276 F. A. Perret — Floating Islands of Halemaumau. 



into contact with the liquid, basic lava at high temperature, 

 chemical reaction ensues with evolution of gas from all points 

 of the surface. This action is certainly effective in the reflo- 

 tation of a sunken mass and may be compared with that seen 

 when chips of zinc are dropped into a beaker of dilute acid 

 and the evolved gas, clinging to the bits of metal, easily raises 

 them to the surface. 



The great buoyancy of gas bubbles in the liquid lava is 

 amply demonstrated by the formation of imposing fountain 

 domes, as described in a previous paper on that subject. 



As to the re-fusion of the mass of the island, it will be 

 evident that, under the normal condition of activity, the tem- 

 perature of the immediate surface of the lake will be insuffi- 

 cient to accomplish this and instead, as we have seen, it is the 

 lake material itself which tends to solidify in contact with the 

 floating mass. But the prism of rock undoubtedly extends to 

 a considerable distance below the surface, where the tempera- 

 ture and the chemical activity of the lava are much greater, 

 and there the contact surfaces must be slowly eaten away. This 

 action will be accelerated if, as frequently occurs, the mass of 

 the island intercepts a path of upstreaming bubbles of juve- 

 nile gas whose temperature and chemical activity are very high 

 and whose mechanical and fountain-forming powers as well 

 will be directed against the opposing rock. There are also 

 reasons for believing that certain portions of the rock act cata- 

 lytically, facilitating the evolution of gas bubbles from the 

 lake lava, the island thus conforming itself into a veritable 

 focus of activity. 



By some such means it would seem that a large cylindrical 

 cavity was eroded vertically upwards from the bottom of the 

 submerged mass to a point under the saddle between the two 

 principal portions of the island and this became the seat of 

 what was undoubtedly one of the most remarkable phenomena 

 ever observed in this place of wonders, but of which the dis- 

 cussion must be deferred to a subsequent paper on the circu- 

 latory system of the lava lake. It may suffice to state here 

 that the cavity was open at the surface level on either side of 

 the saddle in the form of a grotto (fig. 2) within which fountain 

 action produced a lower level of lava, as explained in the pre- 

 ceding paper on Lava Fountains. The result of this was an 

 inpouring of the outside, surface liquid which still further 

 lowered the inner level until a magnificent, free flowing, 

 double cascade was formed and maintained for hours at a time. 

 When it is realized that the surface lava was plunging down- 

 ward from the level of the lake into a void in the submerged 

 portion of an island which was, itself, floating freely in the 

 same lake, it will be admitted that we have here a sufficiently 

 interesting subject of investigation. 



