PETROLOGICAL ABSTRACTS AND REVIEWS Age 
great size and is superheated sufficiently to allow of extensive assimila- 
tion of the overlying acid rock. It works its way up partly by piecemeal 
stoping of the roof, but a point is finally reached when a great part of the 
latter will cave in and founder in the fluid magma below. The assimila- 
tion of the surrounding rocks alters the composition of the latter, giving 
rise to a differentiation which results in the collecting of the acid phases 
at the top. The upper portions of the magma, which are usually 
liparitic, may overflow and produce a formation somewhat similar to 
a fissure eruption, but differing from it in being acid and in occurring in 
one thick flow rather than in several superposed sheets. The 600-meter 
sheet of rhyolite in Yellowstone Park is explained in this way. The 
fact that the geysers require more heat than such a flow can well supply 
is also accepted as a confirmatory indication that the rhyolite is not a 
true flow, but is merely the top layer of a great batholith which is basic 
in its lowest levels. 
The paper, however, deals chiefly with the question of central erup- 
tions, and the first problem presented is that of the opening and localiza- 
tion of the vent. If there be a pre-existent fissure it is doubtless suscepti- 
ble of enlargement by the superheated lava; this may be accomplished by 
the solution and mechanical removal of the wall rock by the lava, or 
by the melting and explosive abrasion effects of the magmatic gas. If 
the latter become segregated and explosive, a diatreme, or tube sur- 
mounted by a funnel results, and this may also form in unfissured rock. 
In other cases the magmatic gases, highly heated and under great pres- 
sure, may collect in the hollows or cupolas of the roof of the batholithic 
chamber, thereby localizing and intensifying the stoping action of the 
magma. ‘The latter may thus work the last few kilometers of its way 
to the surface unaided by a fissure. 
A large part of the remainder of the paper is based upon an analysis 
of conditions in Kilauea. Mathematical calculations prove that an 
enormous amount of heat is lost at an open vent, and that the heat lost 
by radiation is over fifty times as much as that lost by conduction. 
Ordinary convection between the main feeding chamber and the vent 
is then proved inadequate to sustain heat in the latter. The conception 
of two-phase convection is then advanced: the magma rises continually 
because of its vesiculation, and, having discharged its gas at the surface 
of the lava lake, sinks again down the pipe to the main feeding chamber. 
This is the explanation of the currents in the Kilauea Lake and of the 
Old Faithful fountains which are thought to be located directly over the 
pipe, and to be due to the ever more rapid rise of the vesiculate magma as 
jt nears the surface. Figures for the loss of density of the lava, after 
