200 ROCHESTER ACADEMY OF SCIENCE. [Jan. 9, 



fication of the beds forming the edge of the plains where they adjoined 

 the ice-sheet, and attributes it to the upflow of subglacial waters 

 bringing with them the sediments which make the plain and reach to 

 a considerable distance, having in their lower i)ortion, on the 

 greater part of their area, the forwardly dipping stratification that is 

 characteristic of deltas or of dc[)osits swept by torrential currents into 

 the slowly Mowing broad exi)anse of Hooded rivers. It seems to me, 

 however, more probable that the back-set beds were formed by the 

 downward and backward transfer of sand from the surface of the 

 plain, to nil in succession the small spaces from which the ice-sheet 

 was gradually withdrawn. 



Because the summer melting of the North American ice-sheet in 

 the Champlain epoch, or closing stage of the Glacial period, was far 

 more rapid than that of the Alaskan glaciers at the present day, the 

 previously existing small subglacial stream-courses were inadecjuate 

 for the transportation of the large supplies of englacial drift then set 

 free, by which, indeed, the subglacial tunnels appear to have been 

 mostly obstructed and closed. The waters of the glacial melting and 

 of accompanying rains therefore flowed, as I believe, in channels on 

 the ice surface, often near their mouths more like cafions than like 

 ordinary land valleys, there depositing the eskers and kames. 



My studies of the Pinnacle hills and Fittsford esker series, of the 

 very massive kame deposits forming the greater part of the outermost 

 terminal moraine on I-ong Island eastward from Roslyn ('), of the 

 large kame called the Devil's Heart, rising in a somewhat conical hill 

 175 feet above the adjoining country south of Devil's lake in North 

 Dakota, and of the esker named Bird's hill, seven miles northeast of 

 Winnipeg {"), seem to me to demonstrate, beyond all doubt, that 

 their material, and probably likewise that of eskers and kames gener- 

 ally, was supplied by supergiacial streams from the plentiful englacial 

 drift, and could n(jt have been brought from drift beneath the ice by 

 subglacial drainage. 



(i). Am. Jour. Sci., Ill, Vol. XVIIl., pp. 84-88, Aufj., '87.;. 



(2). Gcol. and Nat. Hist. .Survey of Canada, Annual report, new series, Vol. IV, for 1888- 

 89, pp. .58-42 E, with section. 



