'54 



NA TURE 



[August ig, 1922 



diameter, surrounding which is a silicate shell some 

 900 miles in thickness. The silicate shell is largely 

 ultra-basic and basic. Lying on the thick shell of 

 basaltic rock, which girdles the whole earth, is a com- 

 paratively thin and discontinuous layer of more siliceous 

 rock-matter (granite and gneiss), on which the sedi- 

 mentary rocks have been formed in and around the 

 continental areas. According to the conception of a 

 solid earth so constituted, we clearly have no satis- 

 factory basis for defining the earth's " crust " in terms 

 of the kind of rock of which it is made up, and unless 

 it can be shown that, at some convenient and fairly 

 uniform depth, the rock-substance of the earth under- 

 goes a critical change in its physical condition at the 

 temperature and pressure prevailing there, the only 

 available alternative is to define the " crust " in a 

 more arbitrary manner in terms of depth. 



One way of doing this is to limit its thickness, as 

 some authors do, to that outer portion of the earth of 

 which we may be said to have observational knowledge. 

 The maximum depth at which rocks observable at the 

 surface of the earth have been formed is quite an 

 important geological problem from the economic as 

 well as from the scientific viewpoint, and one that 

 appears never to have been treated adequately. It 

 ma}-, however, perhaps safely be inferred that, by 

 observation of surface geological features, we have a 

 knowledge of the earth down to a depth of more than 

 5 miles, but considerably less than 10 miles. 



In their most recent estimate of the average com- 

 position of the earth's " crust," Drs. F. W. Clarke and 

 H. S. Washington, of the United States Geological 

 Survey, give its average composition down to depths 

 of 10 and 20 miles. The detailed statement of their 

 results has not yet been published, but is to be issued 

 as a Professional Paper by the U.S. Geological Survey. 

 Pending the publication of the detailed report, how- 

 ever, they have given a brief account of their results 

 in the Proceedings of the National Academy of Science 

 (1922, vol. 8, p. 108). 



The method adopted by them for ascertaining the 

 average composition of the Iithosphere is to take the 

 average of trustworthy analyses of igneous rock speci- 

 mens collected from various parts of the earth's surface. 

 They have included 5159 analyses. Averages are 

 given separately for the igneous rocks of the United 

 States ; North America other than the United States, 

 including Greenland ; Central and South America ; 

 Europe ; Africa and Asia ; Australasia, Polynesia and 

 Antarctica. In computing the averages for these 

 various regions the sum total of each constituent was 

 divided by the total number of analyses of specimens 

 from the region dealt with. In calculating the com- 

 position of the earth's " crust " as a whole, the pro- 

 portions of the Iithosphere, hydrosphere and atmosphere 

 for a depth of ten miles were taken as follows : — 

 Iithosphere 93 per cent., hydrosphere 7 per cent., and 

 atmosphere 0-03 per cent. The Iithosphere is assumed 

 to be made up as follows : — igneous rocks, 95 per cent. ; 

 shale, 4 per cent. ; sandstone, 0-75 per cent. ; and 

 limestone, 0-25 per cent. Figures are given for the rarer 

 as well as for the commoner elements. 



The following is the result obtained for the average 

 chemical composition of the igneous rocks of the 

 earth : — 



NO. 2755, VOL. I io] 



SiO„ 



Al.,63 



Fe„0 3 



FeO 



Mg( I 



CaO 



Na 2 



K 2 



HX> + 



CO„ 



Tid 2 



ZrO a 



P 2 o s 

 CI 



Average Igneous Kock. 



Per cent. 

 59-12 

 15-34 

 3-08 

 3-So 

 .3 49 

 5-08 



3-13 

 I-I5 



O-IOI 



1-050 



0-039 



299 



0-048 



The following table shows the estimated percentages 

 of the commoner elements in the Iithosphere, hydro- 

 sphere and atmosphere : — 



Elements in the Lithosphere, Hydrosphere, and 

 Atmosphere. 



Oxygen 



Silicon 



Aluminium 



Iron 



Calcium 



Sodium 



Potassium 



Magnesium 



Hydrogen 



Titanium 



Chlorine 



Phosphorus 



Carbon 



Manganese 



Sulphur 



Barium 



Chromium 



Zirconium 



Vanadium 



Strontium 



Fluorine 



Nickel . 



Nitrogen 



Cerium, Yttrium 



Copper . 



Lithium 



Zinc 



Cobalt . 



Lead 



Boron . 



Glucinum 



ioo-ooo 100000 ioo-ooo ioo-ooo 

 Average composition. Ten-mile crust, hydrosphere, 



and atmosphere. 

 Average composition. 



atmosphere. 

 Average composition. 



sedimentary rocks. 

 Average composition. 



Twenty-mile crust, hydrosphere, 

 Ten-mile crust, igneous and 

 Ten-mile crust. Igneous rocks. 



A serious defect in the method of procedure on which 

 the above estimates by Clarke and Washington are 

 based is that it makes no allowance for the relative 

 magnitude of the different kinds of rock of which the 

 lithosphere is composed. They admit this defect, but 

 claim that any errors involved are likely to be com- 

 pensating (Journ. Franklin Inst., 1920, vol. 190, p. 770). 

 Their claim can scarcely be allowed, however, even for 

 the outer 10 miles of the " crust," and still less can it 

 be allowed down to a depth of 20 miles. 



As to the relative proportions of the rocks composing 



