450 COSMOS. 



and the richer they are in alkalies. It remains, however, 

 very uncertain, according to Rammelsberg's researches,* 

 whether the tumefaction is to be ascribed to the volatiliza- 

 tion of potash or hydrochloric acid. It is probable that 

 similar phenomena of inflation in trachytes rich in obsidian 

 and sanidine, in porous basalts and amygdaloids, in pitch- 

 stone, tourmalin, and that dark-brown flint which loses its 

 color, may have very different causes in the different mate- 

 rials themselves. An investigation which has now been long 

 looked for in vain, founded on accurate experiments, exclu- 

 sively directed to these escaping gaseous fluids, would lead to 

 an invaluable extension of our knowledge of the geology of 

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

 operation of the sea-water in subterranean formations, and 

 to the great quantity of carbureted hydrogen belonging to 

 the commingled organic substances. 



The facts which I have brought together at the end of this 

 section, the enumeration of those volcanoes which produce 

 pumice without obsidian, and those which yield a great deal 

 of obsidian and no pumice the remarkable, not constant, 

 but very diversified association of obsidian and pumice with 

 certain other minerals, early led me, during my residence in 

 the Cordilleras of Quito, to the conclusion that the forma- 

 tion of pumice is the result of a chemical process, which may 

 be verified in trachytes of very heterogeneous composition, 

 without the necessity of a previous intervention of obsidian 

 (that is to say, without its pre-existence in large masses). 

 The conditions under which such a process is performed on a 

 large scale are perhaps founded (I would here repeat) less on 

 the diversity of the material than on the gradation of heat, 

 the pressure determined by the depth, the fluidity, and the 

 length of time occupied in solidification. The striking, though 

 rare, phenomena presented by the isolation of immense sub- 

 terraneous pumice quarries, far from any volcanic structures 

 (conical and befl-shaped mountains), lead me at the same 

 time to conjecture! that a not inconsiderable perhaps even, 

 in regard to volume, the greater number of the volcanic 

 rocks have been erupted, not from upraised volcanic struc- 



* Rammelsberg, in Poggend., AnnaL, bd. Ixxx., 1850, s. 464, and 

 fourth supplement to his Chemische Handworterbuch, s, 169 j compaue 

 also Bischof, GeoL, bd. ii., 2224, 2232, 2280. 



t See above, p. 291, 311, 312-316, 322-325. For particulars re- 

 specting the geographical distribution of pumice and obsidian in the 

 tropical zone of the New Continent, see Humboldt, Essai Gcoqnostiqite 

 stir k Gisement des Roches, etc., 1823, p. 340-342, and 344-347. 



