REPOET ON THE COMPOSITION OF OCEAN-WATER. 
117 
of air until, according to the laws of absorptiometric exchange, the gas remnant in the 
liquid ought to have been reduced to a small fraction of a milligram. The actual carbonic 
acid was then determined by manipulating the solution in my modification of Classen’s 
apparatus like an acidified solution of a carbonate. The sulphate of magnesia solution 
was prepared from a salt purified as explained on page 106, and contained 13 '5 grms. of 
MgS0 4 7 H 2 0 per litre. The sulphate of lime solution was prepared by neutralising a 
measured volume of normal sulphuric acid (lc.c. = £[H 2 S0 4 ] mgrms.) with the calculated 
proportion of pure precipitated carbonate of lime, diluting to a convenient volume and 
filtering. To determine the strength of the solution, a measured volume of it was pre- 
cipitated by chloride of barium, and the sulphate of baryta weighed. It was found to 
correspond to 1*973 grms. of (anhydrous) sulphate of lime per litre. 
The results of the carbonic acid determinations were as follows : — 
Milligrams of Carbonic Acid per litre of Liquid. 
I. II. III. 
Substance in solution, Nil. Magnesium sulpliate Calcium sulphate 
Carbonic acid, 2‘5 2 ’8 3T 
From these experiments it would appear that each of the three solutions contained a 
small quantity of its gas in a state of qmsf-chemical combination, and that this quantity 
was the same in the three cases. But it is extremely difficult in such experiments to 
exclude every trace of adventitious carbonic acid, and the carbonic acid actually extracted 
from the 300 c.c. of licquid operated upon amounted to only 0*76 to 0*94 mgrm. ; no more 
than the excess of carbonic acid which we obtained in the three experiments with sulphate 
of magnesia and carbonic acid added as carbonate of soda (page 116). Quantities not 
much, if at all, below those found in the present case I often obtained in blank trials 
with absolutely carbonic acid free materials. I suspect that these irrepressible traces of 
carbonic acid came out of the india-rubber stoppers used for making the gas-tight joints. 
In any case my conclusion from the present experiment is that the alleged affinity of 
carbonic acid for sulphate solutions (or pure water) has no existence. 
The following table gives the total quantity of carbonic acid actually found in a 
selection of Challenger waters, contrasted with the carbonic acid calculated from the 
alkalinity on the assumption of the “ alkali ” being all bicarbonate unmixed with free 
carbonic acid. In each case 250 c.c. of water were taken for analysis, and the carbonic 
acid titrated with baryta-water and hydrochloric acid, 1 c.c. of each being equivalent to 
