chap. in. Obsidian Formation, 73 



seem to mark the passage, from the perfectly glassy to 

 the stony and crystalline beds. Beudant's account ] of 

 his ' perlite lithoide globulaire ' in every, even the 

 most trifling particular, might have been written for 

 the little brown sphaerulitic globules of the rocks of 

 Ascension. 



From the close similarity in so many respects, be- 

 tween the obsidian formations of Hungary, Mexico, 

 Peru, and of some of the Italian islands, with that of 

 Ascension, I can hardly doubt that in all these cases, 

 the obsidian and the sphserulites owe their origin to a 

 concretionary aggregation of the silica, and of some of 

 the other constituent elements, taking place whilst the 

 liquefied mass cooled at a certain required rate. It is, 

 however, well known, that in several places, obsidian 

 has flowed in streams like lava ; for instance, at Teneriffe, 

 at the Lipari Islands, and at Iceland. 2 In these cases, 

 the superficial parts are the most perfectly glassy, the 

 obsidian passing at the depth of a few feet into an opaque 

 stone. In an analysis by Vauquelin of a specimen of 

 obsidian from Hecla, which probably flowed as lava, the 

 proportion of silica is nearly the same as in the nodular 

 or concretionary obsidian from Mexico. It would be 

 interesting to ascertain, whether the opaque interior 

 portions and the superficial glassy coating contained 

 the same proportional constituent parts : we know from 

 M. Dufrenoy 3 that the exterior and interior parts of 

 the same stream of lava sometimes differ considerably 

 in their composition. Even should the whole body oi 

 the stream of obsidian turn out to be similarly com- 



1 Beudant's ' Voyage,' torn. iii. p. 373. 



2 For Teneriffe, see Von Buch, ' Descript. des Isles Canaries,' pp. 

 184 and 190 ; for the Lipari Islands, see Dolimieu's ' Voyage,' p. 34 ; 

 for Iceland, see Mackenzie's ' Travels,' p. 369. 



3 ' Memoires pour servir a. une Descript. Geolog. de la France,' 

 torn. iv. p. 371. 



