82 W. UPPIAM — DERIVATION OF KAMES, ESKERS, AND MORAINES. 



eroded from the land and slowly borne forward in the very sluggishly 

 moving basal portion of the ice. 



Nearly all the rock surface of glaciated countries first suffered much 

 erosion by the ice-sheet, before the deposition of its drift. This implies 

 that all of the drift on these great areas has been at some time engiacial, 

 since the base of the moving ice studded .with its enclosed drift was the 

 agent abrading the rock and producing its striae. Observations of widely 

 different glacial striation on adjoining parts of the same rock exposure, 

 the two sets of striae being on surfaces of slightly different inclination 

 which join each other with a bevelled edge or angle, convince me that 

 a very definite plane divided the bottom of the ice-sheet with its enclosed 

 drift bowlders, pebbles, and sand grains, which acted as graving tools, 

 from underlying stationary drift accumulations, not less than from the 

 immovable bed-rock. There was, at least in the places of these crucial 

 observations, no pushing or dragging forw^ard of the drift beneath the 

 ice. Several of these localities of adjacent differently striated bevelled 

 rock surfaces I have noted in southwestern Minnesota, and others in 

 Somerville, Massachusetts.* In all these places the contrasted directions 

 of glaciation are apparently referable to deflections of a continuous ice- 

 current, rather than to any withdrawal and subsequent new advance of 

 the ice-sheet. During some considerable interval between the times of 

 different courses of the glacial current, a very thin layer of stationary 

 drift covering a part of the rock surface protected it from striation while 

 the later ice erosion engraved its striae on the closely contiguous part of 

 the same ledge. Here, and I think likewise generally, the only trans- 

 portation of drift took place within the moving ice-sheet, not by any 

 sliding or rolling under it. Where all the underlying rock is glacially 

 worn and striated, we must therefore conclude that all the drift earlier 

 or later Avas carried into the ice. 



How the continental ice-sheet could thus erode all its rock-bed, ex- 

 cepting areas near the limits of its farthest extension, and yet finally 

 leave the same rock-bed, upon many large tracts, completely and deeply 

 enveloped by the subglacial till, 1 have attempted to show^ in the accom- 

 panying idealized section (figure 1) of a portion of this ice-sheet in the 

 Mississippi basin. On the southern part of the drift in that region no 

 continuous moraines were accumulated, and I attribute their absence 

 principally to the attenuated condition of the ice there and its lack of a 

 steep border. During the glacial retreat, wherever the vicissitudes of 

 the wavering climate caused the chiefly waning ice-border to remain 



* Geology of Minn., vol. i, 1884, pp. 504-5 and 549-50. Proc. Boston Soe. Nat. Hist., vol. xxvi, 1893, 

 pp. 33-42. Compare with Professor Chainberlin's memoir, " The Roek-Scorin^s of the great Ice 

 Invasions," U. S. Geol. Survey, Seventh An. Rep., lor 1885-86, pp. 147-248, especunliy pp. 175, 17G, 

 200-207. 



