MAGMATIC DIFFERENTIATION OF IGNEOUS ROCKS 623 



forsterite to 94.5 per cent fluid. With continued heating the 

 forsterite dissolves in the course of time, so that all is transformed 

 into fluid at 1577°. 



On cooling a melt of MgSiOa there first crystallizes (at 1577°- 

 1557°) a little forsterite, which on continued (and slow) cooling is 

 resorbed under new formation of solid MgSi03 (at high tempera- 

 ture clinoenstatite) . 



The inversion point, at a pressure of one atmosphere, between 

 clinoenstatite and enstatite amounts to about 1100°. The point 

 cannot be settled exactly. The above reference to the treatise of 

 Andersen and Bo wen is for pure MgSiOa (^t one atmospheric 

 pressure) . 



For (Mg, Fe) SiOj, (Mg, ^^) SiOj and FeSiOj the following must 

 be taken into consideration: 



Mg2Si04 has melting point 1890°. 



FezSiO^ has melting point about i loo**. 



For the intermediate mixtures there are intermediate melting 

 point intervals (Fig. 22). 



Fe2Si04 and Mg2Si04 with relation about in the middle between 

 Mg and Fe, has a considerably lower melting point interval than 

 1550°, and consequently cannot be separated in the solid face at 

 the first mentioned temperature from a (Mg, Fe) SiOj melt. This 

 makes it probable that the splitting by one atmosphere's pressure of 

 solid pure MgSiOj into olivine-mineral and fluid may be transferred 

 to (Mg, ^^) Si03 with only a little Fe, but not to metasilicate with 

 predominant Fe or with a middle relation between Mg and Fe. 



Clinoenstatite (respectively clinobronzite) , as well known, 

 sometimes occurs in meteorites, but has never been determined 

 certainly in terrestrial igneous rocks. If the metasilicate in the 

 rocks originally had existed in the clino-modification (clino-enstatite, 

 -bronzite, -hypersthene) we might, according to the experiences from 

 the meteorites, suppose that the original mineral, in any case, 

 occasionally would still have been in existence. But while this, 

 as far as we know up to date, never is the case, the explanation 

 seems to be that the crystallization of the Mg, Fe-metasilicate 

 in the common igneous rocks — of composition between 0.08 FeSiOj, 

 0.92 MgSiOa and about 0.4 FeSiOj, 0.6 MgSiOj — always took place 



