TRUE VOLCANOES. 449 



pclago of the Gallapagos. Wc have already remarked in 

 another place that cones of cinders are wanting in the mighty 

 volcano of Manna Loa, in the Sandwich Islands, as well as 

 in the volcanoes of the Eifel,* which once emitted lava 

 streams. Though the island of Java contains a series of 

 more than forty volcanoes, of which as many as twenty-three 

 are still active, yet Junghuhn was only able to discover two 

 points in the volcano of Gunung Guntnr, near Bandong and 

 the great Tengger Mountains,! in which masses of obsidian 

 have been formed. These do not appear to have given occa- 

 sion to the formation of pumice. The sand lakes of Dasar, 

 which lie about G828 feet above the mean level of the sea, 

 are not covered with pumice, but with a layer of rapilli, de- 

 scribed as being obsidian-like, semi-vitrified fragments of ba- 

 salt. The cone of Vesuvius, which never emits pumice, gave 

 out, from the 21th to the 28th of October, 1822, a layer 

 eighteen inches thick of sand-like ashes, consisting of pulver- 

 ized trachytic rapilli, which has never been mistaken for 

 pumice. 



The cavities and air-holes of obsidian, in which crystals of 

 olivin, probably precipitated from vapors, have formed — as, 

 for example, in the Mexican Cerro del Jacal — are sometimes 

 found, in both hemispheres, to contain another kind of in- 

 closures, which seem to indicate the manner of their origin 

 and formation. In the wider portions of these long-extended, 

 and for the most part very regularly parallel cavities, frag- 

 ments of half-decomposed earthy trachyte are found imbed- 

 ded. Beyond these the cavity runs on in the form of a tail, 

 as if a gas-like elastic fluid had been developed by volcanic 

 heat in the still soft mass. This phenomenon particularly 

 attracted the attention of Leopold von Buch when he visited 

 the Thomson collection of minerals at Naples, in company 

 with Gay-Lussac and myself, in the year 1805.$ The infla- 

 tion of obsidian by the operation of fire, which did not escape 

 attention in the early period of Grecian antiquity ,§ is cer- 

 tainly caused by some such development of gas. According 

 to Abich, obsidians pass the more easily into cellular (not 

 parallel-porous) pumice, the poorer they are in silicic acid 



* See above, p. 367, and notes, p. 302-304. 



f Franz Junghuhn, Java, bd. ii., s. 388, 592. 



% Leopold von Buch, in the Abhandl. der Akademie der Wiss. zu 

 Berlin, for the years 1812-1813 (Berlin, 1816), s. 128. 



§ Theophrastus de Lapidibus, s. 14 and 15 (Opera cd. Schneider, t. i., 

 1818, p. 689 ; t. ii., p. 426 ; and t. iv., p. 551), says this of the "lipa- 

 rian stone" (kiTrctpcrtog). 



