130 GEOLOGY OF THE HIGII PLATEAUS. 



cases, for want of ability to show the contrary, it may be accounted a suf- 

 ficient explanation, and in general it cannot be questioned, that in most 

 volcanoes this identical action plays a more or less important part. Scoria, 

 pumice, and volcanic dust have unquestionably this origin; but the whole of 

 the extravasation is not so accomplished. The outpour of lava is a very 

 different matter. It is comparatively calnr'and quiet in its flow, like water 

 welling forth from a spring; sometimes boiling, bubbling, and spurting a 

 little, but never boisterous or obstreperous. It continues its flow for days 

 and sometimes weeks, but at length ceases and comes to rest. 



A careful examination of the details of volcanic eruptions leaves^the 

 impression that they are pressed up by the weight of rocks which overlie 

 their reservoirs, and that their extravasation is merely a hydrostatic prob- 

 lem of the simplest order. The conception of a liquid inclosed in a cavity 

 beneath the surface and opening to the outer air through a stand-pipe 

 requires some discussion when we come to apply it to volcanic eruptions. 

 Our conceptions of the constrained motion of liquids are derived from 

 experiments upon small quantities of them in small vessels ; but when we 

 come to such enormous volumes as are disgorged by volcanoes, a consider- 

 ation arising from mere magnitude enters into the scheme — a consideration 

 which has no bearing in relation to small volumes. This is the strength 

 of the receptacle. It is a well-known principle in mechanics that the 

 relative strength of a body is inversely proportional to its size. Thus, 

 where we have similar bodies subject to forces which are proportional to 

 their own masses, the resistance to detrusion is proportional only to the 

 square of their linear dimensions. It is this relation which limits the span 

 of an arch or the length of a truss. Now, if we could conceive the contents 

 of one of these subterranean lava reservoirs to be suddenly annihilated, 

 so great must be their dimensions that the rocks above would instantly 

 sink into the cavity, just as the rocks above a coal-mine do on small provo- 

 cation. A small cavity, on the other hand, might persist. Now, the point 

 I wish to illustrate is that the strength of the retaining-walls of a lava 

 reservoir are relatively so weak, in consequence of the large dimensions, 

 that their effect is very nearly the same as it would be if the lava were 

 overlaid by another liquid with which it could not commingle. It is the 



