WATER AND VOLCANIC ACTIVITY DAY AND SHEPHERD. 285 



study of the material collected in May, the determination of the exact 

 proportion of water to the other gases present must await another 

 favorable opportunity. It may perhaps be added that a complete 

 equipment for another attempt lies ready at the laboratory of the 

 Volcano Research Association on the crater rim, but the lava lake 

 disappeared completely from view soon after the December descent 

 was made and has not again reappeared. 



Although the continuation of the field studies must await the 

 gracious pleasure of the most fickle of goddesses, it need not delay the 

 prosecution of the laboratory study of the relations between the gases 

 already found or the preliminary discussion of the results thus far 

 attained. Moreover, in the discussion which follows, evidence will be 

 offered that the composition of the gases varies within considerable 

 limits, so that the precise proportions of the gases which go to make 

 up the exhalation at any particular moment may prove to be of less 

 importance than was at first believed. 



CHEMICAL. SrrUDY OF THE MATERIAL COLLECTED. 



From a physicochemical viewpoint, the study of volcanic activity 

 centers first on the nature of the participating ingredients, then on 

 the condition of equilibrium or the progress of the reactions taking 

 place between them, as the case may be. At the time of our two 

 visits all the three states of matter — gaseous, liquid, and solid — were 

 found represented. Gases were emitted constantly in great volume, 

 and displayed nearly all the great variety of cloud forms which have 

 been so frequently described in volcano literature except the violently 

 explosive type, which has been rarely or never seen at Kilauea since 

 the advent of the white man (1820). There was a liquid lava basin 

 of oval shape some 600 by 300 feet, inclosed by a lava dike or rampart 

 built up from the surrounding floor of the basin by the tumultuous 

 spattering and splashing of the lava lake (pi. 5). Both floor and 

 rampart are frequently overflowed when the lake is high, and again 

 great masses of it fall into the lake and are redissolved when it is 

 low. The floor of the pit at the time of our first descent in May, 1912, 

 had been completely overflowed but three days before and was rea- 

 sonably level. The fresh lava had solidified to a depth of some 10 

 inches and was abundantly solid to walk on, but was still uncomfort- 

 ably hot and the cracks were still glowing. 



Surrounding this floor are the walls of the pit, some 200 feet high 

 at the time of our first descent and made up of the exposed edges of 

 successive earlier overflows (pi. 6), which indi^adually rarely ex- 

 ceeded 2 or 3 feet in thickness. The (Halemaumau) pit as a whole 

 was about 1,500 feet in diameter, roughly circular in plan, and with 

 nearly perpendicular walls except for the talus pile at the base, which 

 extended about half way up the wall. All these dimensions vary 



