182 JV. L. JBowen — Crystallization of Haplobasaltic, 



not long since the discussion of the theory of the crystalliza- 

 tion of igneous rocks was carried out hy describing the crystalli- 

 zation of some simple binary eutectic mixture such as salt and 

 water, and winding up with the statement that the crystalliza- 

 tion of igneous rocks is in some measure analogous. A better 

 grasp of the theoretical aspects of the problems and a consider- 

 able number of exact investigations of silicate melts themselves 

 have now carried the matter well beyond this stage. 



In the present investigation the mixtures treated are suffi- 

 ciently close to certain natural magmas, in essential mineral 

 composition, that I have, with more or less propriety, referred 

 to them as haplobasaltic, haplodioritic and related magmas. 

 (Note also analyses compared in Table VI.) The names seem 

 further justified by the number of facts and principles govern- 

 ing the crystallization of their natural analogues that are 

 brought out by the investigation of these artificial melts. 



One fact stands out very clearly, viz., the great difference 

 between the crystallization of the mixtures that have been 

 described and crystallization in a system with a ternary eutectic. 

 There is no eutectic between diopside and any intermediate 

 plagioclase, still less between plagioclase and a complex 

 pyroxene solid solution such as augite. 



There is, then, nothing to be gained from the search for the 

 gabbroic eutectic, the dioritic eutectic and so forth. Apprecia- 

 tion of the non-existence of such eutectics is an important mat- 

 ter. The existence of a gabbro eutectic would mean that there 

 is a certain definite, lowest-melting mixture of calcic plagio- 

 clase and augite, with a little magnetite, towards which the 

 liquid, during the crystallization of any gabbro, always pro- 

 ceeds and beyond which it can not pass. The sinking of 

 crystals or the squeezing-out of residual liquid would avail 

 nothing in enabling the liquid to pass the eutectic temperature 

 and composition, the eutectic being a necessary end-point. A 

 gabbro magma might give rise locally to anorthosite and 

 pyroxenite by the sorting of crystals according to their densi- 

 ties, but, by crystallization-differentiation, it could never give 

 diorite or syenite or granite. 



On the other hand, if it is realized that the crystallization of 

 gabbro is in large measure analogous to that of haplogabbro 

 the possibility of the derivation of an igneous rock series 

 becomes apparent. Haplogabbro magma of composition 

 AbjAn 2 50 per cent-diopside 50 per cent, may, if it is cooled 

 very rapidly, give rise simply to haplogabbro with 50 per 

 cent plagioclase crystals of the uniform composition A^An, 

 and 50 per cent diopside. Yet if the same original magma is 

 cooled more slowly, different results may be obtained.* The 

 plagioclase crystals may be zoned, with compositions ranging 

 * See discussion of figures 11 and 13. 



