192 N. L. BOW EN 



have a reaction relation to the Hquid, not a mere subtraction rela- 

 tion. Each separated mineral tends always to change into a later 

 member of the reaction series. This change of composition is 

 effected by reaction with the hquid, and according to the oppor- 

 tunity for reaction the Hquid is entirely used up, in some cases 

 sooner, in others later, and only then is soHdification complete. 



Thus we see that rock series cannot be partitioned off into 

 such divisions as gabbro, diorite, etc., each having a eutectic of 

 its own. All of these belong to a single crystaUization series, to 

 a single polycomponent system, which is dominated by reaction 

 series. 



Even after what is ordinarily termed complete consolidation 

 of the rock many changes of a reaction nature may occur between 

 the minerals. Usually a small amoimt of Hquid is a factor in these 

 changes which are commonly described under the head, meta- 

 morphism. Certain aspects of these changes have been described 

 by V. M. Goldschmidt in a paper on metasomatism.' It is impor- 

 tant to note in the present connection that metasomatic reactions 

 are not confmed to the metamorphic stage but are, as we have seen, 

 a constant feature of the whole process of crystalHzation of the 

 magma. Eskola has presented evidence for beHeving that, at 

 times, the original crystallization of a magma may itself take place 

 ■under the conditions that are ordinarily prevalent during meta- 

 morphism. The rock then presents a "mineral facies" identical 

 with that of a metamorphic rock of the same composition formed 

 under the same conditions. Totally different reaction series would 

 be concerned in the crystalHzation of these rocks and we have not 

 gone far toward an understanding of these, though Eskola has 

 done much to set us on our way.^ The variation of conditions 

 necessary to produce different facies is of a different order of 

 magnitude from the variation of conditions here considered. 

 This is largely a variation of the rate of cooling and we are here 

 concerned mainly with the diverse rocks in the common and defi- 

 nitely estabHshed sequence that can be produced under the 

 conditions of his diabase facies principally. With adequate 



' Economic Geology, Vol. XVII (1922), p. 105. 



^ P. Eskola, Norsk Geologisk Tidskrifl, Vol. VI (1920), pp. 143-94. 



