204: Jaggar — Yolcanologic Investigations at Kilauea. 



that the violent downsncking at the grottoes and fountains 

 mnst be accounted for bv some process other than chemical 

 condensation. 



The only conceivable process other than gas condensation, 

 in view of the obviously excessive discharge of gas and libera- 

 tion of heat which occur at these places, to account for the 

 loss of volume which produces downward suction, is a partial 

 de-vesiculation of the lava, a loss of gas which decreases its 

 bulk, and to that extent increases its density. It is plain that 

 if the furnace of a grotto stack could reheat the vesicidated 

 lava for any depth, we would have our convection reversed 

 and so would become involved in paradox. 



It is evident that the reheating does not prevent convectional 

 downflow of the liquid, and this would clearly not be the case 

 if we had to deal with mere thermal convection, rather than 

 two-phase convection. But in the union of the oxygen of 

 crust vesicles, with the combustible gases of viscous lava 

 vesicles, to induce violent reactions which expel great quan- 

 tities of burned and unburned 2:as vertically, the hard 

 crusts must be comminuted, and the melt must be largely 

 robbed of its vesicles. The burned gases escape in large bubbles, 

 while the local heating and lowering of viscosity favor expan- 

 sion and escape of the small bubbles. Therefore loss of gas 

 plays an important part in shrinking the heated lava of the 

 grottoes and sinkholes. The expansion of the combustion 

 products, moreover, is a cooling effect. All of this may 

 account for the lowering of the level in the sinkholes, when 

 extraordinary oxidation is maintained by the cascading process. 

 Furthermore, the faster the shrinkage within a sinkhole, the 

 heavier the cataract resulting and the more voluminous will 

 be the direct suction of air downward. This in turn increases 

 oxidation and so the process of acceleration is self-propagating 

 to the limit of combustible gases available. 



Multiple FountaUiing of 1910 to 1912. 



Before leaving the question of fountaining, mention must be 

 made of the very remarkable distributive fountaining, gener- 

 ally accompanying subsidence, in Halemaumau lake in 1910 

 and 1912. This may be considered a time when the lava 

 column had reached its highest level and was entering upon a 

 general decline that reached its lowest in 1913. There was 

 a sudden spurt to nearly the 1910 level in December-January, 

 1911-12, and there was another rise to a lower level in July, 

 1912. The characteristics of this extreme activity were abund- 

 ance of bubble fountaining and numerous large fountains, 

 rushing and streaming surface currents in changing directions, 

 and frequent lines of traveling fountains when two or more 

 such currents would meet and oppose each other. In July of 



