REPORT ON THE COMPOSITION OF OCEAN-WATER. 
1G1 
actual work, and not intended for publication, expanded into a ratlier lengthy and 
troublesome research ; and before going further, I at once acknowledge the very able and 
zealous manner in which I was supported by Mr. Robert Lennox throughout its unex- 
pectedly slow progress. 
Bunsen, in determining the coefficients named, proceeded as follows : — He first 
determined the coefficient for nitrogen synthetically by means of his absorptiometer ; 
that is, by reading the contraction suffered by a measured volume of the gas when shaken 
over mercury with a measured volume of previously boiled water at a definite tempera- 
ture and pressure. The same method when applied to oxygen gave bad results, because 
the oxygen was found to act promptly, even on the purest mercury, when shaken up with 
it and water. To avoid this difficulty, Bunsen, by a series of separate experiments, 
determined the composition of the gas which is absorbed by water when it is again 
and again shaken up with fresh air free from carbonic acid under the ordinary pressure and 
at a certain constant temperature, until absorptiometric equilibrium is sure to be established. 
Large known volumes of such saturated water were boiled to expel the gases, which were 
then analysed. The result was that at all temperatures from 0° to 23° C. the absorbed air 
contained 34'91 volumes of oxygen and 65'09 volumes of nitrogen in 100 volumes.* 
Hence it was easy to calculate the coefficient of absorption of oxygen from that found for 
nitrogen at the same temperature. 
This method, of course, would apply to sea-water as well as to pure, but I had not a 
Bunsen’s absorptiometer in my possession, and besides I thought that, for my special 
purpose, I had better use a method which adapts itself more directly to the one which 
Mr. Buchanan had used for extracting the dissolved gases from his sea-water samples. 
I accordingly drew up the following general scheme for the determinations. Starting 
with a sufficiency (say from 1 to 1-| litres) of water which in general should be free from 
gases, shake up this water repeatedly, at a constant temperature (maintained by means of 
a water-bath), with pure nitrogen, or pure oxygen, or with air deprived of carbonic acid, — 
in general with a mixture of m 1 volume of oxygen, and m 2 = (1 —m^ volume of nitrogen, 
taking care to renew the gas-atmosphere in the bottle after each shaking, and so go on until 
absorptiometric equilibrium has certainly been established. Then boil a known volume 
(from f to 1 litre) of the saturated water in a Jacobsen’s apparatus, collect, measure, and 
analyse the extracted gas. Taking t Q and B 0 as symbols for the temperature and baro- 
metric pressure at the time of the absorption ; t x and B x the corresponding values at the 
time of the analysis (which, of course, was always carried out by means of my apparatus, 
in which all gas volumes were read moist at the pressure Bj + tt mm., see page 145 ) ; t 0 
and Tj the tension of vapour of water at t 0 and t x respectively ; supposing wq and nu 
to be the volumes of oxygen and nitrogen in a unit volume of the ctir operated upon 
(which includes the cases of pure oxygen or nitrogen), n i and n 2 to have a similar meaning 
* Bunsen in his book reports three experiments — one at 1°, one at 13°, one at 23 C. 
(PHYS. CHEM. CHALL. EXP.- — PART 1. 1884.) 
E t 
