Jaggar — Yolcanologic Investigations at Kilaiiea. 205 



1912 this process reached its maximum, when in a small oval 

 lake about 600 feet (180 meters) long, depressed 200 feet 

 (60 meters) below the rim of the pit, there were several 

 luindred large fountains and many more small ones (fig. 15). 

 All of these were playing at once with a roaring noise, intense 

 incandescence, tumultuous rush of surface currents, and bom- 

 bardment on the bank, first on one shore and then on the other. 

 - To account for this phase of activity we must suppose that 

 convectional downiiow was dominant, a shallow, air-iilled and 

 superficial convectional circulation was very rapid, and that 

 accordingly oxidation was at a maximum. This was the sum- 

 mer when the Day and Sheplierd gas collection w^as made and 

 excess hydrogen was found among the combustible gases. 

 Temperature measurement with Holborn-Kurlbaum pyro- 

 meter gave 1185° C. This was the time when Day and 

 Shepherd found greatest emission of gas (maximnm fountain- 

 ing) to coincide with higliest measured temperatures. There 

 had been repeated inward tumbles of great quantities of air- 

 filled bench rock in 1910, 1911 and 1912, the lava each time 

 recovering so as to rise and submerge this talus debris. It 

 seems likely that the excessive fountaining during the season 

 of sinking of 1912 was due to the breaking up by subsidence 

 of this submerged air-filled debris, and this, coupled with and 

 contributing to extreme rapidity of two-phase convectional 

 circulation, as outlined above, brought about an admixture of 

 much oxygen with combustible gases. Vast quantities of air 

 were drawn down directly with downfiow currents, under the 

 banks and at the traveling fountains, owing to the speed of the 

 torrents, which increased to cascades in August. It is possible 

 that at certain seasons the supply of some unstable gases such 

 as hydrogen is larger than at other times, and the heating effect 

 greater. Routine photospectroscopic work on flame spectra 

 might prove this. 



JDlsappearance of Liquid Lava during Low Levels. 



These terms of violent effervescence coincided with culmina- 

 tions of rising and the beginnings of sinking. With the low 

 levels of September, 1912, and of the entire year following 

 May of 1913, gas pressure dwindled to a minimum and the pit 

 at some TOO feet (213 meters) of depth was floored with frozen 

 lava surrounded by talus slopes. There was much sulphur 

 fume and water vapor, but even the hissing of confined gas at 

 times ceased completely. This condition, merely because the 

 lava column had retired topographically to a depth one-sixth of 

 the height of the mountain above sea-level, could hardly be 

 attributed to depression alone. Why, in other words, should 

 depression induce inactivity? Why should not the same 

 violent fountaining, and vast discharge of gas and heat, go on 



