156 Jounston-Lavis—The Eruption of Vesuvius in April, 1906. 
these large bubbles bursting at the surface of the fluid rock, coarse lava spray is 
detached and ejected, and constitutes the lava-cakes that are constantly scattered 
around the vent by the explosion (bursting of gigantic gas-bubbles), The 
collection of these lava-cakes composes the principal materials around the vent, and 
builds up the cone of eruption. While the outer slope of this cone of eruption is 
usually that of the angle of repose of such materials, the inner face is vertical and 
the lava-column is sufficiently high to support it, or solder its constituents together. 
When a radial rift splits the cone, as so frequently occurs, there will be a sudden 
fall of the lava in consequence of its filling such a cleft. Still more, if that cleft 
reaches the surface of the cone at some spot, and fluid rock issues, there will bea 
still greater fall of this central lava-column. The support of that column will be 
removed from the inner surface of the chimney; and the cone of eruption and the 
edges will begin to crumble in. This destructive action will be further accelerated 
by the escape of vapour from the lowered surface of the lava. In fact, I shall 
show presently that the amount of vapour in such a case is even increased. The 
escape of these gases under some constraint in the now empty portion will add to 
the degradation of the walls, which continue to crumble in, gradually excavating a 
conical hollow which we call the crater. 
If the lava-surface has fallen very low, the explosions of gas may be insufficient 
to eject lava-cakes beyond the new crater edge; but the crumbling materials will 
be ground and more or less pulverized and ejected, so constituting the dust- 
forming stage with the expulsion of these accessory ejecta. There are all 
gradations between the cone and crater-forming stages, as there are between the 
ejection of essential materials, as lava and scoria cakes, or of old broken-up 
fragments and dust which constitute the accessory ejecta. 
By the constant ebullition of lava in the chimney we have two sources of 
the vapour. Part will be derived from the upper lava itself, which it had 
acquired at greater depths and brought up with it, while other vapour is no 
doubt the excess of that which the magma deeper down in the canal continuously 
evolves, and which would rise through the upper mass. Provided the output is 
slow, then there will be a tendency, by the increase in viscosity of the upper 
portion of the lava column, to augment the resistance to the escape of vapour 
from greater depths. Then the tension of the magma will increase in the deeper 
parts. A struggle will then take place between the increasing viscosity of the 
upper portion of the column which might go on to complete solidification, and the 
rising tension of the magma deeper down. If sucha state of things has been going 
on for some time—we will suppose not to the extent of solidification of the upper 
part of the column—and if by the formation or reopening a lateral outlet takes 
place, and most of the obstructing plug is drained out, removing the pressure from 
that of the high tension magma below, it is obvious that violent vesiculation and 
