REPORT ON THE COMPOSITION OE OCEAN-WATER. 
145 
in tlie graduated tube in reference to the scale) is provided with a “ cross ” in the focus 
to make sure of the accuracy of the adjustment. The burette is then read by means of the 
telescope, and the height of the barometer and the temperature of the bath noted down at 
the same time. In other words, the volume of the gas is determined at the temperature of 
the bath, and the pressure B + tt, where B stands for the height of the barometer, and it for 
the small excess of the capillary depression in the side tube over that in the measuring tube. 
The measured gas is now blown back into a test tube, and in it, by means of an iron ladle 
fixed horizontally to a thick iron wire, transferred to the well of the pipette trough. The 
pipette is supposed to have been already charged with a small volume of concentrated 
caustic potash, aud to be otherwise full of mercury. It is needless to describe how the 
gas in being sucked into the bulb of the pipette is deprived of its carbonic acid. 
When the absorption appears to be completed, the gas is blown back into the test tube, 
which during this time has been standing, full of mercury, in the well, the flow of liquid 
reagent being arrested when the latter has come to the safe side of the point where the 
vertical side tube of the small auxiliary reservoir is joined on. Mercury is then run 
into the capillary tube from this reservoir, or with the improved form (Fig. VI.), from the 
absorption-bulb itself, to sweep the gas completely into the test-tube, the liquid reagent 
sucked back into the body of the pipette, the gas in the test-tube transported to the 
trough of the measurer, and thence sucked back into the measurer to be measured again. 
The treatment with potash is repeated to make sure that no carbonic acid has escaped 
absorption. The gas freed from carbonic acid is now mixed with a sufficiency of 
hydrogen, again measured, transferred (by means of the test-tube) to the exploder, 
exploded, taken back into the measurer and again measured. All the several quantities 
of gas are measured moist. The whole sequence of operations, with some practice, takes 
little over half an hour, so that, as a rule, corrections for variations in the temperature or 
pressure are unnecessary. Should such variations occur they are small and easily 
allowed for by calculation. Supposing, for instance, in two measurements, I. and II., the 
temperatures to have been t and t + ( A t) respectively, then, to reduce the gas-volume II. 
from t + ( A t) to t we have 
w 273-K 
Vi- V( (+At )X 273 + ^-h( AO ’ 
for which expression we may safely substitute 
v«=v (<tM x 
and if we keep at hand a table of the reciprocals of all the values 273 + 1 that can be 
expected to occur, the calculation takes very little time. Similarly with the variations in 
pressure. I had originally devised a disgregation-indicator similar to Doyere’s regulator, 
(PHYS. CHEM. CHALL. EXP. PART I. 1884.) A 19 
