Jccggar— Yolcanologic Investigations at Kilaiiea. 183 



lakes of Kilauea are truly shallow pools, with feedino; conduits 

 beneath of small size. Where these lakes are confined in the 

 cylindrical crater of Halemaumau and within it are surrounded 

 by or include benches and crags of solid lava developed by 

 overflow and accretion, these latter are merely subaerial exten- 

 sions of the semi-solid lava column, which forms the bottom of 

 the lake, and through which feeding conduit tubes are main- 

 tained open. This lava column is incandescent within, graded 

 in viscosity from high to low in transition to the liquid lake 

 through its bottom, from high to solid in transition upward to 

 the crags, and probably from high to lower viscosity in depth 

 w^here the convectional circulation must become adjusted to a 

 uniform and relatively slight radiation outward into the retain- 

 ing walls. 



It might be asked, how do the hmer bench overflows, as a 

 part of and contributory to the upward growth of this non- 

 liquid lava column, differ from any other surface lavas? The 

 answer is that they are surface layers of a continuously mobile 

 body ; that by their weight they subside and displace the mobile 

 incandescent matter into which they grade downward ; away 

 from them this semisolid paste actually pushes upward locally 

 so as to maintain a balance among islands, crags, benches and 

 lake bottoms ; there is hence a slow circulation, itself essen- 

 tially convectional, in the matter which supports the liquid 

 lake, in addition to the much more rapid convection continu- 

 ously stirring the lake itself. Where such overflows become 

 buried for hundreds of feet of such compensated subsidence, 

 under other flows of the same kind, remaining integral parts of 

 the lava column and reheated within it, in marked contrast to 

 the older rock Avails adjacent, they must be considered a part 

 of the live lava column and of the mechanism of its circulation, 

 and quite distinct from surface lavas which become frozen and 

 dead. (See description of collapse of June 5, 1916, above.) 



jSum?nary, Duplex Lava Column. 



I have attempted above to develop proof that there are two 

 ])hases of the mobile lava cohimn generally visible within 

 Halemaumau, the one apparently liquid, the other apparently 

 solid. Investigation has shown that the lifjuid lake is largely 

 a gas-charged "froth, and that the rocky benches are only a 

 surface hardening or crust upon a stiff but mobile and incan- 

 descent column of lava. The liquid phase maintains a shallow 

 saucer, fed by a few spring holes beneath, in the top of the 

 serai-solid phase. Both phases are in circulation. Heavy crusts 

 founder with much effervescence in the liquid lake while quiet 

 upwelling of lighter lava goes on on the opposite side, and sur- 

 face streaming" pours from the locus of upwelling to the locus 

 of foundering. In the case of the semi-solid bench lava, over- 



