276 PROCEEDINGS OF SECTION C. 



as glasses, than more fluid liquids. The most familiar illustrations of 

 this law among igneous rocks are the persilicic (rhyolites) lavas, which 

 often form glasses (obsidians). The question has been raised by- 

 Crosby, and others, whether an earthquake happening when a magma 

 was in a sufficiently supersaturated metastable condition might not 

 induce crystallisation of some of the constituent compounds. 



Crystallisation may begin with different degrees of supersatura- 

 tion of the liquid, and would proceed at diSerent rates according to 

 the degree of supersaturation, being more rapid the greater the con- 

 centration. It would also be more rapid the greater the molecular 

 diffusibility, that is, the lower the viscosity of the liquid, and the 

 greatei* the rate of cooling, so long as this does not increase viscosity 

 too rapidly. Gradual or slow crystallisation at comparatively few 

 centres would yield relatively few, large crystals; whereas sudden, 

 rapid crystallisation from many centres would produce many small 

 ones. 



High fluidity in solutions would permit easy diffusion of separat- 

 ing molecules toward crystallising centres, favouring the growth of 

 relatively large individuals. High viscosity would retard diffusion and 

 favour the growth of many small crystals. This is well illustrated by 

 the laboratory experience of Day with the crystallisation of various 

 lime-soda-feldspars. Thus it was found that 100 grams of liquid 

 anorthite crystallised completely in ten minutes to fair-sized crystals, 

 and it required quick chilling to prevent its crystallisation and to 

 produce glass. A mixture of equal parts of anorthite and albite (Ab^ 

 An^) required a gradual cooling extending over several days to effect 

 complete crystallisation. Whereas liquid albite could not be induced 

 to crystallise through days of cooling in an open crucible. Comparing 

 the size of the crystals of anorthite produced in ten minutes with 

 those of oligoclase-andesine (Abg Auj), which were produced by gradual 

 cooling through two days, the fonner were from 3 to 5 mm. thick, 

 the latter about O'OOo mm. thick. That is, the more liquid anorthite 

 produced crystals 1,000 times as thick in about one three-hundredth 

 the time. An apparent ratio of 300,000: 1. 



The rate of separation of solid from liquid also depends on the 

 solubility and the amount of any substance in solution. The greater 

 each of these factors the more rapid the rate of crystallisation and 

 tlie larger the ciystals, other things being constant in compared cases. 

 This law has been expressed definitely by von Pickardt, as follows : — - 

 " The velocity of ciystallisation (separation in solid phase) is 

 diminished by the addition of foreign substances to the liquid phase 

 of a substance, the diminution of the velocity being the same for 

 equimolecular quantities of all substances." 



The order of succession in the separation of different kinds of 

 minerals from molten magma is a subject upon which there has been 

 some difference of opinion among petrologists. It has been clearly 

 demonstrated that the order is not an invariable one. The laws relat- 

 ing to the order of separation of solids from mixed solutions have been 

 definitely determined for solutions of various compounds in one 

 another, and the general principles are applicable to the study of 

 igneous rocks. The attention of petrographers has been called to 

 these laws by Vogt. 



