuable evidence as to the sequence of the deposits, etc.,. 

 give no decisive indication as to their character. 



On the outlying secondary hills the loess is not so thick, and 

 a systematic study of these has given unexpectedly interesting 

 results. The accumulations there, unlike those of lower levels. 

 in which all the different local formations are represented, are 

 composed almost exclusively of limestone from the tops of the 



glacial hypothesis. But as shown above, while on one side the deposit may rest on a 

 shelf, on the other it may lie in the axis of a valley, where if circumstances are favor- 

 able it may be covered by late deposits. This asymmetry is, however, quite as serious 

 an objection to torrential action or to the flow of semiliquid material, neither of which 

 could form deposits at a notably higher level on one side of a valley than on the 

 other. At c the bowlder bed rises into a sharp ridge three or four feet high as 

 shown in Sec 6, Fig. 2 (previous article). From c to d there is a nearly uniform east- 

 ward slope, broken only by the present drainage channel. Had there been nothing in 

 the configuration of the ridge to confine the drainage it must have left the bowlder 

 bed near d, to enter the lower level extending southeastward from thence. The dotted 

 half circle drawn through points of approximately equal elevation on either side of the 

 old drainage channel will show how widely the bowlder deposit departs from the 

 normal plane of a water borne deposit. 



There is an implied suggestion in Professor Chamberlin's note prefacing my 

 first article, which requires more specific notice than I have yet given it. It is 

 whether these deposits might not have resulted from landslides or from the lavalike 

 flow of saturated earth during thaws, or the more gradual creep due to repeated 

 thawing and freezing. Perhaps the best answer will be to state the conditions pre- 

 sented by a single case. The deposit shown in Sec. 3 (Fig. 2) of last article lies on a 

 rock shelf and reaches to about forty feet above low water of the Mississippi, a height 

 which is not reached in the valley for several hundred feet back. A 2° slope would 

 not intersect the bottom of the valley at a distance much, if any, short of 1500 feet. 

 To reach its present position from any possible source the material would have had to 

 travel about 3000 feet and make a sharp bend in its course. Its probable source was 

 one of the bluffs shown in Figs. 4 and 5, present article. Limestone is abundant in 

 the deposit, and some of the fragments quite large. 





