306 ANNUAL REPORT SMITHSONIAN INSTITUTION, 11)20. 



that they are actually mineral aggregates and so as to be able to com- 

 pare them one with another on this basis, in spite of the various pos- 

 sibilities as to mineral composition introduced by the varying condi- 

 tions of solidification, the chemical composition shown by analysis is 

 calculated in terms of mineral composition according to one uniform 

 system; that is, one general assumption as to the minerals that are 

 formed, or may be formed, from the particular magma. In this way, 

 all igneous rocks are comparable and classifiable inter sv, both chemi- 

 cally and mineralogically. The details of the procedure and the re- 

 sults of this system of classification can not be gone into here, but may 

 be looked for elsewhere. 34 It will suffice here to say that the general 

 principles which are considered basal are the so-called " affinities " of 

 the various basic oxides for, first, silica, and, second, alumina, which 

 have been given on a previous page and which are deduced from the 

 general knowledge of rock minerals. Carried out along the lines so 

 laid down the results of the calculation from the data of the chemical 

 analysis give a mineral composition which, although ideal, cor- 

 responds with the actual mineral composition in the great majority of 

 cases. 



Some years ago Iddings 35 pointed out that the density (specific 

 gravity) of a rock as calculated from the calculated mineral composi- 

 tion on the assumption that the rock is holocrystalline, corresponds 

 very closely with the actual density. This fact is of great interest ; 

 partly because of its justification of the fundamental basis of the 

 classification, and also because it thus furnishes a uniform means of 

 comparing the densities, not only of particular rocks, but of the 

 average rocks of different regions, and quite irrespective of such 

 factors as those due to porosity or the presence of glass. Following 

 the suggestion of Iddings, I have calculated the average densities of 

 the continents, the ocean floors (represented by the lavas of the vol- 

 canic islands in the Pacific and the Atlantic), and of the igneous 

 rocks of various countries and comagmatic regions, whose average 

 chemical compositions were calculated by Doctor Clarke from the 

 data in Professional Paper 99. 



34 Cross, Iddings, Pirsson, and Washington, A Quantitative Classification of Igneous 

 Rocks, Chicago, 1903; Washington, H. S., U. S. Geol. Survey, Prof. Paper No. 99, 1917. 

 There is a considerable literature on this and other systems of the classification of rocks. 



83 Iddings, J. P., The Problem of Volcanism, p. 123, 1914. For a later and more detailed 

 statement, see Iddings, Amer. Jour. Sci. (4), xlix, p. 363, 1920. 



