source of the water which, converted into steam, takes such a conspicuous part in volcanic eruption is open to several objections.
In almost all land areas the rocks below the surface are saturated with water, the source of which is mainly rain. Excepting that the pressure of the sea on its floor tends to force water into the rocks beneath, there does not seem any good reason for concluding that the earth's crust where covered by the sea is more highly charged with water than the portions beneath land areas.
Another argument for the presence of sea water in volcanoes is that after eruption the country about a volcano is sometimes whitened for many miles with salt, and also that some of the vapors arising from volcanic vents are such as might be expected to occur if the substances contained in sea water were sufficiently heated. It is to be remembered, however, that large bodies of salt derived in some instances from the evaporation of sea water occur among stratified rocks, and also that many sedimentary deposits are saturated with saline water. It thus becomes evident that communication between the conduit of a volcano and the sea is not the only means by which saline water can come in contact with molten rocks.
It is well known that volcanoes as a rule are located near the borders of continents, or on the floor of the sea. This fact is more in harmony, however, with the idea proposed by Dana, that the margins of continents are determined by the location of weak belts in the earth's crust, along which maximum movement takes place, than that the presence of surface water bodies is essential to the existence of volcanoes. In support of this conclusion it may be pointed out that volcanoes of recent date occur in the Great Basin, hundreds of miles distant from the Pacific. The Great Basin is a region of faults, and as much a belt of weakness in the earth's crust as if it had chanced to be situated near the sea.
Owing to the increase of pressure with depth, it is evident that cavities in rocks in which any considerable bodies of water can be stored must become less and less frequent as the distance below the surface increases. As has been shown by Van Hise, at a depth in excess of about thirty thousand feet what may be termed appreciable cavities can not exist. Rocks under pressure become compact, so that deeply seated rocks must be less porous than similar material near the surface. These considerations lead to the conclusion that water-charged portions of the earth's crust are superficial. Hence the water given off by volcanoes in the form of steam, and probably also the gases produced by the dissociation of the elements composing water and the vaporization of the various salts it contains, must reach volcanic conduits in their upper portions. These considerations add strength to the conclu-
