REPORT ON THE COMPOSITION OF OCEAN-WATER. 
171 
at nearly the same temperature to one and the same intermediate temperature. The 
following eight groups were thus formed : — 
Group I . — 7 experiments. Adopted standard temperature (t) = 14 0, 5. Mean value 
of X = 21 '32 ( -r 1000). This divisor must be supposed to be appended to all the Vs 
down to group VIII. 
Group II. — 2 experiments; £ = 13 o, 0; X = 21'68. 
Group III— -l experiment; £ = 20°'4; X=19'05 (not allowed any influence in the 
calculation). 
Group IV. — 2 experiments; i=18 o, 0; X = 20‘09. 
Group V. — 2 experiments; t = 29°'5 ; X=16 - 40. 
Group VI.— 4 experiments ; t= 1°'0 ; X = 28'94. 
Group VII. — 3 experiments; t= 5°*5 ; X = 25’72. 
Group VIII. — 2 experiments; £ = 45 o, 0; X=13’98. 
A few preliminary calculations based on convenient ordinates taken from the curve 
showed that an equation of the form X = X 0 — at + bt 2 gave no sufficient approximation. A 
formula of the form 
worked better ; yet the difference between observed and calculated values in some cases 
was inconveniently great. When the same function was treated according to the method 
of least squares, quite satisfactory results were obtained. In this final calculation 
groups I. and VI. were each allowed two votes ; group VIII., half a vote ; the rest one 
vote each, except group III., which was not taken into account at all. The resulting 
formula was 
whence we have — 
1000 x X = 
1119-4 
37-9 + t’ 
Group. 
t. 
A Calculated. 
A Found. 
VI. 
r-o 
28-78 
28-94 
VII. 
5°-5 
25-79 
25-72 
II. 
13°-0 
21-99 
21-68 
I. 
14°-5 
21-36 
21-32 
IV. 
18°-0 
20-03 
20-09 
III. 
20°-4 
19-20 
19-05 
V. 
29°-5 
16-61 
16-40 
VIII. 
45°-0 
13-50 
13-98 
In a similar manner the values for n x (the volumes of oxygen in unit-volume of 
absorbed gas) were dealt with. The resulting formula, calculated by the method of 
the least squares, was 
100 x? 2 1 = 34-693-0-04545i. 
