l.'iS F. A. Perret — Lava Eruption of Stromboli. 



flowed for hours over snow without melting its way through 

 it, and the masses of liquid lava thrown from the craters upon 

 the snow fields remained virtually upon the surface and later 

 formed "alpine mushrooms" by remaining poised on snow 

 pedestals as the rest of the held disappeared under the sun's 

 ra}'s- — a merely hot rock will rapidly melt its way to the bottom 

 of the snow. At Sakurashima, the water was most heated by 

 the lava where this entered the sea in a tumble of hot blocks — 

 in the case of a liquid stream there was little heating, as has 

 already been shown. 



We may say, therefore, that a flowing lava may exist in con- 

 tact with water without the disintegration of either, thanks ,to 

 the formation of a protective sheath, and this fact helps us to 

 understand the quiet growth of submarine volcanoes. In such 

 cases the only surface commotion need be that due to true gas 

 emission at the central vent. 



In point of fact, a sub-aqueous lava stream comports itself 

 more decorously than a similar sub-aerial one. This is due to 

 an important fact which should here be mentioned. It is not 

 to be supposed, of course, that the protective sheath absolutely 

 and always prevents the entrance of water into contact with 

 hot material. Cracks must form at its outer surface and a 

 little water enter and be vaporized in the act of sheathing the 

 raw places. But that which is thus evolved is simply the 

 vapor of water and this, in the presence of water in mass, con- 

 denses to water again — there is nothing to reach the surface 

 and cause ebullition. At Stromboli, when incandescent masses, 

 detaching from the lava stream, rolled off the beach into the 

 sea, those of which a portion remained projecting above the 

 surface would steam copiously for a considerable time, while 

 others, precisely similar but completely submerged, gave no 

 surface indication of their existence. 



The Explosive Phenomena. 



In contradistinction to that of 1891 the recent eruption was 

 not characterized by important precursory explosive phenom- 

 ena. There was, apparently, an emission of ash preceding the 

 outflow of lava, but no sample of this could be obtained, and 

 the emission was unimportant in point of quantity. But, 

 toward the end of the eruption, on the 18th and 26th of 

 November, two very violent explosions opened the conduits of 

 the crater which had evidently become obstructed by collapse 

 of the walls as a result of the falling lava column. The first 

 of these explosions, occurring a few hours after the first com- 

 plete cessation of the lava flow, seemed a direct result of this 

 new condition — "post hoc, ergo propter hoc." The second 



