126 A. Holmes — Classification of Igneous Rocks. 



three divisions are necessary, those corresponding to albite (up 

 to An 15 ), to oligoclase and andesine (from An 15 to An 50 ), and to 

 labradorite to anortliite (from An B0 to An 100 ). Symmetry seems to 

 demand an anorthite division, and in some cases it would be useful 

 to distinguish oligoclase rocks. Probably the five rang divisions of 

 the soda-lime felspars would give somewhat better results, the present 

 objection to using them being merely the lack of an appropriate 

 nomenclature. In particular, the boundary between andesine and 

 labradorite at An 47 is preferable to that at An 50 . However, these 

 are extensions of the method which will fall naturally into place as 

 more rocks are quantitatively examined and described. 



Secondly, it is proposed to divide the rocks further according to 

 the ratio of orthoclase to albite, or, stating it more generally, according 

 to the molecular ratio of potash to soda. The ratio is conveniently 

 expressed by the percentage of orthoclase in total orthoclase plus 

 albite. In order to give the felspathoid minerals and the micas due 

 weight in the classification, these may also be expressed in terms of 

 the amounts of orthoclase and albite to which they are approximately 

 equivalent. The chief factors required for this conversion are given 

 approximately in the following table : — 



Orthoc%ase=1'0 Albite =ro 



Anorthoclase = 0'4 Anorthoclase =0'6 



Leucite =1"4 Analcime =1*3 



Muscovite =0'7 Nepheline =1"6 

 Biotite =0'6 



Some petrologists may object to the inclusion of biotite as a mineral 

 comparable to orthoclase, but since it may be regarded as containing 

 a leucite-like molecule, it seems desirable to take it into considera- 

 tion. The writer is aware that at this stage the classification ceases 

 to be purely mineralogical, but it is difficult to avoid some such 

 grouping of minerals if the classification is to be one that can readily 

 be tabulated and memorized. To employ every important mineral as 

 a classificatory co-ordinate would demand more dimensions than can 

 be printed or mentally visualized. The point raised touches the 

 question of the objects of the classification, and these may be stated 

 as follows : — 



(a) To attain parallelism with a chemical classification as far as 

 possible. 



(b) To form compartments which shall largely define and limit 

 the types of rocks falling within them. 



(c) To express the relations of rocks to one another ; linear in 

 a chemical and mineralogical sense, and genetic as far as this can be 

 done by a division of space into compartments. 



(d) To enable students of the subject easily to memorize the ever 

 increasing list of rock names. 



{e) To indicate to petrologists where new names are necessary, and 

 where they may, with advantage, be avoided. 



Returning from this digression to a consideration of the actual 

 subdivisions adopted on a potash to soda basis, we may consider 

 those that have already been used. In the C.I.P.W. Classification 

 the five sub-rangs would give limiting positions at Or 12 . 5 -Or 3 7. 5 -Or 6 2.5- 



