74 SCIENCE PROGRESS 



of mineralogical or chemical similarity. Furthermore a modal 

 classification, whatever its difficulties, is the one based on the 

 characters most readily elicited from the rocks themselves. It 

 will, at any rate, be more convenient, save for special purposes, 

 than a normative classification which requires a tedious process 

 of chemical analysis before a rock can be placed in its proper 

 position. 



Difficulties at once present themselves when a quantitative 

 classification on modal lines is considered. The texture of 

 many rocks is such that the mode cannot be elicited at all ; 

 and in the case of merocrystalline and porphyritic rocks, 

 Iddings has justly pointed out the anomalies that result from 

 a classification based simply on the identifiable crystalline 

 constituents regardless of the composition of the minutely 

 crystalline or non-crystalline remainder. 1 Furthermore a dif- 

 ficulty arises in the fact that the same mineral species may 

 vary considerably in chemical composition in different rocks. 

 The phenomenon of "occult" minerals is in this connection 

 to be considered as a special case of variable chemical 

 composition. 



Before any system can be applied it is necessary, therefore, to 

 consider the methods by which the mode may be obtained from 

 various kinds of rocks. With the holocrystalline granular rocks 

 there is very little trouble. The graphic method of quantitative 

 mineral measurement invented by Rosiwal yields sufficiently 

 accurate results for classificatory purposes in the great majority 

 of types. It might here be said that the degree of accuracy 

 required in the estimation of the mode depends on the size of 

 the classificatory divisions it is proposed to erect. If the latter 

 are to be as comprehensive as, say, the subrangs of the American 

 Quantitative Classification, a great degree of accuracy is not 

 necessary. In many cases the experienced petrographer can 

 estimate the quantitative mineral composition of the rock from 

 a careful examination of the thin sections with sufficient accuracy 

 for this purpose. 



The expression of the mode in aphanitic rocks is more 

 difficult, since they are frequently insusceptible to the Rosiwal 

 method. The approximate mode of a quite fine-grained rock, 

 however, may be obtained by this method provided that it is 

 holocrystalline. It must be admitted, however, that with the 



1 Igneous Rocks, vol. ii. 1913, p. 72. 



