EE POET ON THE: COMPOSITION OF OCEAXAVATEK. 
;7 
determine the. halogen ip certain opganiq. bromine compounds, (by' combustion with lime 
or tri-sodic phosphate'"), we obtained veryainsatisfgc^or^ re^plt^, showng that in .the 
presence of large quantities of pyro-phosphoric acid, or even of nitrate of lime, the titration 
by sulphocyanate solution could not be relied on. About the same time my friend Pro- 
fessor Crum Brown informed me that the well-known colour of ferric sulphocyanate was 
destroyed by large quantities of magnesia salts. I therefore instituted a series of 
experiments in order to see to what extent our determinations of chlorine in sea-water 
must be assumed to have been vitiated by the presence in the water of magnesia and 
lime salts. The following two standard solutions were prepared : — 
1. 15 '65 grms. of pure chloride of sodium were dissolved in water and diluted to 
(“ 250 c.c.” =) 26P21 grms. One grm. of solution contained 59‘91 mgrms. of chloride 
of sodium. 
2. 10 ’5 5 grms. of crystallised sulphate of magnesium and potassium (MgK 2 S 2 0 8 6H 2 0) 
were dissolved in water, and 0‘58 grm. of pure carbonate of lime dissolved in dilute 
nitric acid, the two solutions mixed and diluted to 1 litre. 
One-half c.c. of solution 1 and two c.c. of solution 2 form a fair equivalent, in respect 
of contents in chlorine, lime, and magnesia, to 1 c.c. of sea- water. 
310‘4 mgrms. of sodium chloride (in the shape of solution 1) were titrated by Yolhard’s 
method. The resulting mixture was now mixed with 40 c.c. of solution 2, i.e., with about 
twice the quantity of lime and magnesia which, in sea-water, is associated with 310 
mgrms. of chloride of sodium, a little extra silver added, and the point of saturation 
again determined by the method of repetition. 
Sodium chloride found in solution 1 alone, . 310'0 
After addition of solution 2, ... . 309 - 88 
In the following series of experiments the impurities were added to the original salt- 
solution before addition of the nitrate of silver. The same salt-solution and the same 
silver-solution (both approximately deci-normal) were used in the four analyses. 
Analysis (1) (2) (3) (4) 
50 c.c. of Salt, weighing 1ST = . 50-274 
grms. 
50'260 grms. 
50-265 grms. 
50"263 grms. = 1ST. 
Lime as nitrate = . .0 
5 ?. 
0 „ 
28 mgrms. 
28 mgrms. 
Magnesia as sulphate = . .0 
?? 
o 
100 „ 
100 „ 
Excess of deci-normal silver-solution | 
found by titration = . . / 
5 ? 
0-413 „ 
0'442 grms. 
0’435 grms. 
Corrected silver- solution in grms. = S = 5L405 
V 
51-387 „ 
51-399 „ 
51-387 „ =S 
S-N = . . . • 1-131 
?? 
1-127 „ 
1-134 „ 
1-124 „ 
As a matter of arithmetic these values S 
— N may be 
assumed as 
holding for the 
* How we came to use the latter reagent is of no consequence here; it will perhaps be explained in a special 
memoir. . " 
