8o SCIENCE PROGRESS 



the common and significant variation in igneous rocks which 

 has given rise to the conception of alkalic and calcic 

 " branches." 



It is considered therefore that this ratio should replace 

 that used by Iddings, especially as it is quite as easily obtained. 

 From the optical examination it is now possible to determine 

 the composition of the lime-soda felspars quite accurately ; 

 and it is then only necessary to add the amount of the albite 

 molecule to the alkali-felspar already quantitatively determined, 

 and to use the remaining amount of anorthite as the denomi- 

 nator of the fraction, in order to obtain the required ratio 

 between the alkalic and calcic felspars. Complications are 

 introduced by the frequent zonal structure of the plagioclase 

 felspars, making it difficult to estimate the proportions of 

 the albite and anorthite molecules ; and by possible perthitic 

 intergrowths of lime-soda felspars with orthoclase. But the 

 errors introduced by these difficulties in the estimation of 

 the felspar ratio will not commonly be so large as to shift 

 the rock from its rightful place in the classification. The 

 grade of accuracy required in the quantitative estimation is 

 governed by the size of the classificatory compartment into 

 which it is desired to introduce the rock ; as and it is probable 

 that the compartments of the modal classification would not 

 be smaller than the subrang of the American Quantitative 

 Classification, a high grade of accuracy would not be re- 

 quired. 



A third criticism of Iddings' classification is that while it is 

 ostensibly based on the mode it is the norm that is most fre- 

 quently used to determine the place of any particular rock. It 

 must be admitted that this is largely due to the paucity of data 

 at his disposal as to the modes of igneous rocks. It is a diffi- 

 culty that only future work will remove ; and its discussion 

 here may serve to remind descriptive petrographers how vital it 

 is to systematic work to give as many and as accurate estima- 

 tions as possible of the modes of the rocks they describe. In 

 the persalic and dosalic classes of the American Quantitative 

 Classification the ratio of salic to femic constituents is, however, 

 nearly the same as that of felsic to mafic. Similarly the dif- 

 ferences between the ratios of quartz or lenads to felspars, and 

 between the felspars themselves, are negligible from the point of 

 view of classification. This has been ascertained from the 



