INTERMITTENT CHARACTEE OF ERUPTIONS EXPLAINED. 141 



hundreds of thousands and even milUons of years, should the same vent or 

 chister of vents yield so many different kinds of lava? So completely do 

 the facts of volcanology antagonize the primordial character of lavas, that 

 we seem driven to seek an opposite theory of their origin. 



These difficulties cease to be such and become normal phenomena 

 when we take the postulate of local increments of temperature. The re- 

 fusion of rocks becomes a slow and very gradual process. But when the 

 melted rock is ready for issue, it does not follow that a steady stream of 

 lava would keep flowing as long as the temperature continues to rise. We 

 must now take into consideration the mechanism by which the expulsion is 

 effected. This has already been suggested as the weight of overlying rocks 

 crowding in upon the reservoir, and as these rocks are rigid relatively to 

 small reservoirs, there is a limit to the smallness of the eruption. As the 

 quantity of melted rock increases, this rigidity relatively diminishes until 

 rupture takes place and all the lava hitherto accumulated is expelled. The 

 overlying masses are then soldered up for a time, during which more lava 

 is melted, and when the quantity is sufficient a second eruption occurs, and 

 so the intermittent character is established and for a long period maintained. 



This assumption also explains the co-existence of vents at different 

 levels, the presumption being that each vent derives its lavas from inde- 

 pendent layers or maculae, and that several maculae or layers can suc- 

 cessively find issue through the same vent when, the magmas which they 

 contain reach the eruptive condition. 



There is, however, one comprehensive or generalized fact connected 

 with volcanoes which this assumption does not explain by itself, though it 

 is not in any obvious respect inconsist'jnt with it. This is the geographical 

 distribution of volcanoes. It is well known that existing and recently extinct 

 vents stand in the vicinity of the ocean and large bodies of inland water; a 

 few exceptions, however, being known. But it has been repeatedly re- 

 marked that the postulated rise of temperature is asserted to be a proximate 

 cause, itself I'equiring explanation by the pi'oduction of some ulterior excit- 

 ing cause. If we were able to find this ulterior cause, we should then know 

 why volcanoes have their present distribution. It may be proper to remark 

 here that this distribution would lead us to look for that cause in occur- 



