CHAP. XT. BY THE EROSIVE ACTION OF GLACIERS. 311 



by their contents to have come from the upper Alpine 

 valleys above the lakes. Such moraines often rejiose on an 

 older stratified alluvium, made up of rounded and worn 

 pebbles of precisely the same rocks as those forming the 

 moraines, but not derived from them, being small in size, 

 never angular, polished, or striated, and the whole having 

 evidently come from a great distance. These older alluvial 

 strata must, according to M. de Mortillet, be of pre-glacial 

 date, and could not have been carried past the sites of the 

 lakes, unless each basin had previously been filled and 

 levelled up with mud, sand, and gravel, so that the river- 

 channel was continuous from the uj)per to the lower extremity 

 of each basin. 



Professor Eamsay, after acquiring an intimate knowledge 

 of the glacial phenomena of the British Isles, had taught, 

 many years before, that small tarns and shallow rock-basins, 

 such as we see in many mountain-regions, owe their origin 

 to glaciers which erode the softer rocks, leaving the harder 

 ones standing out in relief and comparatively unabraded. 

 Following up this idea after he had visited Switzerland, and 

 without any communication with M. de Mortillet or cog- 

 nizance of his views, he suggested in 1859 that the lake- 

 basins were not of pre-glacial date, but had been scooped out 

 by ice during the glacial period, the excavation having for 

 the most part been effected in miocene sandstone, provincially 

 called, on account of its softness, "molasse." By this theory 

 he dispensed with the necessity of filling up pre-existing 

 cavities Avith stratified alluvium, in the manner proposed by 

 M. de Mortillet. 



I will now explain to what extent I agree with, and on what 



points I feel compelled to differ from, the two distinguished 



o-eoloo-ists above cited. 1st. It is no doubt true, as Professor 



Eamsay remarks, that heavy masses of ice, creeping for ages 



over a surface of dry land (whether this comprise hills, 



21 



