IGNEOUS ROCK-MAGMAS AS SOLUTIONS 243 



That a rock-magma is to be regarded as, in some sense, of 

 the nature of a complex solution was suggested by Bunsen so 

 long ago as 1861. At that time it was not possible to attach 

 any very precise significance to the comparison of rock-magmas 

 with ordinary solutions of salts ; for the properties even of the 

 latter were only imperfectly known, and the fundamental laws 

 of crystallisation from saline or other solutions were yet to be 

 discovered. The rapid advances since made in physical 

 chemistry have placed the question on a different footing; 

 Lagorio, Teall, Morozewicz, Vogt, and others have discussed 

 rock-magmas from this point of view, and it is generally agreed 

 that the only obstacles to framing a complete theory of the 

 crystallisation of igneous rocks on the solution hypothesis are 

 those practical ones which arise from the intractable nature of 

 the silicate and other minerals of which these rocks are composed. 

 The high melting-points, extreme viscosity, and other specific 

 properties of these minerals render the direct experimental 

 method a matter of much difficulty. Before noticing what has 

 already been accomplished in this direction, we may point out 

 some of the evidences from the petrographical side which seem 

 very decidedly to assimilate the crystallisation of rock-magmas 

 to that of those solutions which can be more easily studied. 



In the first place, it has long been recognised that the 

 freezing-points of the rock-forming minerals are not the same 

 in magmas as they are for the minerals tested separately. This 

 is sufficiently evident from the fact that the order in which the 

 several minerals crystallise from a magma is not in general 

 the inverse order of their specific fusibility. For instance, the 

 melting-points of the alkali-felspars, and probably also of the 

 micas and the common hornblendes and augites, lie in the neigh- 

 bourhood of 1 200- 1 2 50 C. For quartz experimenters have 

 given various estimates, mostly about 1700-1800 , while, 

 according to Allen and White, the stable form of silica at 

 high temperatures is tridymite, melting at about 1600 . Never- 

 theless, quartz is usually the last mineral to crystallise in a 

 granite. Before the conception of rock-magmas as solutions 

 had been realised, this fact appeared anomalous ; and discussion 

 of it usually took the line of suggesting that granite is not a 

 true igneous rock, but results from some rather vaguely defined 

 " igneo-aqueous fusion." It cannot indeed be doubted that 

 the presence of water is an important circumstance ; but this 



