1892 - 93 .] Dr J. Gibson on Gom'position of Sea- Water. 317 
must be noted that in calculating Dittmar added on 0"04 
(water = 1000) to Buchanan’s specific gravities in order to undo the 
correction to vacuo which these contain, and so bring the results 
into accord with his experimental conditions. These differences, as 
they stand in Dittmar’s Report, greatly exceed the limits of the 
experimental errors to which Buchanan and Dittmar respectively 
admit their methods to be liable. Buchanan estimates the experi- 
mental error in his determination at not more than ’05 (water at 
4° C. = 1000). This corresponds to ± *035 in x “X? and Dittmar 
calculates his errors in his determinations of \ at certainly not more 
than ± 0'03. Errors, therefore, of ± 0*065 might reasonably be 
expected. ISTow excluding, as Dittmar did, some five or six alto- 
gether abnormal values for x' - x> these differences ranging 
from -1- 0*245 to - 0*231. Errors of -f- 0*100 and upwards occurring 
frequently, while errors of -0*100 are of so very frequent occur- 
rence as to strike the eye at once on a cursory inspection of the 
table. Dittmar calculates that a constant correction of - 0*042 
should be applied to these differences (x' — x) they stand in his 
Report, and formulates the net result of his inquiry by putting 
X ~ X~ ~ 0*042 + S , 
“ where S is a variable quantity, of which the chances are even that 
it is less or greater than 0*06, and about 8 against 2 that it is less than 
0*12.” Very unfortunately Dittmar, for some reason or other, did 
not determine experimentally the density of those waters in which 
he determined x- Having failed to trace any connection between 
geographical distribution and the slight differences in chemical 
composition which he observed in the case of 77 “Challenger” 
samples, which he subjected to a full analysis, and having come 
to the conclusion that such differences, if any, were negligable, 
it is perhaps not to be wondered at that this task was not under- 
taken. One thing Dittmar actually did was to make an invalu- 
able investigation into the thermal expansion of sea-water, which 
made it possible to reduce the specific gravity determinations 
made by Mr J. Y. Buchanan to a common standard temperature. 
The standard temperature adopted was 15*56° C. or 60° F. 
Some years ago, at a meeting of this Society, I gave it as my 
opinion that these difierences (x - x) by Dittmar should not 
