500 Bowen and Andersen — Binary System MgO-SiO„. 



Hawaii, but in that case we are probably dealing with a quite dif- 

 ferent matter, perhaps with the decomposition of silicates and 

 the setting free of silica as cristobalite by the action of vapors. 

 A mixture of total composition MgSi0 3 98 per cent, SiO„ 2 per 

 cent, can be cooled in such a manner that it consists of forsterite 

 and glass. We have then what might be termed an olivine- 

 bearing lava in simplified form, though the actual total compo- 

 sition of the mixture shows no olivine but an excess of free silica. 

 This simple case emphasizes the fact that, in spite of the great 

 usefulness of a purely chemical classification of rocks, a classifi- 

 cation in terms of the actual phases (minerals) present is 

 extremely desirable because these phases are an expression of 

 equilibrium under the conditions of their formation. 



Summary. 



Equilibrium in the binary system MgO-Si0 2 was studied by 

 applying the method of quenching. 



There are two compounds, the orthosilicate Mg 2 Si0 4 and the 

 metasilicate MgSi0 3 , capable of existing in contact with liquid 

 in the binary system. The former crystallizes in a form 

 corresponding with the mineral forsterite and the latter forms 

 crystals similar to enstatite in most properties but of monoclinic 

 symmetry, clino-enstatite. 



Clino-enstatite is the only stable form of MgSiO s encountered. 

 It has no true melting-point but breaks up at 1557° (formerly 

 considered the melting-point) into forsterite and liquid and 

 the temperature must be raised to 1577° before complete 

 solution of the forsterite takes place. 



In an earlier publication from this laboratory, crystals 

 termed a-MgSi0 3 were described as a high-temperatnre form 

 of magnesium metasilicate. They were considered to be the 

 product of inversion of clino-enstatite (/3-MgSiO s ), but the 

 crystals described have now been proved to be a product of the 

 dissociation at 1557° and to be the orthosilicate, forsterite, not 

 a form of the metasilicate. 



On account of the break up of clino-enstatite into forsterite 

 and liquid there is no eutectic between the two compounds, 

 and the liquids show, on cooling, the partial or complete 

 re-solution of forsterite at the reaction point, 1557°, the liquid 

 reacting with the forsterite crystals to give clino-enstatite. 



A discussion is given of the geological significance of this 

 resorption of the olivine, forsterite, by reaction with the 

 liquid to give the pyroxene, clino-enstatite. 



Geophysical Laboratory, 



Carnegie Institution of Washington. 



