746 JOHN JOHNSTON 



may still remain above its point of anhydrous fusion, or it may be maintained 

 fused by heat from other sources. Finally cooled, it is the familiar glassy 

 amorphous mineral. A quick release of pressure entails a quick vaporization 

 of water and a quick loss of heat. The obsidian mass, during and because of 

 the loss of water and the loss of heat, becomes pasty and "rises" like dough 

 during fermentation, and becomes pumice, which is often found overlying 

 obsidian. A quicker release of pressure from above causes the vesicular and 

 vesiculating masses to be projected, and if the vesiculation is carried far and 

 fast, volcanic dust is produced. 



The phenomena discussed in the paragraph just quoted have 

 now been reahzed experimentally with potassium silicates by 

 Morey, who was able to produce at will either a hard or a pumiceous 

 glass merely by altering the mode of cooling.'' 



What has been said about systems containing volatile compo- 

 nents may be applied to elucidate some aspects of the behavior of 

 of a cooling magma. If the crystals separating out initially from 

 a magma situated within a confined space contain none of the vola- 

 tile component, then the concentration of the volatile component 

 in the residue would become continuously greater, and consequently 

 the vapor pressure would increase; this increase might under certain 

 circumstances be large, so that there would be a considerable tend- 

 ency to enlarge the space within which the magma is confined. 

 Accordingly eruption of a magma may be correlated with a compara- 

 tively early stage of its crystallization. On the other hand, the 

 initiation of crystallization may be due to the circumstance that 

 the pressure was relieved — by faulting in the adjacent rock, or 

 otherwise — whereupon the magma began to lose its volatile com- 

 ponents, and consequently to crystallize; this process is precisely 

 analogous to the crystallization of a salt from a solution by boiling 

 off the solvent. 



If no escape is possible, the residue from the crystallization of 

 the main portion of the non-volatile constituents will be a fluid (as 

 distinct from a liquid) solution, containing silicates and probably 

 sulphides, etc., which is so mobile that it can easily penetrate the 

 adjacent rock, producing the phenomena of contact metamorphism 

 and injection. Thus we can account for the very thin veins and 

 dikelets often observed at igneous contacts, which indicate that at 



^ G. W. Morey, Jour. Am. Chem. Soc, XXXVI (1914), 226-27. 



