THE ERKATIC PHENOMENA. 405 



the agent ; floating ice is supposed to have ground and polished the 

 surfaces of rocks, while I consider^ them to have been acted ujjon by 

 terrestrial glaciers. To settle this difference we have a test which is 

 as irresistible as the other arguments already introduced. 



Let us investigate the mode of action, the mode of transportation 

 of icebergs, and let us examine whether this cause is adequate to 

 produce phenomena for which it is made to account. As mentioned 

 above, the polished surfaces are continuous over hills, and in depres- 

 sions of the soil, and the scratches which run over such undulating 

 surfaces are nevertheless continuous in straight lines. If we imagine 

 icebergs moving upon shoals, no doubt they would scratch and 

 polish the rocks in a way similar to moving glaciers. But upon such 

 grounds they would sooner or later be stranded, and if they remain- 

 ed loose enough to move, they would, in their gyratory movements, 

 produce curved lines, and mark the spots Avhere they had been 

 stranded with particular indications of their prolonged action. But 

 nowhere upon arctic ground do we find such indications. Every- 

 where the polished and scratched surfaces are continuous in straight 

 juxtaposition. 



Phenomena analogous to those produced by icebergs would only 

 be seen along the sea-shores ; and if the theory of drifted icebergs 

 were correct, we should have, all over those continents where erratic 

 phenomena occur, indications of retreating shores as far as the erratic 

 phenomena are found. But there is no such thing to be observed 

 over the whole extent of the North American continent, nor over 

 Northern Europe and Asia, as far as the northern erratics extend. 

 From the arctics to the southernmost limit of the erratic distribu- 

 tion, we find nowhere the indications of the action of the sea as 

 directly connected with the production of the erratic phenomena. 

 And wherever the marine deposits rest upon the polished surfaces 

 of ground and scratched rocks, they can be shown to be deposits 

 formed since the grooving and polishing of the rocks, in consequence 

 of the subsidence of those tracts of land upon which such deposits 

 occur. 



Again, if we take for a moment into consideration the immense 

 extent of land covered by erratic phenomena, and view them as 

 produced by drifted icebergs, we must acknowledge that the ice- 



