Lavis — On the Structure of Rocks. 123 



which we should deduce on the above theoretical grounds, com- 

 pletely coincide with earthquake registration. 



The extension of the fissure may have been sufficient to allow 

 of the formation of steam, which may collect together throughout 

 the pasty mass as bubbles ; and, should solidification soon follow, 

 the resulting dyke-metal would present a vesicular or amygdaloidal 

 structure. On the other hand, the expansion may only have taken 

 place to such a point, that no conversion of liquid into gas has 

 taken place, and as a result we should look for, in case of solidifi- 

 cation, a dyke presenting no signs of vesicularity. The finding of 

 a dyke-metal, in which no vesicularity is manifest, is no proof that 

 at some time it may not have had such ; for, were cooling not to 

 follow soon on vesicularization, the renewed gradually increasing 

 pressure would again compel the steam to redissolve in the magma. 

 These facts probably account for the rarity of a vesicular state of 

 granite, though even this is sometimes known to occur as in the 

 island of Mull, and that of the plateau of the Palais du Boi, Lozere, 

 described by Lecoq. 1 



Under the two former circumstances we should expect the first 

 to end in solidification more often than the second ; for, by the 

 conversion of the dissolved water into steam, a very much larger 

 amount of heat would be used up, proportionally of course to the 

 amount of conversion that took place. 



By the progressive extension of the fissure a larger area of 

 igneous rock surface will be exposed to the conditions which have 

 been described, so that the tendency will be towards the more rapid 

 absorption of water, and consequent crisis between tension and re- 

 sistance. Besides, from the larger amount of expansive matter 

 capable of acting, the effects will be more violent each consecutive 

 time. 



The facts are borne out in such examples as Jorullo and Monte 

 Nuovo, and are now probably in progress under Casamicciola, in 

 the island of Ischia. In such examples we find, that for a long 

 period earthquakes occur at distant intervals, but that these have a 

 gradual tendency to follow each other progressively more often, 

 and often increasing in destructiveness at one spot, although the 

 area affected may proportionally become more concentrated. At 



1 Les epoques geologiques de l'Auvergne, I., page 465. 



