270 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1920. 



sibility, so that we have reason to think that there is also, toward the 

 center, a change in actual substance. 



6. It has been often suggested, and is more or less commonly be- 

 lieved, from consideration of the density and magnetic character of 

 the earth and from the composition of many meteorites, that part, 

 at least, of the interior is composed essentially of iron, or of nickel- 

 iron alloy similar to those which constitute the iron meteorites. 



Leaving the interior of the earth for the present we may con- 

 centrate our attention on the outer shell — the so-called " crust " — 

 which is the only portion that is directly open to our study, and 

 which has been compared, with some justice, to a covering of slag 

 or scoria over the interior. In dealing with this we shall consider 

 only its chemical characters, with, toward the end of the paper, 

 some relations between these and the densities of rocks. 



The thickness of this crust is, of course, unknown, probably not 

 uniform, and presumably indeterminate. Following Dr. F. W; 

 Clarke, we may assume, for purposes of computation, an approxi- 

 mate thickness of 10 miles (16 kilometers), this being about the 

 (minimum) aggregate thickness of all known rocks and deposits 

 of the various geological ages that have become exposed to our ob- 

 servation and study through movements in the crust. Incidentally, 

 it about equals the sum of the highest land elevation and the greatest 

 oceanic depth, though no causal nexus is apparent. 



This solid crust is made up almost wholly of igneous rock — that 

 is, rock that has solidified from a hot, liquid (" molten ") condition, 

 either as " plutonic " rocks, at different depths beneath the surface, 

 or in the form of lava flows at the surface. Assuming a thickness 

 of 10 miles, Doctor Clarke 2 has estimated the rock composition of 

 the crust to be about as follows : 



Per cent. 



Igneous rocks 95. 



Shales 4. 



Sandstones 0. 75 



Limestones 0. 25 



100.00 

 Such masses as coal beds or salt and ore deposits are of negligible 



magnitude in studying the chemistry of the crust as a whole, as it 



is purposed to do here, though their presence is of some significance. 



The amount of the coating of soil is absolutely negligible from this 



point of view. 

 When we take into consideration the oceans and the atmosphere, 



Clarke estimates the lithosphere at 93 per cent, the hydrosphere at 



2 Clarke, F. W., The data of geochemistry, U. S. Gool. Survey, Bull. No. 695, p. 33, 1920. 

 The proportion of igneous rock would be still greater with greater assumed thickness of 

 crust. 



