REPOET ON THE COMPOSITION OF OCEAN- WATER. 
105 
portion at least of the carbonic acid which Buchanan had liberated, from its combination 
with sulphates as he thought, was only eliminated from the bicarbonates, thus : — 
R 2 0.H 2 0.IC0 2 + BaCl 2 = BaOC0 2 + 2RC1 + H 2 0 + C0 2 . 
A subsequent critical examination of Buchanan’s paper in the Royal Society’s Proceedings* 
gave me the conviction that the alleged affinity between sulphates and carbonic acid is 
founded upon an incorrect interpretation of what are probably in themselves correct 
observations. 
When at a later date I resumed the carbonic acid question, I conceived the idea of 
determining both this acid and the alkalinity by the following gcisometric method : — 
Take two portions of the given water, and in one convert all the “free” base into bicar- 
bonate by passing in carbonic acid, and then removing the excess by shaking with air. 
Then eliminate the carbonic acid from each by boiling it out in a Jacobsen flask, and 
determining it gasometrically in the boiled out gas, taking care to use hydrochloric acid 
instead of water for producing the vacuum in the bulb and gas-collecting tube. In 
order to try this method a Jacobsen’s flask t of 800 c.c. capacity was charged with gas-free 
water and a known weight of carbonate of soda, equal to about 10 c.c. of carbonic acid, 
and the gas eliminated as just explained. This experiment was made five times, and in 
each case the carbonic acid gas found was short of the calculated volume by about one 
cubic-centimetre. Just as if, after all the boiling under reduced pressure, so much of the 
carbonic acid had been retained by the hot acid liquid. An attempt to prove its 
presence in the residual liquid by addition of baryta water gave no result, because 
enough of the substance of the glass had dissolved in the liquor to cause the formation 
of a precipitate not carbonate of baryta ; and, besides, a direct synthetic trial showed 
that the detection of 2 mgrms. of carbonic acid in 800 c.c. of water is altogether beyond 
the range of the baryta test. I therefore tried to solve the question synthetically :■ — -Two 
litres of ordinary (aerated) distilled water were boiled down in a narrow-necked flask to 
about one-half of the original volume, and a quantity of clear baryta water was then run 
in from a protected Mohr’s burette, without interrupting the boiling, which was continued 
until only a few cubic-centimetres remained. The flask was now allowed to cool and 
suck in air free from carbonic acid. Finally the contents were acidified with hydrochloric 
acid, boiled in a current of air free from carbonic acid, and the out-going air allowed to 
bubble through clear baryta-water. There was a slight but distinct precipitate of car- 
bonate of baryta produced, and a similar result was obtained in two or three repetitions 
of the experiment. It was not so much in consequence of this result, as on account of 
the pressure of other work, that the carbonic acid question was again put aside for a 
* Proc. Roy. Soc., vol. xxii. pp. 192-196, and pp. 483-495 ; 1874. 
t A drawing and description of the apparatus will he found in the chapter “ On the Absorbed Air in Ocean V aUr.’ 
(PHYS. CHEM. OHALL. EXP. PART I. 1884.) H 
