Andersen — System Anortliite-Forsterite-Silica. 453 



of other mixtures (50-60 per cent anorthite) lying within the 

 ternary reaction field of forsterite would give the compositions 

 of olivine gabbros. We know that the reaction field of forster- 

 ite in a complex system must embrace all the mixtures cor- 

 responding to the ternary reaction field and at least some of 

 the complex mixtures. Small changes of mixtures which lie 

 within the ternary reaction field, like those mentioned above, 

 are therefore not likely to move the composition points of the 

 resulting mixtures outside the reaction field in the more com- 

 plex system produced by the changes. As the olivine in prac- 

 tically all oli vine-bearing gabbros and anorthosites show 

 magmatic resorption, the explanation of this phenomenon by 

 the reaction principle seems probable and of general application 

 to these rocks. 



Summary of application to olivine-hearing rocks. — It is 

 clear that similar reasoning to that applied above to the reaction 

 field of forsterite (olivine), may be applied to other fields in 

 any system. Without repeating any arguments the conclusions 

 which may be drawn concerning the crystallization of olivine- 

 and pyroxene-bearing rocks, based on the principles and ex- 

 perimental data represented in the present paper and other 

 papers from this laboratory,* are indicated below. 



1. When olivine occurs in pyroxene-bearing rocks it should 

 in general begin to crystallize before the pyroxene and its crys- 

 tallization should also in general be completed before that of 

 the pyroxene. Exceptions to this rule should be found in cer- 

 tain diopside-bearing rocks. 



2. The olivine crystals of pyroxene-bearing rocks should be 

 more or less corroded as a consequence of the magmatic reac- 

 tion during the cooling. 



3. Some olivine- and diopside-bearing rocks should show two 

 generations of olivine due to crystallization, dissolution and 

 recrystallization during the cooling. 



These conclusions are based on ciwstallization with equili- 

 brium between liquid and solid throughout the period of 

 crystallization. We may also consider a case where the final 

 product of crystallization does not represent equilibrium 

 between the phases involved. 



4. The corroded crystals of olivine may be imbedded in the 

 reaction products (pyroxene, amphiboles) or in other minerals 

 in such a way that they are protected against further reaction 

 with the magma. In this case we may find for example olivine 

 and silica together in the same rock. 



In how far these conclusions, based on generalizations of ex- 

 perimental results, are corroborated by observations on actual 

 rocks will be known by all penologists and need not be pointed 

 out. 



* N. L. Bowen and Olaf Andersen, loc. cit. N. L. Bowen. loc. cit. 



