REPORT ON THE COMPOSITION OE OCEAN- WATER. 
183 
theoretical values calculated from the observed temperature by means of our formulae. 
(4) The oxygen, we should say, a priori, can never quite come up to the calculated 
value corresponding to t 0 , because it is constantly being utilised in processes of oxidation 
and respiration going on within the water. The nitrogen is not subject to this dis- 
turbing influence ; hence we should expect it to accommodate itself more closely to 
the law of gas absorption. And yet, on comparing the volumes of nitrogen found 
with the volumes calculated for the observed t 0 , we find the latter to be in general 
greater than the former. These nitrogen deficits were given in Table XI. Column X. ; 
in Table XIY. I have enumerated these nitrogen-deficits in the order of their magnitudes. 
I have vainly endeavoured to find some relation between them on the one hand, and 
temperature or geographical position on the other. Considering the great frequency 
of values from 0‘4 to 0'5, I am inclined to assume that, in virtue of some general 
constant influence,* this deficit tends to assume some value like, say, 0'42, subject to 
variation in either direction, as seen in the table, which, besides, exhibits numerous 
cases of negative values, i.e., of nitrogen excesses. These latter, although in accordance 
with what we said under (3), are explained more plausibly as resulting from an 
intermixture of surface water with deep-sea water richer in nitrogen. If the variations in 
the nitrogen deficits were owing to accidental causes, then counting off eight entries (i.e., ^ 
of 0‘26 x 62) from the neutral point either way, we should arrive at half the value of the 
probable “error,” and we indeed arrive in either case at the value ±0T, so that the 
probable “ error” would appear to be= ±0'2 = r. But adopting this value, we have 
Deviations under ± 
Number of Cases counted. 
+ Deviations. - Deviations. 
Total. 
Calculated. 
\ r = 
■1 
8 
8 
16 
■2 
11 
9 
31 
f r = 
•3 
13 
11 
42-6 
2 r = 
•4 
18 
14 
51 
3r = 
•6 
21 
21 
59-3 
4 r = 
•8 
24 
26 
61-6 
5 r = 
DO 
25 
30 
62 
GO 
31 
31 
62 
The nitrogen-deficits may be represented by naming the temperature at which a 
water would have to be completely saturated with air to take up the observed volume of 
nitrogen per litre. These values h are given in the last column of Table XIV. I have 
not utilised these values t x , but assuming the actual values of nitrogen to have been 
brought about by incomplete (or super-) saturation at the observed temperature t 0 , I have 
calculated the volumes of oxygen corresponding to the observed quantities of nitrogen 
'• f In air of 760 mm., fully saturated with water at 22°, the dry -air pressure is only (1 — - 026) 760 mm., corresponding 
to a nitrogen deficit of 0'26 or 10 units. 
