THE BEHAVIOR OF INCLUSIONS IN IGNEOUS MAGMAS 547 



reason why forsterite is precipitated for the formation of lime sih- 

 cates requires some siHca but the Hquid cannot be desihcated since 

 it lies on the forsterite saturation surface. Therefore the liquid 

 changes its composition by moving on the forsterite surface toward 

 lower temperature. It does not move directly away from forsterite, 

 however, but curves somewhat in the direction of the composition 

 of the inclusions. If the reaction with the inclusions is strongly 

 endothermic, as it is in this case, the liquid would move down to 

 the forsterite-pyroxene boundary curve where separation of pyrox- 

 ene would occur and the liquid would be entirely used up without 

 ever getting over to such compositions that akermanite would 

 separate from it. However, some part of the inclusions might have 

 been converted to akermanite. Thus, however large a supply of 

 inclusions were available, even in excess of the 1 2 per cent mentioned 

 above, the liquid might never get over to such compositions that 

 any new phase is precipitated from it, but would precipitate only 

 forsterite and pyroxene as in normal crystallization, with, however, 

 a certain increase in the lime content of the pyroxene. 



These considerations lead us to the conclusion that the liquid 

 we have chosen, even if it has a moderate amount of superheat and 

 therefore is capable of directly dissolving a little lime and magnesia 

 from calcite or dolomite, does not suffer a change in the kind of 

 solid phases capable of forming from it. There results only a modi- 

 fication of the composition of the phase of variable composition 

 (pyroxene). And, as a consequence of the heat effect of the solu- 

 tion of inclusions, saturation shortly ensues and thereafter further 

 action upon the inclusions is accompHshed only with concomitant 

 precipitation of the phase or phases with which the liquid is satu- 

 rated, whereby the liquid is constrained to follow a general course 

 not significantly different from the one it would follow were no 

 inclusions present. 



The results obtained in the foregoing enable us to draw certain 

 conclusions as to the effects of natural magmas upon inclusions of 

 various sedimentary rocks. One point that does not seem to be 

 realized is that when a sedimentary inclusion becomes immersed in 

 a magma nothing is added that the magma does not already contain. 

 Both belong to the same poly component system embracing all the 



