MINERALOGICAL CLASSIFICATION OF IGNEOUS ROCKS 77 



is incorrect in theory. Tyrrell 1 says that Iddings' system is faulty 

 in this respect, and suggests uniting all the soda molecules with 

 orthoclase, and comparing the sum with the lime molecules. But 

 to this the objection may be made that it fails to separate the soda- 

 from the potash-rocks. Personally the writer prefers to go one 

 step farther and separate the three molecules, as shown in 

 Fig. 11. If Kf and all the Naf were united, it would make 

 difficulty in the monzonite group where the Ab molecule must 

 be separated from the An. Thus with the potash and soda 

 united, a rock with 50 per cent orthoclase and 50 per cent 

 andesine (Ab 6 oAn 40 ) would give (Or-f Ab) 8o An 20 , while T 'f classified 

 by the ratio of orthoclase to albite plus anorthite it would 

 give Or 50 Plag 50 . The difficulty in determining the albite in most 

 microperthite is not great; the amount can be estimated with 

 little error. 2 Of course this is not possible in anorthoclase, and 

 rocks containing much of this mineral will have to be determined 

 chemically. Ordinarily, however, the amount of soda is too small 

 to change the classification of the rock, even if neglected. In rocks 

 which contain known amounts of soda-orthoclase and plagioclase, 

 the molecules must be separated. Thus a ciminite from the 

 Roman Comagmatic Region 3 contains soda-orthoclase (OreAbi) 

 43.6 per cent and labradorite (Ab x An 2 ) 16. 1 per cent, which 

 gives orthoclase 37.4 per cent and albite 6.2 per cent from 

 the soda-orthoclase, and albite 5 . 4 per cent and anorthite 10 . 7 

 per cent from the labradorite. Uniting these there is orthoclase 

 37.4 per cent, albite 11. 6 per cent, and anorthite 10.7 per cent. 

 This gives Ab S2 An 48 , the point falling just on the Ab side of Ab x Ani 

 or in Order 2, and Kf 63 Plag 37 , which brings the rock in the row of 

 families 3, 8, 13, etc. (Fig. 16). Zonal feldspar may be determined 

 by considering the approximate amounts of each kind and obtaining 

 the average Ab-An value. This will be necessary in but few cases, 

 for ordinarily it may be determined by inspection whether the 



1 G. W. Tyrrell, "A Review of Igneous Rock Classification," Science Progress, 

 No. 33 (July, 1914), 79. 



2 For figures giving a comparison of measured and calculated values see Eero 

 Makinen, Bull. com. geol. Finlande, No. 35 (1913), p. 74; Charles H. Warren, Proc. 

 Amer. Acad. Arts and Sciences, LI (1915), 127-54. 



3 H. S. Washington, op. cit., p. 65. 



