272 W. UPHAM — GLACIAL LAKES IN CANADA. 



its last remnants, moisture-laden winds doubtless carried portions of it across 

 Baffin's bay and Davis strait to be deposited again in the ice-sheet that still 

 covers the interior of Greenland. 



Proportion of Englacial Drib-p supplied to the Deltas of 



Glacial Lakes. 



Osars, kames, and terminal moraines, the comparatively loose upper part 

 of the till, and the beds of gravel, sand and clay which form valley terraces 

 and plains in drift-covered regions, appear to be derived from the englacial 

 drift that had been gathered up into the ice-sheet from the surface over 

 which it moved, especially from the sides of hills and mountains, being 

 thence borne forward in the lower portion of the ice. My observations of 

 the Coteau des Prairies at lakes Benton, Shaokatan, and Hendricks in south- 

 western Minnesota,^ and of the osar at Bird's Hill. Manitoba,t indicate that 

 the amount of englacial drift contained in certain favored portions of the 

 ice-sheet, as near its border where conspicuous moraines were being formed, 

 or on tracts to which glacial currents converged, was equal to a general drift 

 sheet 30 to 40 feet thick. During the departure of the ice, its englacial drift 

 became superficial by the melting of its upper part, until, when only a thick- 

 ness of a few hundred feet or less of the ice remained, its surface was covered 

 by the drift which it had held ; and such a condition probably extended 

 many miles back from the margin of the ice-sheet during its entire retreat 

 from the glaciated area. The latest glacial melting and the attendant 

 rains therefore caused much of the englacial and finally superficial drift to 

 be washed away by streams and deposited as stratified or modified drift, 

 some of the coarsest portions being laid down in ice-walled channels as osars 

 and kames, but far the greater part being borne beyond the ice-margin to 

 form flood-plains of rivers and deltas of lakes or of the sea. 



The " White silts " of British Columbia were doubtless in large part de- 

 rived directly from the englacial drift of the Cordilleran portion of the 

 ice-sheet ; and the relationship of the topography to the recession of the ice 

 determined for the various districts where these silts are developed whether 

 their deposition was fluvial or lacustrine. And the direction of glacial 

 drainage across water-sheds affords reliable information of the course of 

 retreat of the ice-border on the same areas. For example, the old river-bed 

 described by Dawson as descending to the southward and westward between 

 Dease lake, tributary by the Liard and Mackenzie to the Arctic ocean, and 

 the Tanzilla river, tributary by the Stikine to the Pacific, proves that the 

 ice-border there retreated northeast ward. J 



* Geology of Minnesota, vol. 1, 1884, pp. 603-4. 



fGeol. and Nat. Hist. Survey of Canada, Annual Report, vol. IV, 1888-89, pp. 38-40 E. 

 tGeol. and Nat. Hist. Survey of Canada, Annual Report, vol. Ill, 1887-88, p. 69 B. 



