MODES OF GLACIAL ACTION. 201 



Modes of Deposition of the englacial Drift. — Again, on this topic, which, 

 like the ways of transportation of the drift, would need too much space, 

 for its adequate presentation here, I may refer to my several papers before 

 cited. It is desirable, however, to call special attention to the discrimina- 

 tion by Hoist of the enclosed drift and the subglacial drift of the Green- 

 land ice-sheet * In New England and in Minnesota, the Dakotas, and 

 Manitoba, I find corresponding divisions of the sheet of till, the lower 

 part being shown by its hard and compact condition and by other 

 characters to be the ground moraine of the ice-sheet, while the upper 

 part, which is comparatively loose, with more plentiful large bowlders, 

 appears to have been dropped from an englacial or superglacial position 

 when the ice melted. 



Russell, in the later paper already quoted from, discusses in a very 

 instructive manner, by analogy with the products of the Malaspina 

 glacier, the origin of Pleistocene eskers and kames, glacial flood-plains of 

 gravel and sand, and somewhat similar high lacustrine plains and ter- 

 races. All the explanations which he thus gives are undoubtedly appli- 

 cable generally or at least in some localities to the drift formations of the 

 Glacial period ; but, as noted in my foregoing paper on the Pinnacle hills 

 and Pittsford eskers, near Rochester, New York, I think that mainly the 

 retreat of the Pleistocene ice-sheet was so rapid that its drainage then 

 was almost wholly by superglacial streams, in whose open channels the 

 eskers of the northern United States and Manitoba, so far as I have 

 studied them, appear to have been formed. 



The accumulation of drumlins, which has been so difficult of explana- 

 tion, seems to me referable to convergent currents of the ice-sheet during 

 its retreat amassing its englacial drift in these peculiarly moulded hills, 

 this drift having been first exposed on the ice surface by ablation, like 

 that on the outer part of the Malaspina glacier, then enveloped again in 

 the ice as a drift stratum, and finally carried into these accumulations 

 beneath the ice.f 



Origin of Forest Beds between Deposits of Till— On extensive tracts of the 

 Mississippi basin, most notably in Indiana, Illinois, Iowa and south- 

 eastern Minnesota, the drift sheet comprises in many places, and some- 

 times wellnigh in every well of whole townships, a horizon marked by 

 prostrate trunks of trees, fallen leaves or rushes and grass of swamps, or 

 occasionally by layers of peat, above and beneath which the sections pass 

 through till. In different districts the forest beds differ in their relation- 

 ship to the terminal moraines, so that they appear to represent more 

 than one stage of ice advance after at least a considerable retreat. Per- 



*Am. Naturalist, vol. xxii, pp. 589-598 and 705-713, July and August, 1888. Am. Geologist, vol. viii, 

 pp. 383, 384, December, 1891. 

 |Am. Geologist, vol. x, pp. 339-302, December, 1892. 



