• ) 



90 H. S. WASHINGTON ISOSTASY AND ROCK DENSITY 



that, while a few of the averages are admittedly defective because of the 

 paucity or the selective character of the data, yet the great majority (in- 

 cluding all of the averages given above except those for Antarctica and 

 the Atlantic) are probably fairly representative. In any case, as has been 

 said before and as may be repeated, the averages given in these pages are 

 based on the most recent and the largest set of reliable data that are as yet 

 at hand and are the only such averages available, so that they may be 

 accepted provisionally as a proper basis for some discussion and con- 

 clusions. 



The average density of the earth's crust, so far as it is exposed to our 

 direct observation and as derived from the chemical averages given here, 

 is 2.792, calculated on a water-free basis, while it would be 2.737 if allow- 

 ance is made for about 1 per cent of "water on ignition" as found in the 

 analyses. 



The § a value is notably higher and the 8 W value is lower than that 

 given in a former paper, 15 namely, 2.77, in which the water was reckoned. 

 It is thought, however, that the present value (2.792) is the more correct, 

 for several reasons. The selection of values for the mineral densities has 

 been more careful, the water is disregarded, and this 8 a value approaches 

 more closely to that which is obtained for the average when the conti- 

 nental and oceanic averages are weighted for areas. The result of one 

 such calculation, in which the areas of the Pacific and South Atlantic 

 basins (not the islands alone) were taken into account along with the 

 respective areas of the several continents, gave the density as 2.8506 

 (sp. vol. = .3508). If the areas of the North Atlantic, Indian, and 

 Antarctic oceans were also considered, with proper allowance for their 

 prevailingly basaltic rocks, the density would be still higher, probably 

 about 2.95 — that is, slightly below the value given here for the Pacific. 



It will be seen that the average chemical compositions of the different 

 continents and ocean basins (represented by their islands) differ among 

 themselves, but in different ways and to various extents. The averages 

 for Europe and North America most closely resemble that for the earth ; 

 that for South America is much like these, but is slightly higher in silica, 

 while the maximum silica is reached in Asia. The Australian average 

 resembles the preceding in its silica, but is distinctly higher in iron oxides 

 and lower in alkalies, while that for Africa, which is distinctly lower in 

 silica, is decidedly higher in soda. With Antarctica we get a much more 

 femic composition, which is still more marked in the Atlantic. and espe- 

 cially the Pacific averages. It is not the place here to discuss these and 



15 H. S. Washington : Jour. Franklin Inst., vol. 190, 1920, p. 807. 



