QUATERNARY AGE. 275 



void of Loess — as the correct explanation. One of these is that in 

 such sections the Loess that borders on to an exposed Drift region 

 is exceptionally full of the remains of elephants and mastodons. As 

 if these animals had come down to the water to drink and to wal- 

 low, and had become mired and perished. This is proposed, how- 

 ever, as only a provisional explanation. 



Another observation depended on by Richthofen to substantiate 

 "his theory is the depth at which root holes are found in the Loess. 

 He supposes these to occur at such a depth that the grasses that oc- 

 cur at the surface could not possibly have penetrated the Loess to 

 such a depth, and that therefore they must have flourished when 

 •this deposit was thinner than at present. Subaerial filling up would 

 account for their presence, as they would be growing during the 

 •whole period of the accumulation of the Loess. To this it may be 

 replied that roots descend from the surface through the Loess to an 

 -enormous depth. In 1S68 I measured the depth of a root of the 

 Buffalo berry (Shepherdia argophylla), at the edge of the St. John's 

 timber, in Dakota County, and found it to extend fifty-five feet be- 

 low the surface in undisturbed Loess. Near the same point, I 

 traced another root from near the bottom of the Loess in a slide for 

 thirty-nine feet to a stock of grass (Androftogon furcatus). West 

 of old Fort Calhoun the roots of the common blue-grass have pen- 

 etrated the Loess to a depth of from five to fifteen feet. A sumach 

 {Rhus glabra) near by was found to send dow r n roots to a depth of 

 fifteen feet. South of Plum Creek, in the Loess canyons, roots of 

 the lead plant {AmcH'pha canescens), can be traced in the Loess for 

 from ten to twenty feet. Prof. J. E. Todd has also observed in the 

 Iowa Loess the roots of other grasses to descend to depths of from 

 six to twenty-five feet.* Moreover, these root marks inosculate in 

 every direction, and become fewer the deeper we descend, with 

 some notable exceptions. There are horizons in the Republican 

 Valley, far below the present surface, where the old root marks oc- 

 cur in exceptional numbers. As these fossil root marks are now 

 more or less completely filled with either lime carbonate or oxides 

 of iron, they are readily distinguished. To understand the probable 

 reason for these phenomena, on the theory of the subaqueous origin 

 of the Loess, the following sections are given. The first is taken 

 from along the sides of a canyon leading into a tributary of the 



♦Proceedings of tin- American Association for tin- Advancement of Science, 

 Louis Meeting, August, i*7s. 



