BY THE CONDITIONS OF THEIR EXTRUSION. 5 



speed of ejection if continued for a month, would set free a mass of gas equal to the atmo- 

 sphere covering over fifty thousand square miles of the earth's surface. The 

 actual weight of gases thrown out in a month of the most vigorous eruption, must in many 

 volcanos exceed one thousand million tons, or a weight greater than half a cubic mile of 

 ordinary volcanic rocks, or a large fraction of the mass of such a cone as ^tna. This esti- 

 mate, which being only approximative is not worth exhibiting in detail, makes it clear that 

 the volume of the escaping gases and their power is sufficient to propel the lava from the 

 depths where it is found to the surface. 



2. The gases erupted from a volcanic cone, and to a certain extent the lavas, are prin- 

 cipally derived from a great horizontal distance from the point of escape. 



This proposition seems to me to stand on a tolerably sure footing. It is clear that the 

 gases which escape from an ordinary volcanic cone during its formation would, even in 

 their solid form, occupy a much greater mass than the cone itself. Moreover, we know 

 that these gases are mainly the gases of water, and that compact sedimentary I'ocks cannot 

 contain more than about two or five per cent, of this fluid. To find a source for the water 

 which escapes from a volcanic cone we must assume it to have been derived from a great 

 horizontal area about the cone. Assuming that in the case of Vesuvius the stratified rocks 

 whicli give rise to the gases are limited to a depth of one or two hundred thousand feet (a 

 conclusion to which we are led by the relatively low temperature at which the lavas are 

 extruded), then we must believe that a part of the supply of gas is derived from distances 

 of hundreds of miles horizontally from the vent. K we reckon the average diameter of 

 the crater in all its history, at one half a mile, which is probably much within the facts, 

 and assume that the whole time iu which the crater has been discharging gases at a high 

 pressure since its beginning to have amounted to an aggregate of only five years, pro- 

 bably a very small estimate, then it would have discharged in vapor nuich more water than 

 could be contained in the rocks over an area of something like forty thousand square miles. 

 So if we assume that the gases of volcanic eruptions are principally of water, and that 

 this water was contained in the rocks as it is ordinarily contained there, then we must 

 admit that the feeding ground of a volcano extends over a very wide area. 



The attitude of the rocks about a volcano comes in to support this conclusion in a very 

 striking way. While in certain cases there is a subsidence of the beds to be noticed very 

 near the crater, the general level of the region about the crater, even the largest, 

 has never shown any distinct evidence of subsidence. When we remember that the 

 cones of a volcano are the seats of a very rapid erosion, owing to the high angles 

 of the slopes, the incoherent nature of the materials, their generally low specific gravity, 

 and the torrential rainfalls that accompany great eruptions, causing the cones to 

 wear down at an average rate of many feet in a thousand years, and also take into 

 account the vast bulk of the gaseous emanations, it' is clear that luiless the supply 

 of ejected matter came from a great distance on either side of the volcano, we wovdd 

 not find this absence of sinking about the cone. 



Moreover the well determined interaction between certain volcanos hundreds of 

 miles away from each other, shows that the gases must have this horizontal 



