62 BUNSEN ON THE FORMATION OP 



tuffs of the Faroe Islands, and elsewhere, admits of the most 

 simple explanation by the reducing action of this hydrogen. 



It is not very easy at first sight to conceive from whence the 

 alkali might have been derived which could thus give rise to the 

 formation of tuff in Iceland. Considering the variable per-cen- 

 tage of alkali in the pyroxenic rocks compared with the more 

 constant proportions of their other constituents, we might be in- 

 duced to assume that a separation of these alkalies took place in 

 the rocks while in a state of igneous fusion. The idea of such a 

 separation of alkalies is indeed not to be unconditionally rejected. 

 It is well known that most salts are decomposed at high tempera- 

 tures. If their acids are considerably more volatile than the 

 bases, either basic salts are formed, or the bases are liberated 

 while the acid is volatilized. The compounds of sulphuric, car- 

 bonic, nitric, arsenious acids, &c. undergo, with few exceptions, 

 this decomposition. If, on the contrary, the acid is less volatile 

 than the base or the salt itself, as is the case with the ammonia- 

 cal salts, then it is the base which is volatilized while the acid 

 remains behind. The alkaline silicates may very probably suf- 

 fer the latter kind of decomposition ; for if silica and caustic pot- 

 ash or soda are heated together upon a platinum wire by means 

 of a galvanic current to near the melting-point of the metal, the 

 alkali is volatilized by a temperature at which the silica does not 

 even begin to melt. It would appear therefore that in lavas 

 which are heated to a temperature so elevated as to become 

 liquid enough to be shot out in large parabolic curves from the 

 opening of the crater, it is not only possible, but even probable, 

 that a separation of alkalies by heat takes place, the more espe- 

 cially if it is remembered that carbonic acid or aqueous vapour, 

 never wanting in such volcanic processes, must have given rise 

 to the formation of hydrated alkalies and carbonates, compounds 

 which are so volatile that their sublimation may be directly ob- 

 served in some technical operations. There are indeed instances 

 in which the separation and volatilization of the alkaline consti- 

 tuents even of silicates may be detected. In the coal and flux 

 used in the blast furnaces in England, which are remarkable for 

 their intense heat, the whole quantity of the alkali present is in 

 combination with silica. Nevertheless, at the foot of these fur- 

 naces, where for years together a temperature prevails nearly 



