ON ITS EXTERIOR. VOLCANOES. 447 



loses colour, may have very different causes in dif- 

 ferent substances ; and an examination long looked for 

 in vain, based on appropriate and exact experiments 

 made exclusively on the escaping gases, would lead to 

 an inestimable enlargement of the chemical geology of 

 volcanoes, if regard were at the same time paid to the 

 operation of sea-water in submarine formations, and to 

 the quantity of carburetted hydrogen in the intermixed 

 organic substances. 



The facts which I have brought together at the close 

 of this section the enumeration of volcanoes which pro- 

 duce pumice without obsidian, or much obsidian without 

 pumice, and the remarkable, not constant, but, on the 

 contrary, very various association of obsidian and pumice 

 with certain other minerals led me, long ago, during 

 my sojourn in the Cordilleras of Quito, to the persua- 

 sion that the formation of pumice is the result of a che- 

 mical process, which may take place in trachytes of very 

 various composition, without the necessary intervention 

 of obsidian (i.e. without the necessary preceding existence 

 of obsidian in large masses). The conditions under 

 which such a process goes forward on a great scale are, 

 probably, I would here repeat, dependent less on 

 diversity of materials, than on degrees of heat, of pres- 

 sure determined by depth, of fluidity, and on length of 

 period of solidification. The highly noteworthy although 

 rare phsenomena presented by the isolated occurrence of 

 vast subterranean pumice-stone quarries, apart from any 

 " volcanic framework," lead me to entertain the conjec- 

 ture^ 46 ) that a not inconsiderable portion perhaps, 

 according to volume, the larger portion of volcanic 

 rocks have been emitted, not from elevated volcanic 



