278 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1913. 



(4) As the cloud rises past the rim of the crater on the leeward 

 side, the walls about the crater, being comparatively cold, should be 

 wet with the condensed vapor, Avhereas in fact these walls remain 

 quite dry. 



(5) A train of glass tubes was lowered over the rim of the crater 

 for a few yardg on the side where the cloud was emerging, and 

 through these tubes (some 250 feet distant from the nearest liquid 

 lava, it may be remarked) air and the vapors carried by it were 

 pumped for several minutes, but no trace of condensed moisture 

 appeared on the inside walls of the tubes. Examination with a 

 hand iens revealed the fact that the tube walls were quite thickly 

 covered with crystallized salts, some of which were stated to be 

 hydrates or to be hygroscopic, but this was deemed to be due to 

 original moisture ( !) carried on the tube wall before the beginning 

 of the experiment. No analyses of the gases or of the solid salts are 

 given. 



(6) A dew-point hj-grometer carried along the rim through the 

 smoke cloud showed a lower humidity within the cloud than in the 

 clear air just outside of it. 



Before proceeding to recount our own experience with these phe- 

 nomena, it may be as well to express our belief that nearly all of 

 these observations, both of Green and Brun, may be perfectly true as 

 recounted above, and still offer no proof that the volcano exhales no 

 water vapor. 



THE EXPLANATION OF THE VOLCANO CLOUD. 



Green's observation that the great white cloud appears but inter- 

 mittently may be explained by a somewhat closer observation of the 

 conditions of formation of the cloud without assumptions of any 

 kind about its possible water content. For example, we noted, dur- 

 ing several months of constant observation, that the visible cloud 

 does not rise directly from the surface of the liquid lava, but rather 

 from cracks in the inclosing banks,^ shattered, as they always are, 

 by alternations of heat and cold as the liquid lava rises and falls in 



iCf. plates 6, 7, 8, and 11. 



Observations confli-matory of the conclusion that the smoke cloud when present does 

 not rise from the liquid lava, but from the shattered floor and talus surrounding the 

 basin, have been recorded by other writers. 



For example, Prof. W. T. Brigham, director of the Bishop Museum,, Honolulu, who 

 for 50 years has been one of the most careful observers of volcanic phenomena in the 

 Island of Hawaii, writes as follows: (Volcanoes of Kilauea and Mauna Loa on the Island 

 of Hawaii, Honolulu, 1901») : 



Page 28 : " * * * It should be noticed how small the supply of steam in the active 

 outpour of Kilauea really is." 



Page 28 : " When the pit is empty of molten lava, the smoke is often most abundant."' 



Legend to plate 4r) : " Lava pool below the rim of Halemaumau. « * * Little 

 vapor rises from the portion which is active." 



Legend to plate 50 : " There is little escape of steam from the lake surface." 



William Lowthian Green (Vestiges of the Molten Globe, vol. 2) writes (p. 170) : 

 " Smoke, vapors, and gases seem to arise from the orifices of eruption and orifices in the 

 neighborhood of molten lavas on Hawaii, and not from the lavas themselves." 



