122 A. Holmes — Classification of Igneous Rocks. 



same symbol would be much more closely akin than is at present the 

 case. If no normative felspar were present, the rang symbol could 

 be omitted, and the sub-rang symbol would then express the kind 

 of felspathoid in the rock. A fivefold classification of the soda-lime 

 felspars by the rang method would probably be equally as good as the 

 sixfold classification usually adopted. 



A natural line of division, similar to that employed to separate 

 quartz-bearing from felspathoid-bearing rocks, may be drawn between 

 quartz-bearing and olivine-bearing rocks. 1 To express the presence of 

 olivine in the norm of a rock the number corresponding to the ratio 

 of pyroxene to olivine (section of grad) might be added to the 

 symbol. As an example of the kind of normative symbol here 

 advocated, the following may be taken: II, 5j, 4, 4, 1. The 

 characters of the rock could then be read off immediately : — 



II signifies that salic minerals are dominant ; 



5i ,, ,, felspathoid is accessory ; 



4 ,, ,, labradorite is present ; 



4 ,, ,, soda is dominant over potash ; 



1 ,, ,, pyroxene is dominant over olivine. 



Clearly the rock would be an olivine-essexite or its equivalent, 

 according to the texture and mode of occurrence. 



A purely mineralogical classification, as already pointed out by 

 many petrologists, cannot hope to attain complete parallelism with 

 a normative classification. The mere existence of such alferric 

 minerals as hornblende and pyroxene assures that unfortunate fact. 

 Nevertheless, all the co-ordinates of the CLP. W. Classification 

 (with the modification suggested above) are available for a modal 

 classification. 



(«) The ratio of felsic to mafic minerals is an important character- 

 istic, but certainly not sufficiently so to take first place. If the 

 mineral composition is quantitatively stated, then the ratio is known 

 and may be used for various purposes, such as a subsidiary means of 

 classification, or as a test of rival theories of differentiation. The 

 terms leucocratic and melanoeratic should, in the opinion of the writer, 

 be used without exact quantitative significance, to express variations 

 within a single body of rock, or within a series of associated rocks, 

 relative to the average, or (assumed) parent rock type. For quanti- 

 tative divisions the modal terms f els >c and mafic are available, and with 

 the C.I.P.W. limits and terminology rocks may he described as 

 perfelsic, Mo-M 12 . 5 (using M for mafic constituents with percentage 

 suffixes) ; dof'elsic, M 12 . 5 -M s7 . 5 ; niafelsic, M 37 . 5 -M 62 . 5 ; domafic, 

 ^62.5-^87.5; and permafic M 87 . 5 -M 100 . (See also Shand, 1916, 

 pp. 400-4.) Tyrrell's suggestion to group the first two of these 



1 The occasional presence in oversaturated rocks of fayalite with tridymite, 

 or other forms of silica, need not invalidate this statement, provided that the 

 term " olivine-bearing rocks " is understood to imply magnesian olivine and to 

 exclude fayalite. The presence of the latter in a rock as a rare accessory does 

 not affect petrographical nomenclature, and as Shand has pointed out, it exists 

 in the company of free silica because an orthorhombic pyroxene of the 

 composition Fe Si Os is apparently incapable of a separate existence. (Shand, 

 1914, pp. 491-2 ; see also Washington, Journ. Geol., xxii, p. 16, 1914.) 



