226 REACTION OF THE INTERIOR OP THE RAHTJI 



noes, or a short time before the elevation of a volcanic 

 island, we can give no explanation. 



When it is asked, what it is that burns in volcanoes, 

 what excites the heat, fuses the earths and metals, and im- 

 parts to lava currents of great thickness ( 223 ) a heat which 

 lasts for many years, the question assumes by implication, 

 that the presence of materials capable of supporting combus- 

 tion is indispensable in volcanoes, like the beds of coal in 

 subterranean fires. According to the different phases which 

 chemical science has passed through, bitumen, pyrites, or a 

 humid mixture of pulverised sulphur and iron, pyrophoric 

 substances, and the metals of the earths and alkalies, have 

 been successively assigned as the cause of active volcanic 

 phsenomena. Sir Humphry Davy, the great chemist to 

 whom we owe the knowledge of these latter most inflam- 

 mable metals, has himself renounced his bold chemical 

 hypothesis in his last work, " Consolation in travel, and last 

 days of a Philosopher/'' which cannot be read without a sen- 

 timent of melancholy. The higji mean density of the earth' 

 (5.44), compared with the much inferior specific gravity of 

 potassium (0.865), and of sodium (0.972), or of the metals 

 of the earths (1.2), the absence of hydrogen in gaseous 

 emanations from the fissures of craters, and from lava cur- 

 rents which have not yet cooled ; and lastly, many chemical 

 considerations, oppose- themselves to the earlier conjectures 

 of Davy and of Ampere. ( 224 ) If hydrogen were disengaged 

 by the eruption of lava, what prodigious quantities of that gas 

 must have been set free in the memorable eruption at the 

 foot of the Skaptar- Jokul, in Iceland, described by Macken- 

 zie and Soemund Magnussen, which lasted from the l|th 

 of June to the 3d of August, and covered very many 



