68 ALBERT JOHANNSEN 



into circles of known size irregular fragments cut from pieces of 

 black paper which bore definite ratios to the circles. 1 Of course, if 

 one has often measured constituents by the Rosiwal method his 

 estimates are likely to be fairly good. 



The system here presented is not intended as a substitute for any 

 chemical system. But, as so well expressed by Clarke, "Even if it 

 [the C.I.P.W. system] should be finally adopted by all petrologists, 

 some form of classification like that now in vogue would have to be 

 retained with it. Good analyses cannot be obtained for every 

 rock which the geologist is called upon to determine, and in many 

 cases he must be content with the results of a microscopic examina- 

 tion." 2 And it is also true that for rocks which show considerable 

 decomposition the microscopic method is far more likely to give 

 good results than the chemical. 



As an objection to a quantitative mineralogical system, such as 

 is here proposed, it will be said that it is not always possible to 

 determine the exact composition of rocks with a glassy base or 

 extrusive rocks of the alkali series. But the percentage of inde- 

 terminable rocks is comparatively small, and for these there still 

 remain, if necessary, chemical methods for determining the compo- 

 sition of the base. Most glassy rocks are leucocratic, and a recalcu- 

 lation into the minerals which would have crystallized had the 

 conditions been right is easy. Since the majority of these glassy 

 rocks are rhyolitic, one is no worse off in adopting a quantitative 

 classification than at the present time, when they are called rhyo- 

 lites from microscopic examination. In such cases it would not 

 be objectionable to make use of tentative names which could be re- 

 vised after chemical analyses have been made. In a later paper the 

 author hopes to present a method for determining quantitatively 

 even these rocks with very little difficulty. Certainly 95 per cent 

 of fresh igneous rocks can be classified microscopically. When 

 rocks are completely decomposed, no determinative system, 

 chemical or mineralogical, will help. 



The plutonic rocks must necessarily form the type families of 

 any mineralogical classification of igneous rocks, and extrusive 



1 The actual percentages in the figures are \, 1, 2, 5, 10, and 50. 

 3 F. W. Clarke, "Data of Geochemistry," U. S. Geol. Surv., Bull. 616, (Wash- 

 ington, 19 16), p. 432. 



