August 0, 1009] 



SCIENCE 



163 



do little towards solving any large prob- 

 lems. "When, however, such a mass of 

 fairly homogeneous analyses is at hand as 

 that developed in the laboratories of the 

 survey, it becomes possible to study the 

 geochemistry of the igneous rocks in a 

 broad way, and to shed light upon more 

 than one troublesome question. It is seen, 

 for instance, that the minor constituents of 

 the rocks represent a wide diffusion of 

 many chemical elements, whose presence 

 was formerly regarded as insignificant. 

 From an average of the analyses the mean 

 composition of the igneous crust of the 

 earth can be determined, and it then ap- 

 pears that titanium, hitherto imimportant, 

 is really the ninth in point of abundance 

 among all the chemical elements. Oxygen 

 comes first, forming almost one half of all 

 the matter present in the outer ten miles of 

 the earth's crust ; then follows silicon, about 

 25 per cent. ; then aluminum, iron and cal- 

 cium, in the order named. Magnesium, 

 sodium and potassium follow, in about 

 equal proportions, then hydrogen, provided 

 that we include the ocean in our computa- 

 tions, and then titanium, amounting to 

 about four tenths of one per cent. Man- 

 ganese, phosphorus, sulphur, chlorine, flu- 

 orine and carbon are important; barium, 

 strontium and zirconium are found in 

 readily determinable proportions; vanadi- 

 um, chromium and nickel are by no means 

 negligible. The remaining elements, even 

 including nitrogen, probably amount to less 

 than one per cent., taken all together; a 

 conclusion which is perhaps surprising, but 

 is thoroughly well sustained. The heavy 

 metals, with the sole exception of iron, are, 

 in their total combined amount, statistically 

 less important than titanium alone. 



This statistical analysis of the igneous 

 rocks has been verified by other workers, 

 and its details are fully published else- 

 where.^ It was first made public in 1889, 



and it has since been largely extended and 

 utilized as a basis for other computations. 

 By combining the figures with those repre- 

 senting the known composition and mass of 

 the ocean, it can be shown that a shell of 

 the average igneous rock one third of a mile 

 thick and completely enveloping the globe, 

 would furnish all the sodium of the sea, and 

 Professor Joly has gone even further and 

 used these data as a measure of geological 

 time.^ Given values for the rate at which 

 rivers supply salts to the ocean, and as- 

 suming that rate to have been constant, the 

 calculation is a matter of simple arithmetic. 

 If the salts of the ocean and the alkalies of 

 the sedimentary rocks were all derived from 

 the decomposition of igneous rocks, then a 

 shell of the latter less than half a mile 

 thick would yield all the sodium required. 

 This estimate is a maximum, and serves to 

 show how slightly the surface of the earth 

 has been eroded during geological time. 

 The greater part of this erosion, of course, 

 was concentrated over the continental areas, 

 being probably insignificant in the depths 

 of the ocean. The land erosion may have 

 been as much as two miles in thickness or 

 four times the average for the entire globe. 

 Going still farther, and using composite 

 analyses of the sedimentary rocks, it can be 

 shown that the half mile of decomposition 

 has yielded determinable proportions of 

 shales, sandstones and limestones. The ap- 

 proximate values for a ten-mile thickness 

 of the earth's crust are, igneous rocks, 95 

 per cent. ; shales, 4 per cent. ; sandstones, 

 0.75 per cent. ; limestones, 0.25 per cent. ;'* 



'See especially U. S. Geol. Survey Bull. 330, 

 pp. 21-33. 



* Sci. Trans. Roy. Soc. Dublin, 2d ser., 7, 30, 

 1899. 



^ For other estimates, based upon the same fun- 

 damental data, see \ an Hise, " Treatise on Meta- 

 morphism," p. 940; and Mead, Journ. Geol., 15, 

 238. 



