REPORT ON THE COMPOSITION OF OCEAN-WATER. 
211 
The action of hydrochloric acid (and consequently also that of dissociating chloride of 
magnesium) on a mixture of carbonate and sulphate of bartya cannot be presumed to be 
absolutely nil ; besides, as part of the bicarbonate in sea-water must be presumed to exist 
as lime and magnesia salt, it is questionable whether all the carbonic acid of the normal 
carbonate was always precipitated as carbonate of baryta ; hence, Mr. Buchanan’s 
method cannot be relied on as having always eliminated exactly that part of the 
carbonic acid which is present over and above that of the normal carbonate, and we 
cannot be sure even of the error always lying in the same direction. My critical trials 
tend to confirm this suspicion. But these trials, so far as they accommodated themselves 
in all strictness to Buchanan’s method, are very few ; while I have no doubt that 
Buchanan’s experiments on board ship were executed with his characteristic exactness 
and with business-like uniformity. Hence his results (as reproduced in my memoir) are, 
after all, very valuable ; because, apart from their giving, in the worst case, a minimum 
limit for the total, they may fairly be assumed to give an approximation to the loose 
carbonic acid (meaning what was there over and above the normal carbonate). 
Admitting this, their general result may be formulated by saying that ocean-water 
from any place or depth contains its surplus base in the form of normal carbonate com- 
bined with additional carbonic acid, which latter in the majority of cases falls short of, 
in a minority of cases comes up to, and very rarely exceeds, that which would produce 
bicarbonate (RHC0 3 ) . 
I might now proceed to a more detailed discussion of Buchanan’s numbers, but 
prefer first to refer to a few synthetical experiments of my own, the primary object 
of which was to ascertain whether or not it was expedient for me to endeavour to 
add to Buchanan’s work by determinations of the carbonic acid in my numerous samples 
of Challenger water. These experiments of mine are fully described in my memoir 
(page 108 et seq.) ; I content myself here with stating the most important result, wdiich 
was that an imitation sea-w r ater, prepared from pure common salt, Epsom salt, and bicar- 
bonate of magnesia (which I then rightly or wrongly looked upon as sufficiently repre- 
senting an ocean-water), when shaken with air in the summer, at w T hich season our 
experiments was made, loses carbonic acid, and on subsequent repetitions of the opera- 
tion continues doing so, the bicarbonate being thus brought nearer and nearer to the 
composition of normal carbonate. After this result I saw that there would be no use 
in my analysing many of my water samples for carbonic acid, and it was more for the 
purpose of satisfying my curiosity concerning the extent to which the bicarbonate- 
dissociation in my long preserved samples of sea-water had gone, than with any 
oceanographic object, that I carried out those few determinations which are tabulated 
on page 118 of my memoir. To my surprise, five out of the thirteen samples analysed 
still contained their surplus base as fully saturated bicarbonate.* 
* A noteworthy result, because in a sense it confirms my critical trials on Buchanan’s method, which tend to show 
that this method, even as a mode of measuring the loose carbonic acid, is infected with a negative error. 
