CLASSIFICATION OF IGNEOUS ROCKS 279 



speak, a foreshortening by which rocks that differ considerably 

 in composition may be represented by the same point, so that 

 the natural groupings are more or less obscured. The difficulty 

 may to some extent be solved by constructing a number of 

 diagrams which, though based on the same general principle, 

 have the components differently selected or combined, and 

 correspond to projections of solid figures on different planes. 



The natural groupings which future research will, I believe, 

 map out, must stand in close relation to the principles on which 

 the separation of rock magmas takes place, and will guide us 

 to a simple and natural classification of igneous rocks which, 

 while essentially chemical, will indicate as far as possible the 

 origin and genetic relations of the different rocks with which it 

 deals. 



The task of framing such a classification must be postponed 

 to a future day — which may not, however, be far distant — when 

 the principles of the segregation of igneous magmas are better 

 understood than they are at present. 



In the meantime it would be the wisest course to employ a 

 system of classification on a mineral basis, which, like that 

 outlined by Professor Bonney, 1 takes the holocrystalline rocks 

 as the types, and considers the less completely developed 

 or the glassy forms with the same chemical composition as 

 modifications of them. 



The general adoption of the quantitative system of classifica- 

 tion which has been described in the present paper would, I 

 fear, tend to obscure the true issues, and to obstruct rather 

 than to facilitate the ultimate evolution of a natural division and 

 nomenclature of igneous rocks ; but petrologists will always 

 owe a debt of gratitude to its authors for making the first 

 definite reconstructive movement towards the radical reform 

 of the present system, or rather want of system. 



There is something to be said for a purely chemical classifica- 

 tion, but it is undesirable to introduce a number of divisions and 

 names that must soon be discarded, and a systematic use of 

 diagrams showing the chemical composition would serve every 

 purpose of a detailed classification on the same lines. 



In any case, if we wish to facilitate the introduction of a 

 scientific classification, and desire that the results of our work 

 should ultimately find their natural place amongst its divisions, 



1 Op. ciL, pp. 67-78. 



