6 4 



ALBERT JOHANNSEN 



for classification. Elsewhere he uses chemical data to classify 

 rocks which he defines in mineralogical terms; for example, the Si0 3 

 percentage must lie between certain limits, or the sum of the alkalies 

 must be less than the alumina, etc. 



Another objection to the present system of classification is the 

 fact that rock terms have been used loosely or with different 



meanings. Thus dolerite, 

 originally applied to a 

 coarse basalt, has been used 

 for any dark rock, and in 

 England is used for rocks 

 which we call diabases. 

 The term diabase in the 

 United States means a 

 dike-rock with an ophitic 

 texture, yet it was origi- 

 nally used for Paleozoic 

 basalts and is still so used 

 in various countries. 

 Basalt has been applied 

 to plagioclase rocks with 

 augite and olivine irrespec- 

 tive of the kind of feldspar, 

 to labradorite-pyribole 1 



Fig. i. — One hundred and nine so-called 

 "granites." Open circles are rocks of Class i, 

 and dark circles rocks of Class 2. The double 

 circle is the mean of Daly's granites re- 

 computed into the probable modal minerals. 



rocks with or without olivine, to the darker labradorite-pyribole 

 rocks, to post-Tertiary extrusives of gabbroic magma, etc. 



The loose usage of terms by different writers with respect to 

 the mineralogical compositions of rocks is well brought out by 

 Figs. 1 and 2, in which the three corners of the triangles represent 

 respectively quartz, potash-feldspar, and plagioclase. In Fig. 1 

 are plotted 109 so-called "granites," taken, not from old descrip- 

 tions, but from comparatively recent ones in which the actual 

 mineral compositions were determined by the Rosiwal method 

 or by careful estimation by the various writers themselves. In a 

 few cases the rocks were doubtless named on the basis of their 



1 A general term for the members of the pyroxene and amphibole groups (Albert 

 Johannsen, op. cit.). 



