Pleistocene to Pve-pleistocene of Mississippi Basin. 377 



terrace, if we correlate rightly, may be traced nearly to the 

 junction of this valley with that of the Mississippi. The 

 streams tributary to the Ohio from the north, have their sys- 

 tems of terraces corresponding to those of their main, and the 

 streams tributary to the Mississippi — so far as they head within 

 the area of second glacial drift — are likewise terrace-bordered. 

 At various points in many of these valleys, the terraces have 

 been removed by late erosion, so that the modern flood plain 

 extends from bluff to bluff. 



The Mississippi Yalley south of the latitude where the ter- 

 races occur, is known to have a filling of considerable depth 

 and of geologically recent age. This filling, as we interpret, 

 is in large part of second glacial origin, is in fact the south- 

 ward extension of the northern terraces, and constitutes the 

 only southern representative of the second glacial epoch. We 

 have never seen the material brought up by the borings in the 

 Mississippi flat, but borings in the Ohio flat, not far from the 

 Mississippi (Mound City, 111.), show sand of unmistakably 

 glacial origin. Here the filling is certainly second glacial. 

 The inference that the corresponding filling of the Mississippi 

 is of the same age, lies close at hand, quite apart from the 

 argument drawn from the southward decline of the Mississippi 

 terraces, and their disappearance beneath the flood plain.* 



* Since this article was written, the junior author has revisited some of the 

 localities referred to in the first paragraph of p. 364. and at a season especially fa- 

 vorable for observations bearing- on the question here raised. The lack of homo- 

 geneity in the loess is much more conspicuously shown when the face of the 

 section is wet. than when it is dry, and it was mostly in the latter condition that 

 the loess sections of this region had been seen heretofore. The new observations 

 seem to leave little doubt that the loess, outsrie the drift limit, is to be referred 

 to two episodes of deposition. The loess sections just east of Forrest City. Ark., 

 seem most conclusive. Here there are two loesses, separated by a thin belt of 

 humus-stained soil. This soil constitutes a sharply marked horizon throughout 

 considerable parts of the extensive railway cuts near the station. It is not 

 everywhere equally conspicuous, apparently containing varying amounts of organic 

 material ; but well-defined or ill-defined, it is a very persistent feature of the 

 Forrest City cuts. The loams above and below this soil nre identified as loess by 

 their mineralogical constitution, by their fossils and by their Mndchen. The upper 

 surface of the lower loess sometimes shows unmistakable signs of oxidation and 

 weathering, just below the soil. Laboratory tests for organic matter sustain the 

 determination made in the field, that the coloring matter of the old soil is of 

 vegetable origin. At scores of points in Illinois. Kentucky, Tennessee, Arkansas 

 and Missouri, a similar division of the loess occurs, as distinctly marked as in 

 some parts of the Forrest City seciion, where the less clearly marked soil is 

 traceable into direct continuity with that which is unmistakable. 



The lower division of the extra-drift loess we refer to the first episode of the 

 first glacial epoch. 



