﻿Vol. 66.] CLASSIFICATION OF IGNEOUS ROCKS. 485 



IV. Classification by Eutectics. 



Recognition of the fact that magmas are solutions implies the 

 existence of eutectic mixtures in them. The low melting-point of 

 the eutectic suggests that partial crystallization necessarily brings 

 the residue of the magma nearer to the eutectic proportion in com- 

 position. The study of magmas in the light of physical chemistry 

 has, then, naturally led certain investigators to hope that in this 

 relatively new field of rock-properties there may be something 

 available for fundamental genetic or natural classification. Becker, 

 Vogt, and Harker have made the most important suggestions in 

 this direction. 



Becker's suggestion. — G. F. Becker was the first to express 

 the idea of classification by eutectics. 1 He regards the eutectic as 

 the solvent in the magmatic solution and hence of greater im- 

 portance than the variable materials in excess, which, in porphyries, 

 are developed as phenocrysts. This idea is not in accord with the 

 conception now current that the magma is a mutual solution of its 

 several components. The following quotations indicate the way in 

 which Becker would use eutectics in classification, and the special 

 purpose subserved by so doing : — 



• If we knew all about magmas, it seems fairl} r certain that we could 



define a number of eutectic mixtures, each, when heated above its meltinc 

 point, yielding an infinite variety of solutions corresponding to rocks of au 

 infinite variety of compositions 



; From this point of view the groundmass of rocks would be more interesting 

 and important than the phenocrysts, while it has usually been studied with 

 less care, because of the greater difficulties in the way of mineralogical deter- 

 mination. The groundmass would either consist substantially of the eutectic 

 mixture, or afford a closer approximation to it than does the whole rock. 



' It is difficult to imagine that the comparatively small number of elements 

 which enter largely into the composition of massive porphyritic rocks should 

 form any very great number of independent eutectic mixtures ; and it seems 

 to me that it would be possible to elaborate a eutectic classification of those 

 rocks which have consolidated from the liquid state — I mean the porphyries — 

 each rock-group representing a series of solutions in one eutectic liquid. Such 

 a classification would also have certain geological advantages over others, for 

 the composition of the groundmass of rocks largely determines their orogenic 

 signi ficance 



' Even among the rocks which represent solidified fluids there is a class not 

 subject to such a classification as is here proposed. It seems an inevitable 

 conclusion from the laws of precipitation that there must be many rocks which 

 have been formed by fractional crystallization 



'Now, such fractional precipitates are essentially impure. Either in nature 

 or in the laboratory, they represent neither the substance dissolved nor the 

 solvent or mother liquor, but only fortuitous mixtures of the two — crystals of 

 precipitate, including and entangling variable quantities of mother liquor; 

 crusts, which vary in composition from millimetre to millimetre. Rocks of 

 such origin appear to me insusceptible of any strict classification and fit only 

 to throw a dim light on the qualitative composition of the magma, which they 

 represent, indeed, but only partially and irregularly. Were a eutectic classifi- 

 cation worked out, it would probably be easy to recognize these impure partial 

 precipitates, which would then receive the scant attention they deserve.' 



1 ' Report on the Geologv of the Philippine Islands' 21st Ann. Rep US 

 Gool. Surv. pt. iii (1901) pp. 487-614 (pp. 519-20). 



