THE ROSIWAL METHOD FOR MINERALS 215 



the hornblende and was probably included in the analyzed horn- 

 blendic material. The two minerals are, therefore, calculated 

 together. There can be no question about the success of the 

 Rosiwal determinations in this case. Cross, Iddings, Pirsson, and 

 Washington' say: "It is possible .... to determine very closely 

 the proportions of the minerals present " This determina- 

 tion proves either the accuracy of the Rosiwal method or, since it 

 is only a single example, an almost exact compensating error. 



Among other petrographers who have found the microscopic 

 methods of sufficient accuracy are Barrell,^ who determined his 

 constituents by a somewhat simpler, though less accurate, method. 

 He says: 



The field of view of the microscope is divided into quadrants by the cross- 

 hairs, and these are mentally divided into sectors which are thirds, fifths, or 

 smaller fractions. By taking such a power of objective that the component 

 to be estimated is represented in the field of view by a fair number of crystals, 

 the latter can be mentally collected together and packed into one quadrant, 



and the fraction which it fills estimated In the case of holocrystalline 



rocks, where the writer has been enabled to check this method of estimating 

 the mineral and chemical composition by comparison with laboratory analyses, 

 it is found that the error in any one element should he less than i per cent and is often 

 less than one-half per cent. [The italics are ours.} 



Robinson^ compared the Rosiwal measurements of a porphyritic 

 dacite with the composition calculated from an analysis and found 

 that 



there is a difference of 3 . 2 per cent between the total salic components, includ- 

 ing the excess of alumina in the calculated mode, and of 2 . 6 per cent between 



the dark components The comparison will give an idea of the accuracy 



of the Rosiwal method as applied to porphyritic lavas containing small amounts 

 of minerals of highly variable size. 



Tyrrell'' says: 



The graphic method of quantitative mineral measurement invented by 

 Rosiwal jdelds sufficiently accurate results for classificatory purposes in the 



' Ihid., p. 204. 



^Joseph Barrell, "Microscopical Petrography of the Elkhorn Mining District, 

 Jefferson County, Montana," XXII An. U.S. Geol. Siirv., Pt. II, 1901, pp. 512-13. 



3 Henry Hollister Robinson, "The San Franciscan Volcanic Field, Arizona," 

 U.S. Geol. Surv., Prof. Paper y6, 1913, p. 121. 



1 G. W. Tyrrell, "A Review of Igneous Rock Classifications," Science Progress, 

 1914, pp. 74-75- 



