STUDY OF IGNEOUS HOCKS. 281 



bulk of the whole ; by the common intrusion of magma along fracture 

 planes and along those of structural weakness, rather than at random 

 through rock masses; and most conspicuously, by the evidence of 

 equilibrium with the atmosphere maintained by lava in volcanic 

 craters. Open vents are known to exist for centuries without great 

 extrusion of rock magma, as at Stromboli. The stresses which produce 

 condensation of volume in proportion to depth, and the results of 

 expansion of volume, are, therefore, somewhat evenly balanced. 



The effect of expanding gases is shown in the explosive character 

 of many eruptions, and the periodic character of all eruptions from 

 open vents (volcanoes). It must increase the volume of all magmas 

 as pressure is relieved. Its effectiveness must increase with the 

 amount of gas in the magma, Avhich may result from diffusion of gas 

 from greater depths of magma, and also from accession from adjacent 

 rocks under favourable conditions. 



Spasmodic eruption may follow sudden yielding of overlying rocks 

 to long-continued stresses, as in the case of massive, or fissure, erup- 

 tions when there may have been no considerable explosive action of 

 gas; or it may result from an accumulation of gas pressure sufficient 

 to rupture overlying rock masses. Eruption is then accompanied by 

 abundant evidence of explosion. Both causes undoubtedly operate 

 tcgether in most cases. 



In so far as magmatic eruption is a result of volumetric expansion 

 of the magma, due to relief of pressure, the shrinkage of volume due 

 to cooling will retard eruption, or eventually stop it. Ciystallisation 

 will operate in the same direction. In proportion as eruption is due 

 to expansion of dissolved gas, the escape of gas from magma, or the 

 reduction of supply, will lessen the force of eruption, or eventually 

 put an end to it. The supply of gas from great depths may be reduced 

 by the gradual diffusion of whatever is in a position to be appreciably 

 diffused; or the supply from rocks adjacent to intruded magma may 

 be cut off by the closing of pores in these rocks through meta- 

 niorphism ; poi'ous rocks becoming dense and almost impervious to 

 gases. In these ways eruptive action initiated by crustal readjust- 

 ment after continuing for variable periods may come to an end: 

 Keadjustraent of stresses may recur from time to time in any region, 

 either at such widely-remote periods that the volcanic activities 

 associated with each readjustment constitute distinct and separate 

 periods of action ; or at such frequent intervals that the results of 

 several profomid movements are combined to fomi a prolonged period 

 of complex volcanic eruptions. 



Independence of action at neighbouring volcanoes, either as to 

 pei'iod of eruption, volume of magma erupted, explosive or quiet 

 character of action, or relative height of lava column in conduit of 

 volcanoes, follows from local variation in the factors entering into 

 the process of magma eruption, such as : the volume of magma in- 

 volved in each conduit extending to profound depths, the shape of the 

 conduit, the temperature of the magma, the rate of cooling, the 

 amount of gas diffused in any given time, the character of the sur- 

 rounding rocks, and the stability of the surrounding rock mass as a 

 complex whole. The chemical composition of the magma is also a 

 factor involved in the activity of a particular volcano. But the 

 composition of the magma is also a feature by which volcanoes may 



