618 GEOLOGY. 



after they have escaped from the crater, when they give them off from 

 their expanded surfaces in the open air, where there is no restraint to 

 develop explosiveness ; (4) some flows are so massive that they cool to 

 the chief gas-discharging point only after they are spread out on the 

 surface, when quiet escape is possible; (5) probably a main occasion 

 of the very violent explosions lies in the fact that the lavas have begun 

 to crystallize while yet in the duct of the volcano. The crystals, in 

 forming in the magma, exclude the gases from themselves, and this 

 excluded portion overcharges the remaining portion of the lava. This 

 process continues as the Ifiva rises and grows cooler until the gases 

 acquire great volume and explosive force. This view is sustained by 

 the fact that the pumice and ash of such extraordinarily explosive 

 eruptions as those of Krakatoa and Pelee contain many small crystals 

 which had certainly formed before the explosive inflation took place. 

 Incipient crystallization does not, however, appear to be a universal 

 accompaniment of explosive action. 



Spasmodic action. — The discharge of the gases is spasmodic, and 

 usually consists of a succession of distinct explosions. Sometimes 

 these succeed one another at rather constant and frequent intervals, 

 as in Stromboli, where the explosions follow one another at intervals 

 of three to ten or more minutes. In many others the outbursts are 

 rhythmic, while in others the spasms are distant and irregular. 



Kinds of gases. — Steam is the chief volcanic gas. Its constituents, 

 hydrogen and oxygen, are also present in the free state, and are per- 

 haps the result of the dissociation of the steam at the very high tempera- 

 tures of the lavas. Carbon dioxide is probably next in abundance. No 

 positive statement as to the relative amounts of the subordinate gases 

 can be made because of the obvious difficulties of obtaining anything like 

 a representative analysis of the gases concerned in the great volcanic 

 eruptions. The materials for the analyses which have been made were 

 derived chiefly from little secondary or '^parasitic" vents, or from side- 

 wall crevices, through which the volcanic gases rise. Such vents prob- 

 ably derive their gases from the very border of the main mass, where it 

 is most subject to the influence of waters and gases from the adjacent 

 walls, and it is uncertain how far they are truly representative of the 

 gases in the interior of the lava itself. The data now at command 

 seem to indicate that carbon dioxide increases greatly in relative abun- 

 dance as volcanic action dies away. Great quantities of this gas are 



