120 A. Holmes — Classification of Igneous Rocks. 



based on the assumption that there are at least a few such groupings, 

 then instead of fitting them into ready-made compartments the 

 divisions of a satisfactory classification should he drawn around them. 



The other feature, to which many criticisms have been directed 

 (e.g. Evans, p. 270 ; and Harker, Natural History of Igneous Rocks, 

 pp. 363-4) — that widely different types of rocks may fall into the same 

 sub-rang — is due partially to the dependence of names on textures, 

 and partially to incorrect applications of the modal nomenclature, 

 much of which, it must be confessed, is unnecessarily vaguely 

 defined. Some glaring examples are exposed in vol. ii of Iddings' 

 Igneous Rocks: e.g., ' 'an olivine diabase " with 10'7per cent normative 

 quartz (p. 144); a "hornblende granite" with 56-8 per cent 

 labradorite and only 7*2 per cent of orthoclase (p. 146) ; and 

 a " basalt" with 40 per cent of nepheline and leucite and no soda- 

 lime felspar (p. 306). 



The relation of the rang classification of soda-lime felspars to those 

 already in use is shown by the following figures, which express 

 the percentage of anorthite in each case. 



Another reason for the lack of similarity between the rocks of 

 certain sub-rangs, and one for which the C.I.P.W. Classification is 

 itself wholly responsible, is displayed in the diagram on p. 118. In 

 this, the subdivisions formed by rangs and sub-rangs are drawn to 

 scale, and across them the distribution of the different soda-lime 

 felspars is plotted. In most cases it will he noticed that two or 

 three, and in a few others even more, of the recognized varieties of 

 plagioclase are present within a single compartment. In Toscanose, 

 for example (I, 4, 2, 3), the felspar may be albite, oligoclase, or 

 andesine, and it is precisely this sub-rang to which attention has 

 been directed by Harker (pp. 363-4) as exemplifying the extra- 

 ordinary degree of non-parallelism between modal and normative 

 classifications. Since in the former the kind of soda-lime felspar is 

 often an important diagnostic, it was an inevitable consequence that 

 the latter, as arranged in its present form (in which the co-ordinates 

 of rang and sub-rang are not independent of each other), should 

 traverse the older nomenclature in an unnatural way. For this 

 reason it is suggested for consideration by the authors of the 

 Quantitative Classification that the molecular ratio of salic alkalies 

 to lime should be replaced by the molecular ratio of felspathic soda 

 to lime. If this were done, each component of the magmatic symbol 

 would convey a definite mineralogical meaning, and rocks having the 



