MINERALOGICAL CLASSIFICATION OF IGNEOUS ROCKS 75 



represents all rocks, there are no sharp division lines except at 5 and 

 95 or thereabout. The central division points could equally well 

 be 50-50 or 35 and 65. On the whole, the writer thinks the 50-50 

 line best, but leaves this question open for the present. 



Lincoln 1 makes three divisions, leucocratic, mesocratic, and 

 melanocratic, according to the percentages of light constituents, 

 with division lines at 0-33-67-100; and in the expanded series, 

 five divisions at 0-4-33-67-96-100. 



Iddings 2 separates his rocks on the ratios o-f-f-100; that is, 

 into rocks with less than 37^ per cent dark, between 37^- and 62^ 

 per cent, and with more than 62^ per cent. This makes the first and 

 third groups very large. Even the C.I.P.W. general subdivisions 

 of 0-12^-371-62^-87^-100 would make the first and last groups 

 too large, for rocks with 12^ per cent of dark constituents (see 

 Fig. 7 with 10 per cent) certainly are not leucocratic. Further- 

 more, a division at 12^- or 37^ per cent at the leucocratic end 

 is not so logical as at 5 per cent (cf. Fig. 12). Shand 3 makes his 

 divisions at 100-97-90-80-70-60-50-40-30-20-10-3-0 per cent 

 light minerals. These, however, are too many for the purpose of 

 classification, the essential difference between rocks with 60 and 70 

 per cent of dark constituents, for example, being insignificant. 

 From the curves in Fig. 12 there appears to be little choice between 

 dividing lines at 33, 35, 37I, or 50. If there is any, it is in favor of 

 So-50. 



Orders.— Having divided the rocks into four (or five) classes 

 according to the amount of dark constituent, they may be divided 

 into orders on the basis of the plagioclase. 



In determining the kind of plagioclase in a rock, it has been 

 quite customary to give the Ab-An percentage in simple round 

 numbers, such as Ab 2 An 3 , A^An!, etc. This produces an excessive 

 number of rocks at these points, as is clearly brought out in Fig. 13, 

 which is less valuable for that reason. As may be seen, there are 

 crests at Ab, Ab 3 An r , Ab 2 An T , AbjAn^ Ab 3 An s , and Ab An I00 . 

 Having no other marked crests in the curve indicating natural 

 division lines, the writer has taken the points 0-5-50-95-100, 



1 F. C. Lincoln, op. cit., 556. 



2 J. P. Iddings, op. cit., II, 150, 308. 3 S. J. Shand, op. cit. 



