384 H. 8. WASHINGTON ISOSTASY AND ROCK DENSITY 



are merely shifted up or down, but the relations remain practically 

 unchanged. 



The "norm'' of the chemical average is then calculated according to the 

 now well known methods employed in the quantitative system oi# classifi- 

 cation of rocks. s The chemical composition is thus reduced to terms of 

 the mineral molecules which the chemical constituents present may, and 

 in most cases do, form on solidification of the molten magma as a non- 

 crystalline rock. 



Several points of special importance are to be noted. In the first place, 

 all rocks or rock averages are treated alike by a uniform method which 

 does not admit of the influence of the personal equation ; the results are 

 thus mutually consistent, having been arrived at by identical means. In 

 the next place, the mineral molecules to which the calculations lead are 

 few and simple, and, which is of still greater importance, they differ 

 among the various averages only in their relative amounts, and thus refer 

 all the averages treated to the same conditions of pressure, rate of cooling, 

 and other possible factors. Thus the same magma which on the surface 

 would yield a mixture of leucite and sodic orthoclase would at depth, 

 under intrusive conditions, yield a mixture of nephelite and purely 

 potassic orthoclase, the two mixtures having different densities. Under 

 the system here employed both would be referred to the same mixture 

 (the second), differing only in the proportions of the constituents; so 

 that the disturbing factor of qualitatively varying mineral composition 

 is eliminated and their respective densities are strictly comparable 

 inter se. 



Again, and a point of very considerable importance, all differences due 

 to texture, such as the presence or absence of glass, porosity or cracks, 

 and other such features caused by different states of physical solidifica- 

 tion, are wholly eliminated. Glass, when present, lowers the density of a 

 rock, as do pore spaces; so that the presence of these would introduce a 

 serious error, making the rock lighter than a dense noncrystalline rock 

 of identical chemical composition. Inasmuch as we know that glass is 

 not present at all, or at most in but insignificant amounts, in the deeply 

 seated rocks, in which also the great pressures would tend to obliterate 

 pore spaces, calculations based on the norm yield results which apply 

 best to the rocks as they occur at great depth — that is, to the greater 



s Cf. Cross. Iddings, Pirsson, and Washington : Quant. Class, igneous rocks. Chicago. 

 1903, pp. 186-203. 



Iddings : Igneous rocks, vol. i, 1909, pp. 433-444. 



Finlay : Introduction to study of igneous rocks. 1913, pp. 149-192. 



Washington : U. S. Geol. Survey Professional Paper no. 99, 1917, pp. 1162-1165. 



