332 GEOLOGY 



lished. The Keewatin sheet pushed northwestward to the mouth of 

 the Mackenzie, and probably to Banks Land, northward and northeast- 

 ward to the Arctic Islands, 1 and eastward to Hudson Bay, and into 

 confluence with the Labradorean sheet. The latter pushed northward 

 into Ungava Baj^, eastward into the North Atlantic, and southeast- 

 ward into the Gulf of St. Lawrence. 



One of the most marvelous features of the ice dispersion was the 

 pushing out of the Keewatin sheet from a low flat center, without 

 even a suggestion of a mountainous nucleus, 800 to 1000 miles west- 

 ward and southwestward over what is now a rising and semi-arid 

 plain, while the mountain glaciation on the west, where now known, 

 pushed eastward but little beyond the foothills. 



There were probably some important variations from the present 

 altitudes which influenced the spread of the ice. The western region 

 was probably relatively lower, and the eastern relatively higher than 

 now; and while there is no question but that topography is an influ- 

 ential factor in controlling the movement of glacial ice, it is probable 

 that differences of precipitation on the different sides of the ice-sheets, 

 and the consequent differences of topography of the ice-surface were 

 still more important. Differences in the mobility of the ice, due to 

 differences of temperature were also probably effective. In general, 

 it is probable that the factors of growth and mobility take precedence 

 over the topography of the bed in determining the course of movement 

 where thick and extensive bodies of ice are involved, for they not only 

 determine the distribution of the material that is to move, but they 

 develop an ice topography, and sometimes a quasi-fluency, which 

 may become the controlling factors in the movement. 



The Cordilleran ice-sheet 2 is less simply defined. Much of it occu- 

 pied a plateau hemmed in by mountains, and plateau glaciation was 

 complicated by extensive mountain glaciation of alpine type. In 

 some sense, the whole Cordilleran ice-sheet was the product of a con- 

 fluence of mountain glaciers deploying on the intervening plateau; 

 but there appears to have been plateau glaciation not solely dependent 

 on contributions of ice from the mountains. The southerly lobes of 



1 Dawson, G. M., Ann. Rept. Geol. Surv. of Can., Vol. II, 1886, pp. 56-58 R. 



2 Dawson, Ann. Geol., Vol. VI, p. 162, Geol. Surv. of Can., 1888, and Trans. 

 Roy. Soc. of Can., Vol. VIII, Sec. IV; Tyrrell, Geol. Surv. of Can., 1890, E, pp. 1-240, 

 and McConnell, idem, D, pp. 24-28. 



