THE REACTION PRINCIPLE IN PETROGENESIS 191 



ever, they become less distinct, in the aluminous pyroxenes and 

 amphiboles a certain amount of interlocking begins and they 

 finally merge into a single series.' This is expressed diagrammati- 

 cally by a convergence of the series, with a dovetailing of the mineral 

 names at first, and finally a joining of the two series by the arrows 

 converging upon potash feldspar. Just where the two series merge 

 completely is more or less a question, but it is given closely enough 

 for our present purpose in the figure. 



That the series, oHvines-pyroxenes-amphiboles-biotites, con- 

 stitutes a reaction series is well attested in many rock varieties. 

 By this is meant that liquid reacts with ohvines to produce pyrox- 

 enes, with pyroxenes to produce amphiboles, and with amphiboles to 

 produce biotites. In the continually increasing water content of the 

 series it is related to the series K2Si03-K2Si03-^H20-K2Si03-H20. 

 The continuous reaction series of the plagioclases is perhaps the 

 best understood series of rock minerals. This is fortunate, for 

 the series happens to be of particular importance in that it runs 

 through a wide range of conditions and compositions in the rock 

 series. We simply have a continual enrichment of the liquid in 

 alkaline feldspar with the separation of the potash variety of 

 alkaline feldspar as a separate phase when it has exceeded its 

 solubiHty in the plagioclase mixture. With the formation of pot- 

 ash feldspar in the one series and of biotite in the other, the two 

 series are now so intimately intermingled as to constitute a single 

 series. 



There is a Httle of the nature of eutectic crystallization in the 

 • crystalHzation series given in the foregoing. At early stages and 

 as between the two series there is some suggestion of the eutectic 

 relation in that a member of one series lowers the melting "point" 

 of a member of the other series. Moreover the one or the other 

 begins to separate first according to which is present in excess over 

 certain fixed proportions. There the analogy with eutectic crys- 

 tallization ends for the simple reason that there is no eutectic, 

 no inevitable end-point where final soHdification must take place 

 when the liquid has attained a certain composition. The minerals 



' The two series are bridged at the very outset by spinel but this has, on the whole, 

 no great practical importance. 



