July 17, 1914] 



SCIENCE 



101 



gions is needed before definite conclusions may 

 be reached regarding' the actual nature of such 

 provinces and their significance with respect 

 to the dynamical history of the earth" 

 (page v). 



Part I. of the volume, devoted to the syste- 

 matic description of igneous rocks, is intro- 

 duced by an important review of their char- 

 acters and the difficulties attending their 

 systematic treatment. It is pointed out that 

 the rock type " is subjective, inherent in the 

 petrographer, not the rock." Igneous rocks 

 form a continuous series, hence their syste- 

 matic treatment must be arbitrary as to di- 

 visions, but it does not follow that all methods 

 of classification are equally good or bad. 



One of the most important features of Id- 

 dings's work is the influence it seems destined 

 to have on the development of systematic pe- 

 trography. That the user of the book may 

 fully appreciate this it seems desirable to sur- 

 vey tlie present situation of the science, on 

 lines not especially emphasized by the author. 

 . The current system of petrography, called 

 the qualitative system by Iddings, classifies 

 igneous rocks chiefly by their important min- 

 eral constituents. For granular rocks there is 

 more and more effort to recognize the quanti- 

 tative development of certain minerals, while 

 nephelite, leucite and others are given great 

 weight, almost regardless of their abundance. 

 In porphyritic rocks, however, especially if 

 micro-, crypto- or hypocrystalline, the pheno- 

 crysts alone are given classificatory value in 

 many cases (Eosenbusch system). The 

 groundmass is practically ignored by many 

 petrographers in naming rocks. 



The fundamental importance of chemical 

 composition of igneous rocks is universally 

 recognized. Mineral composition is an ex- 

 pression of the chemical, though less directly 

 than once supposed. The natural ambition of 

 the petrographer to express chemical compo- 

 sition in his mineralogical classification is 

 frustrated by the fact that he can not ascer- 

 tain the mineral composition of many rocks 

 at all accurately and by the variable chemical 

 composition of most rock-making minerals. 

 Only the inherent and perhaps insuperable 



difficulties of the problem have prevented the 

 formulation, before now, of a satisfactory 

 mineralogical classification, and it is equally 

 certain that efforts to improve that system will 

 go on with at least some ineasure of success. 



The quantitative system, of which Iddings 

 is one of the authors, permits an accurate 

 classification of igneous rocks, desirable for 

 many purposes, wherever the chemical compo- 

 sition can be ascertained. It is self-evident, 

 however, that the data for the quantitative 

 classification of the vast majority of rocks can 

 not be obtained and that another system of 

 general applicability must be used concur- 

 rently with it. The most evident and familiar 

 character-giving feature of igneous rocks is 

 their mineral composition, and it seems clear 

 that some improved form of the current sys- 

 tem must always remain the one for general 

 purposes. But to satisfy the natural demands 

 of scientific men this system must be given 

 much greater precision and consistency than it 

 now possesses. The approach to a real system 

 must come by introducing greater precision in 

 definition, using quantitative mineral compo- 

 sition as far as practicable and expressing in 

 the most feasible manner a correlation be- 

 tween chemical and mineral factors. The 

 petrographic system of the future for gen- 

 eral purposes will be an evolution from the 

 unsatisfactory one now current through the 

 trying out of many propositions and a selec- 

 tion of the best. 



A quantitative factor is now being intro- 

 duced into the mineralogical system in various 

 ways. Monzonite, granodiorite and other 

 major terms illustrate this, and a large num- 

 ber of new rock names of lesser importance 

 have been recently proposed in recognition of 

 the abundance or prominence of certain min- 

 erals. But this development is not controlled 

 or guided by definite rules or principles, and 

 until such have been adopted increased con- 

 fusion must be the result. 



Probably all petrographers are ready to wel- 

 come a practical and logical proposition to 

 modify or control the mineralogical classifica- 

 tion by chemical data. Biit it is evident that 

 the application of such a scheme implies chem- 



