INTRODUCTION 377 



gravity, after correcting for latitude and elevation, and the converse ; and 

 it has long been recognized that the material underlying the ocean floors 

 is much denser than that which underlies the continental masses. Isos- 

 tasy therefore demands that the crust be non-homogeneous; in other 

 words, the material beneath the various surficial areas of different alti- 

 tudes must vary in chemical and mineralogical composition ; for only thus 

 can we explain the differences in density demanded by the theory and by 

 the gravity measurements. It is true that the material beneath an area 

 of high relief, such as a mountain range or a high plateau, is' under 

 greater pressure than that which underlies an ocean floor. But, if we 

 could assume the same kind of rock in each case, the change in density 

 brought about by the differing pressures is not only opposite in direction 

 to that demanded by observation and theory, but is of an order of magni- 

 tude far too small to be appreciable, as we know from studies on the com- 

 pressibility of rocks. 3 Similarly, rise or fall of the isogeotherms, which 

 would also bring about differences in density, would likewise be inade- 

 quate to produce differences in density which would correspond with 

 what have been observed. 



The earth^s crust is composed of igneous rocks to the extent of about 

 95 per cent if it is assumed to be 10 miles thick, according to Clarke's 4 

 estimate, or about 97.5 if it is 20 miles thick. The metamorphic rocks, 

 gneisses and schists, are reckoned in with those of igneous origin, while 

 the remaining 5 or, respectively, 2.5 per cent of sedimentary rocks (shales, 

 sandstones, and limestones) form a very thin veneer on the surface. In 

 dealing with a crust of such a thickness as is involved in the idea of 

 isostasy and the depth of the "zone of compensation," therefore, we are 

 justified in assuming it to be wholly of igneous rock. Now, study of the 

 igneous rocks of the earth, of which about 10 miles or so in depth are 

 accessible to our observation, shows that the chemical and mineral hetero- 

 geneity demanded by the theory of isostasy to account for the varying 

 densities does actually exist; furthermore, that the differences in com- 

 position are of the order of magnitude to meet the needs of isostasy and 

 are uniformly in the right direction. 



It is the object of the present paper to discuss the bearing of petrology 

 on the problem of isostasy ; to explain briefly the petrological heteroge- 

 neity, both qualitative and quantitative, involved in the notion of comag- 

 matic regions; to describe a method for calculating the areal densities 



3 Dr. L. H. Adams has kindly calculated that, assuming at sea level a non-porous 

 granodiorite of density 2.80 and compressibility 0=1.9 X 1CH 5 per megabar. the density 

 at a depth of 60 kilometers (the isopiestic level) would he 2.889. 



4 F. W. Clarke : Data of geochemistry. U. S. Geoi. Survey Bull. G95. 1920. p. 33. 



