Andersen — System Anorthite-Forsterite-Silica. 451 



take place. It is seen that the system need not be complex if 

 only one of the boundary curves has a certain convex shape in 

 relation to a conjugation line. The system anorthite-forsterite- 

 silica does not show any instances of recurrent crystallization. 



It is evident, then, that magmatic resorption and recurrent 

 crystallization in igneous rocks may be explained as conse- 

 quences of simple cooling of the magma. This way of explana- 

 tion by the reaction principle is not a new theory, set up with 

 the sole purpose of explaining a certain phenomenon, but is a 

 simple deduction from a great general principle, the phase rule, 

 amply supported by experiments. In each case, therefore, 

 where the question arises as to the proper explanation of mag- 

 matic resorption and related phenomena, the reaction principle 

 should receive due consideration before resorting to any of the 

 theories which imply changes of the physical conditions, due 

 to exterior influence, or of the chemical composition of the 

 magma as a whole during the solidification. 



Application of the reaction principle to olivine-bearing rocks. 



It would be premature now to discuss the application of the 

 reaction principle to rocks in general with the object to prove 

 or disprove its validity. A discussion of the special application 

 to some olivine-bearing rocks may, however, be justified because 

 of the experimental data available in this case. Such a dis- 

 cussion will also throw some light upon the problem of the 

 general application. 



Jfagmatic resorption of olivine. — It has been shown* that 

 in the two-component system forsterite-silica, forsterite is 

 wholly or in part dissolved on cooling in all mixtures where it 

 appears as primary phase. If now nothing were known beyond 

 this fact we might still draw some conclusions concerning the 

 behavior of forsterite in systems containing more components. 

 We see that the relations of forsterite would not undergo any 

 sudden change if a third component were added to the system 

 forsterite-silica. In some part of the ternary system the qual- 

 ity of forsterite to dissolve on cooling must be maintained. 

 This part, the reaction field of forsterite, may embrace only a 

 fraction of the entire ternary forsterite field, but may equally 

 well include all this field and also parts of adjoining fields, as 

 in the system anorthite-forsterite-silica. This obvious conclu- 

 sion is based on what we may call the principle of continuity, 

 which assumes a gradual transition between a ternary system 

 and its binary systems, the latter being considered ternary sys- 

 tems with infinitely small amounts of the third component. 

 * See N. L. Bowenand Olaf Andersen, loc. cit.. p. 496. 



