F. A. Perret — Floating Islands of Ilalemaumau. 281 



demolition — and the entire structure soon collapsed and sank 

 out of sight beneath the now nnbroken surface of the fiery 

 lake. 



"I have no doubt that the explanation of this phenomenon is 

 as follows : The island was kept from sinking by the large, 

 flat area of black crust and, on the breaking up of this, it 

 began to sink. In so doing, a quantity of gas was evolved 

 when the hot lava covered the rock and this blew the great 

 cylindrical bubble which then continued to support the island, 

 although below the surface, for the final tour of the lake. On 



Fig. 5. 



Fig. 5. The great inflated cylinder of black glass which formed over the 

 sunken island, Aug. 18, 1911. 



the demolition of the great bubble the entire mass sank below 

 the surface." 



To this should be added that the cylinder (fig. 5) was evi- 

 dently inflated and of great buoyancy, as shown by its almost 

 wholly emergent flotation. The material was the black, 

 vitreous rock which results from a rapid cooling of the lake 

 lava. That the original island remained beneath the floating 

 cylinder was proved by its partial emergence for a few 

 moments (when it was also photographed) during the passage 

 across the lake. The point of rock came up as clean and dry 

 as a piece of wood out of mercury, showing that, while too hot 

 to permit of the liquid lava's remaining and solidifying upon 

 it, it had not yet reached the state of incandescence and there- 

 fore could not amalgamate with the lake material. This was 

 after nine days of immersion, but always near the surface. 



Of the island itself nothing more was seen, but on August 



