164 COSMOS. 



the internal lava; and the deficiency, or, at all events, 

 very rare occurrence of burning hydrogen gas during the 

 eruption (which the formation of hydrochloric acid,* am- 

 monia, and sulphureted hydrogen certainly does not suffi- 

 ciently replace), has led the celebrated originator of this 

 hypothesis to abandon it of his own accord, f 



According to a third view, that of the highly-endowed 

 South American traveler, Boussingault, a deficiency of co- 

 herence in the trachytic and doleritic masses which form 

 the elevated volcanoes of the chain of the Andes, is regard- 

 ed as a primary cause of many earthquakes of very great 

 extent. The colossal cones and dome-like summits of the 

 Cordilleras, according to this view, have by no means been 

 elevated in a soft and semi-fluid state, but have been thrown 

 up and piled on one another when perfectly hardened, in 

 the form of enormous, sharp-edged fragments. In an ele- 

 vation and piling of this description, large interstices and 

 cavities have necessarily been produced ; so that by sud- 

 den sinking, and by the fall of solid masses which are too 

 weakly supported, shocks are produced. f 



that the radicals of silica, alumina, lime, and iron are combined with 

 chlorine in the interior of the earth," and the penetration of sea-wa- 

 ter does not appear to him to be improbable under certain conditions 

 (p. 419, 420, 423, and 42G). Upon the difficulty of a theory founded 

 upon the penetration of water, see Hopkins, I3rit.Assoc.Hfp., 1847, p. 38. 



* According to the beautiful analyses made by Boussingault on the 

 margins of five craters (Tolima, Furace, Pasto, Tuqueras, and Cum- 

 bal), hydrochloric acid is entirely wanting in the vapors poured forth 

 by the South American volcanoes, but not in those of Italy (Annaks 

 de Chiutie, tome lii., 1 833, p. 7 and 23). 



f Cosmos, vol. i., p. 236. While Davy, in the most distinct man- 

 ner, gave up the opinion that volcanic eruptions are a consequence of 

 the contact of the metalloid bases with water and air, he still assert- 

 ed that the presence of oxydizable metalloids in the interior of tho 

 earth might be a co-operating cause in volcanic processes already com- 

 menced. 



J Boussingault says : " I attribute most of the earthquakes in the 

 Cordillera of the Andes to falls produced in the interior of these 

 mountains by the subsidence which takes place, and which is a conse- 

 quence of their elevation. The mass which constitutes these gigantic 

 ridges has not been raised in a soft state ; the elevation did not. take 

 place until after the solidification of the rocks. I assume, therefore, 

 that the elevated masses of the Andes are composed of fragments 

 heaped upon each other. The consolidation 1 of the fragments could 

 not be so stable from the beginning as that there should be no 

 settlements after the elevation, or that there should be no inte- 

 rior movements 'in the fragmentary masses" (Boussingault, Sur Ics 

 Tremblemens de Terre des Andes, in Annales de Chimie et de Phy- 

 sique, tome IviiL, 1835, p. 84-86). In the description of his mem- 



