IGNEOUS ROCK-MAGMAS AS SOLUTIONS 241 



rocks must be ascribed, is a differentiation, not of free oxides, 

 but of silicate and other compounds identical or closely com- 

 parable with the known rock-forming minerals themselves. We 

 seem to be justified, therefore, in concluding that when a mineral, 

 such as olivine, crystallises from a magma, it is not formed by 

 the union of the oxides, magnesia and silica, in the act of 

 crystallisation, but already existed as magnesium orthosilicate 

 in the fluid magma. The only oxides which existed in the free 

 state in the magma, at least near the point of crystallisation, are 

 those parts, if any, of the silica, alumina, ferric oxide, and water 

 which appear in the solid rock as quartz, corundum, haematite, 

 and free water. For the rest, the magma is a mixture of various 

 silicates, with a small amount of other compounds, such as phos- 

 phates and sulphides, known as constituents of igneous rocks. 



It does not follow that the compounds which exist in a rock- 

 magma at high temperatures are necessarily identical in general 

 with those represented by the familiar rock-forming minerals. 

 Indeed there are several considerations which induce us to 

 qualify such conception in some important particulars. In the 

 first place, many of the common minerals of igneous rocks belong 

 to isomorphous series. It is not to be supposed that such a 

 mineral as labradorite, which is an intermediate member of such 

 a series, exists as such in a rock-magma. The two end-members 

 of the series exist there separately, and the proportions in which 

 they are represented in the labradorite crystals are not the same 

 as the proportions in which they are present in the magma. A 

 like consideration affects other well-known groups of minerals. 

 Secondly, important examples are known, and possibly others 

 are to be discovered, of dimorphism, and even polymorphism, 

 among the silicates and other rock-forming minerals. 1 One 

 form may be stable below a certain temperature and another 

 form above that temperature. If the inversion-point falls within 

 the range of magmatic temperatures, the transformation may 

 take place in the magma ; and it must do so, when the proper 

 temperature is passed, if we can assume that equilibrium is 

 always attained. Of the two principal forms of silica, quartz is, 

 according to Allen and White, stable below 8oo° C., and tridymite 



1 See especially Day and Shepherd, " The Lime-Silica Series of Minerals," 

 Amer. Journ. Set'., (4) vol. xxii. (1906) pp. 265-302 ; Allen, Wright, and Clement, 

 " Minerals of the Composition MgSi0 3 ; a Case of Tetramorphism," ibid., pp. 

 385-433. 



