278 F. A. Ferret — Floating Islands of Halemaumau. 



tively favorable latter conditions, the eruptive forces succeed 

 in establishing a vent at the earth's surface, it may well be 

 considered improbable in the former case, but the suggestion 

 conveyed by this smoking, floating island remains, nevertheless, 

 an interesting possibility. 



The eventual sinking of the islands is generally brought 

 about by a process whose mode of action is as instructive as it 

 is unlooked for and surprising. It might certainly be supposed 

 that a solid mass in free flotation would be raised and lowered 

 with the rise and fall of the level of the liquid in which it is 

 suspended, yet such is not the case here. It is but another 

 exemplification of the fact, upon which the writer has often 

 had occasion to insist, viz. that a column of active lava must 

 not be considered upon the basis of a simple liquid. It is true 

 that a steady and continued rise of the lava will float an island 

 upward, but in the case of those comparatively rapid oscilla- 

 tions of level which are frequent at the "turn of the tide*' and 

 especially so during a general subsidence of the lava column, 

 the surface of the lake may rise or fall to the extent of from 

 one to several meters without moving a large island vertically 

 by so much as a single millimeter. The reason for this is that 

 these rapid alternations in the level of the lake are due to a 

 greater or lesser supply of gas — or of gas-charged lava — from 

 below causing the surface to rise very much after the manner 

 of boiling molasses. The lava at a certain distance below the 

 surface, and in which the bulk of the island is presumably sus- 

 pended, is not much affected and the island remains stationary 

 while the rising surface emulsion overflows its shore, where- 

 upon it sinks to some extent. If, instead of rising, the surface 

 level is lowered, the island remains for a time with its line of 

 shore emergent, but if the lava column as a whole is in process 

 of subsiding, the island finally descends a little and is then 

 carried by its momentum farther than it would otherwise have 

 rested and so the next upward oscillation of the surface mate- 

 rial engulfs the shore and still further sinks the whole. The 

 seeming fact that the sinking of the island is effectuated during 

 a general lowering of the lava column after floating many 

 months on a generally rising lava receives additional explana- 

 tion from the probability that the general subsidence is accom- 

 panied — if not caused — by a diminution in the supply of gases 

 from below ; which gases, as we have seen, lend buoyancy to 

 the mass of rock. During the summer of 1911, the lava 

 column reached its greatest height in the middle of July, when 

 it stood seventy meters below the writer's station and dwelling 

 on the east brink. Rapid oscillations began on the 17th and 

 by the 22d the lake was in full subsidence. On the 31st the 

 first great engulf ment of the shore of the island occurred # by 

 which the island was sunk to the base of the solid original 



