REPORT ON THE COMPOSITION OF OCEAN-WATER. 
223 
N. 
N.E. 
E. 
S.E. 
s. 
s.w. 
w. 
NOV. 
1877. 
June, . . ) 
July, . . J 
20-95 
20-71 
20-66 
20-61 
20-56 
20-53 
October, . \ 
November, . f 
21-01 
20-91 
20-80 
20-56 
20-76 
20-78 
Hence it would appear that this natural “constant” is subject to greater variation than 
has hitherto been suspected. But the well-known extensive research of Regnault,* who, 
in 106 samples of air collected in different parts of the world (including 2 from the top 
of Pichincha (4800 metres), and 19 from the polar regions), found the percentage 
to vary only from 20*908 to 20*999, is sufficient to prove that lower values than 20*9 are 
the rare exception, and 21 per cent, may well be adopted for oceanographic purposes 
as holding universally. 
As the pressure of the atmosphere at the sea-level does not differ very greatly from 760 
millimetres, the two gases may be assumed everywhere to press on the ocean, the oxygen 
with a force equivalent to 0*21 times (760— p), the nitrogen with a force equivalent to 
0*79 times (760— p) millimetres of mercury, where p stands for the tension of the 
vapour of water, which of course is very little in the polar regions, while in the tropics 
it may assume greater values up to some 33 millimetres (the tension of steam saturated 
at 30° C.). According to the law of gas absorption, a given volume of sea-water, when 
shaken up with a given volume of air at a given temperature, takes up both gases, the 
dissolved quantity of each being proportional to the product of its coefficient of absorption 
into its partial pressure in the unclissolved residue, and as the coefficient of absorption 
of oxygen is greater than that of nitrogen, the percentage of oxygen in the dissolved air 
is greater, and that in the un dissolved residue is less, than that prevailing in the natural 
air. The ratio of the two percentages obviously depends on the relative volume of air 
used, but we need not trouble ourselves with the somewhat complex general formula, 
because for our purpose it is sufficient to consider the special case which applies to the 
ocean surface, and which presents itself when the shaking is repeated with constantly 
renewed air until the last instalment of air remains unchanged. In this case (which 
for us is the general case) the volume of air dissolved at a given temperature t by 
one litre of sea-water is a quantity A., which is a function only of t and the pressure 
of the atmosphere, and every c.c. of dissolved gas contains n x c.c. of oxygen and 
n 2 c.c. of nitrogen, where n x and n 2 depend only on t, but change very slowly with the 
latter. 
According to my own determinations, as fully reported in the memoir, 
* Comptes rendus, t. xxxiv. p. 863, 1852 ; also Ann. Chim. Phys., ser. 3, t. xxxvi. p. 335. 
