THE BEHAVIOR OF INCLUSIONS IN IGNEOUS MAGMAS 529 



in the paper it will be shown that the observations of many petrog- 

 raphers upon actual rock inclusions strongly suggest just such an 

 action. 



If we turn now to inclusions of plagioclase less calcic than the 

 crystals with which the liquid is in equilibrium we find that quite a 

 different condition obtains. We may begin with the liquid A, 

 again at 1250°, and add to it foreign inclusions of AbiAui already 

 heated to 1250°. It will be noted that we are adding plagioclase of 

 the same composition as that in the liquid since liquid ^ is a mix- 

 ture of AbiAui and diopside. Suppose that the reaction takes place 

 at first between a thin layer of liquid adjacent to the inclusion and 

 an equal weight of the peripheral part of the inclusion. The com- 

 position of this reacting mass is then to be represented by the point 

 F. Suppose further that the reacting mass is a very small portion 

 of the total so that no significant drop of temperature occurs, for in 

 this case the reaction absorbs heat. Equilibrium will be established 

 in this reacting layer sensibly at 1250° when the composition of the 

 crystalline outer crust of the inclusion is Ab3An7 (Z) and that of the 

 adjacent thin layer of liquid is M, which is a mixture of AbaAuj and 

 diopside. If the general cooling of the liquid were proceeding 

 rapidly so that this condition were "frozen-in" we would have a 

 central core of unaltered inclusion of composition AbjAui, an 

 altered crust of the inclusion of composition Ab3An7, a reaction rim 

 (formerly liquid) about this, consisting of a mixture of diopside and 

 Ab2Anx, all surrounded by the main mass consisting of a mixture of 

 diopside and AbiAuj. The reaction rim is not intermediate in com- 

 position between the inclusion and the main mass, nor yet between 

 the altered crust of the inclusion and the main mass. This is a 

 commonly observed feature of reaction rims and has been referred 

 to "selective diffusion," an explanation which is correct but rather 

 misleading. The effect is really due to equilibrium effects between 

 the inclusion and the liquid and such diffusion as occurs is selective 

 only because equilibrium selects certain constituents to become part 

 of the liquid and others to be fixed in the solid phase. In the case 

 we have described, more sodic plagioclase is emphasized in the new 

 liquid formed and more calcic plagioclase is emphasized in the solid 

 product. We shall find it to be a general result of reactions of this 



