GEXKRAL CONCLUSIONS. 297 



which are incapable of ]n-oof. (1.) It is taken for granted 

 that the upper limits of glaciation in tlie mountain ranges 

 of America indicate the thickness of a continental ice- 

 sheet. They probably indicate only the upper limit of 

 the abrasion of local glaciers. (2.) Hence it is computed 

 that the thickness of a continental glacier Howing radially 

 outward in all directions from the Laurentian highlands 

 of Canada, amounted to two miles ; and in connection 

 with this it is stated that the maximum thickness of 

 the great Cordilleran glacier of British Columbia has 

 been estimated to have been about 7,000 feet ; an 

 entirely different thing, and referring to the maximum 

 depth of a local glacier traversing deep valleys. (3.) It 

 is admitted that the assumed continental glacier could 

 not move without an elevation of the Laurentian high- 

 lands to the height of several tliousand feet, of which we 

 have no evidence, for the cutting of the deep fiords 

 referred to in this connection must have taken place in 

 the time of Pliocene elevation of the continents before 

 the glacial period. (4.) Tlie Upper and Lower Boulder 

 drift, so different in their characters, are accounted for on 

 the supposition that the former comes from material sus- 

 pended in the ice at some height above its base, the other 

 from that in the bottom of the ice. In like manner the 

 widely distributed interglacial beds holding remains of 

 land plants of North temperate character, are attributed 

 to such small local occurrences of trees on or under 

 moraines as appear in the Alaska glaciers. (5.) The rapid 

 disappearance of the ice is connected with a supposed 

 subsidence of the land luider its weiglit, though from 

 other considerations we know that if this was dependent 

 on such a cause, it must have Ijeen going on from the 

 first gathering of the ice, so that the required high land 



