624 J- S. L. VOGT 



for the existing pressure at a temperature lower than the inversion 

 point between the clino- and the ortho-modification of the meta- 

 silicate. This view is strengthened by the fact that in the rocks 

 where the metasilicate has crystallized at an early stage it shows 

 the crystallographic contour of the orthopyroxene. 



It is probable that the inversion temperature between the clino- 

 and the ortho-modification rises not quite inconsiderably with the 

 pressure. The difference in density between clinoenstatite and 

 enstatite is indeed very little, but if the inversion heat is minimal, 

 as only a small fraction of i cal., the inversion temperature will 

 nevertheless increase not inconsiderably with the pressure. In 

 the common anchi-eutectic norites, etc., the crystallization interval 

 for the most part lies at about 1275-1200°, which must be lower 

 than the inversion boundary at high pressure between the two Mg, 

 Fe-metasilicate stages. In bronzite rocks and bronzite-carrying 

 olivine rocks a not inconsiderably higher crystallization interval may 

 be presumed, but even in these rocks we cannot observe any struc- 

 tural indication that here originally occurred a clino-modification. 



By melting MgSiOj at 1557° to a liquid in connection with solid 

 olivine— the latter only in a small quantity — and at 1577° only to a 

 liquid, Vi — Va will, in accordance with what happens in all previously 

 examined silicates, doubtless be positive. That is to say, the 

 "incongruent" melting point of MgSiOs, in the same way as of 

 other silicates, must be supposed to rise somewhat with the pres- 

 sure. The splitting of MgSiOj to an olivine (very poor in FeO) in 

 the igneous rocks might thus have taken place at a temperature 

 somewhat above 1557°. But the crystallization of the hypersthene- 

 bearing gabbros, syenites, granites, etc., occurred, as just mentioned, 

 at a considerably lower temperature. The conclusion of this, 

 according to my view, is that the proved crystallization of forsterite 

 in a melt of MgSiOs at a pressure of one atmosphere cannot be 

 transferred to take place at the crystallization of the bronzite- or 

 hypersthene-bearing deep-seated rocks. I cannot agree with more 

 of the opinions maintained by Andersen and Bowen {loc. cit.) 

 and Bowen^ since these investigators have not taken into considera- 

 tion the influence of pressure. 



^The crystallization of haplobasaltic and haplodioritic magmas (19 1.5), and The 

 later stages of the evolution of the igneous rocks]^(i9i5). 



