40 Messrs Lyell and Murchison on the 



dern, so we may expect to find proofs that the obstructions of- 

 fered by them to the course of rivers, were in hke proportion 

 more considerable. 



It must not be forgotten, that in three of the cases under 

 consideration, we have decisive proofs that not only the refilling, 

 but even the first formation, of the ancient valleys, took place 

 long after the commencement of volcanic action in Auvergne. 

 For the original valleys at St Maurice and Perrier, and that re- 

 opened by the Sioule, above the Etang de Fung, had been ex- 

 cavated, in the two first instances, through tertiary strata, and in 

 the last-mentioned case through gneiss, all capped with pla- 

 teaus of basalt. Assuming, then, the existence of lakes caused 

 by currents of lava, it is evident, that the transported materials 

 therein accumulated, were not all introduced into them at one 

 time ; for such contiguous beds as consist almost exclusively of 

 volcanic matter, often differ widely from each other in compo- 

 sition, nor are their contents arranged according to their speci- 

 fic gravity. On the contrary, layers of light pumice underlie 

 others of massive trachytic breccia and conglomerate, or alter- 

 nate, as above Perrier, several times with alluvium containing 

 heavy fragments of basalt. Undoubtedly the separation of pu- 

 mice from all rocks washed down at the same time, may natu- 

 rally be accounted for, since, from its buoyancy, it would float 

 for a while upon the lake, just as it is often observed to cover 

 the sea in the Grecian Archipelago, after submarine eruptions. 

 But as often as we see strata of pumice covered by heavier mat- 

 ter, we must suppose an inter\al during which the former had 

 time to subside. 



But if it appear that the alternating breccias and alluvions 

 were not all formed at one time, so also is it clear that they were 

 not all produced by the same mode of action. The beds of 

 rounded pebbles of quartz, granite, and basalt, resemble those 

 produced by the continued action of the present river. But 

 the angular forms of the fragments in the interslratified trachy- 

 tic breccias bespeak violence in the destroying cause, and in- 

 sufficient time between their fracture and subsequent deposition 

 to have allowed ol" the triturating effects of running water. As 

 they cannot by any characters be distinguished from breccias of 

 !>imilar composition and magnitude entering into the structure 



