THE BEHAVIOR OF INCLUSIONS IN IGNEOUS MAGMAS 553 



The so-called surplus material does not remain indefinitely as 

 such, but by the time we have obtained inclusions very rich in sil- 

 limanite a turning point in the process is reached. Hitherto certain 

 constituents of the schist have become a part of the liquid in virtue 

 of the precipitation of various phases with which the liquid (magma) 

 was saturated. Now reactions between magma and inclusions be- 

 come of such a nature that the precipitation of phases with which 

 the magma is saturated is the sole process, these phases being appro- 

 priately modified by the inclusions. There is now no addition to 

 the liquid. The exact modification of the phases that is produced 

 by sillimanite has already been discussed and equations representing 

 the changes have been written. The net result is an increase in 

 the amount of anorthite and magnesian pyroxene at the expense of 

 lime-bearing pyroxene, and this tends to promote the separation of 

 magnesian pyroxene as a distinct phase, orthopyroxene. Thus the 

 tendency of the magma to give a noritic differentiate is increased, 

 as well as the likelihood of formation of a pyroxenite containing 

 orthopyroxene. These expectations are well matched by the Cort- 

 landt series. 



Apart from possible later differentiation this production of 

 norite and related types is the end result of the action of basic 

 (basaltic) magma on sillimanite-rich inclusions. We may with 

 profit examine the details of the action, that is, the processes going 

 on within and immediately around the sillimanite-rich inclusions, 

 immersed in a magma rich in plagioclase and pyroxene. This exam- 

 ination serves to throw some light on the detailed mineralogy of the 

 inclusions, and in particular on the separation of free alumina, as 

 corundum. 



If a mass of sillimanite were added to some anorthite just above 

 its melting point, it can be readily seen from examination of Figure 

 II that some of the sillimanite would be converted into corundum. 

 This is because the line joining sillimanite and anorthite passes 

 through the corundum field. We may imagine, for example, that 

 the bulk composition of a layer immediately surrounding the silli- 

 manite is 50 per cent sillimanite and 50 per cent anorthite. The 

 mixture represented by this layer, at 1550°, would consist of silliman- 

 ite, corundum, and liquid. If the original anorthite liquid had an 



