212 Tertiary. 



7. Glaciers must descend slopes, and must be backed by large sup- 

 plies of perennial snow. Icebergs act independentl}^ and being water- 

 borne, may work up slopes and on level surfaces. 



8. Glaciers striate the sides and bottoms of their ravines very un- 

 equally, acting with great force and effect only on those places where 

 their weight impinges most heavily. Icebergs, on the contrary, being 

 carried b}'- constant currents, and over comparatively flat surfaces, 

 must striate and grind more regularly over large areas, and with less 

 reference to local inequalities of surface. 



9. The direction of the striae and grooves produced by glaciers de- 

 pends on the direction of the valleys. That of icebergs, on the con- 

 trary, depends upon the direction of marine currents, which is not 

 determined by the. outline of surface, but is influenced by the large 

 and wide depressions of the sea bottom. 



10. When subsidence of the land is in progress, floating ice may 

 carry bowlders from lower to higher levels. Glaciers can not do this 

 under any circumstances, though in their progress they may leave 

 blocks perched on the tops of peaks and ridges. 



He further said, that, in all these points of difl'erence, the bowlder 

 clay and drift of Canada, and other parts of North America, cor- 

 respond rather with the action of floating ice than of land ice. 

 More especially is this the case in the character of the striated sur- 

 faces, the bedded distribution of the deposits, the transport of mate- 

 rial up the natural slope, the presence of marine shells, and the 

 mechanical and chemical character of the bowlder cla}'. 



He also enumerated the following Post-pliocene plants as occur- 

 ring, in nodules, at Green's Creek, and other places in Canada, to-wit: 

 Drosera rotitndifoUa, Acer spicatum^ Potentilla canadensis, Gaylus- 

 saccia. resinosa, Populus halsamifera. Thuja occidentalis, Potamo- 

 geton perfoUatus^ P. pusillus, Equisetum scirpoides. None of the 

 plants are properly Arctic in their distribution, and the assemblage 

 may be characterized as a selection from the present Canadian flora of 

 some of the more hai*dy species having the most northern range. At 

 Green's Creek (near Ottawa) the plaat-beariug nodules occur in the 

 lower part of the Leda clay, which contains a few bowlders, and is 

 apparently, in places, overlaid by large bowlders, while no distinct 

 bowlder clay underlies it. The circumstances which accumulated the 

 thick bed ot bowlder clay near Montreal, were probably absent in the 

 Ottawa valley. In any case, we must regard the deposits of Green's 

 Creek as coeval with the Leda clay of Montreal, and with the period 



