138 
CLARKE. 
alumina. The silica, then, may he reduced by about one- 
fourth the percentage of the titanic oxide, and the alumina 
by the other fourth, plus half the values assigned to P 2 0 5 
and Cr 2 0 3 . Making these corrections, reducing to elemen¬ 
tary form, and recalculating to 100 per cent., we get a rough 
approximation to the mean composition of the solid crust of 
the earth. To this approximation the other unestimated 
elements may be regarded as future corrections of very small 
amount. Combining this result with the mean composition 
of the ocean, and including 0*02 per cent, for the nitrogen 
of the air, we get the final column to illustrate the abundance 
of the elements so far as at present known. Quantities less 
than 0*01 per cent, are left out of consideration. 
'iSolid crust, 98 f 0 - 
Ocean, 7 
Mean , incl. air. 
Oxygen. _ 
47-29 
85-79 
49-98 
Silicon - 
27-21 
_ 
25-30 
Aluminum _ 
7-81 
7-26 
Iron _ __ _ 
5-46 
5-08 
Calcium ___ 
3-77 
•05 
3-51 
Magnesium _ _ 
2-68 
•14 
2-50 
Sodium. __ __ 
2-36 
1-14 
2-28 
Potassium _ __ 
2-40 
•04 
2-23 
Hydrogen. __ 
•21 
10-67 
•94 
Titanium. _ _ ... 
•33 
_ 
•30 
Carbon _ _ _ 
•22 
•002 
•21 
Chlorine _ _ _ 
•01 
2-07 
} ' 16 
Bromine _ _ _ 
•008 
Phosphorus _ 
•10 
•09 
Manganese.. _ _ 
•08 
— 
•07 
Sulphur __ ..... __ 
•03+ 
•09 
•04+ 
Barium _ _ _ 
•03 
_ 
•03 
Nitrogen . __ 
i __ 
•02 
Chromium ... .. 
•01 
•01 
100-00 
100-000 
10000 
Nineteen of the 
elements are here provided 
for with very 
varying degrees of probability, although their order in the 
last column is presumably correct. The uncertainty may 
reach one per cent, in the cases of silicon and iron, one-half 
as much with aluminum and oxygen, and is proportionally 
