THE VOYAGE OF H.M.S. CHALLENGER. 
244 
with a quantity of acid deci-normal solution of nitrate of silver, corresponding to 3'9 per cent. 
,.f the total halogen present. From the report in the memoir it is easily calculated that 
a mixture of O' 1473 grm. of bromide and 2-9510 grins, of chloride of silver would be a very 
close imitation of the “ first precipitates” obtained in the bottom water analyses. In the 
test analyses, accordingly, quantities of the two haloids nearly equal to these were operated 
upon. The chloride of silver was first weighed out in the boat, fused, and weighed again 
within its tube. The proper quantity of bromide of silver (dried at 150° C.) was then 
added from a preparation-tube and spread over the surface of the chloride. 
The two haloids were then fused together, and their conjoint weight redetermined ; 
and the mixture, lastly, was chlorinated repeatedly until the residue was constant 
in weight. Two experiments were thus carried out with the following results : — 
Experiment I. 
Experiment II. 
Weight of fused chloride of silver taken, 
2 ‘95 11 grams. 
2-9487 
grams. 
Weight of bromide of silver added — 
a, By loss of weight of the preparation tube, 
0-1465 „ 
b, By excess of mixed haloid after fusion over 
chloride, . . 
0-1466 „ 
0-1497 
Bromine in O’ 1 466 grams, by calculation from b, 
62-38 mgrms. 
63-70 
mgrms. 
Loss on chlorination, ..... 
350 „ 
35-7 
Hence bromine found, ..... 
62-90 
64-15 
3 ) 
Error, . . . . . . = + 
0-52 
+ 0-45 
33 
We see that the absolute error is about the same as in the case of pure bromide ; but 
just on that account the relative error is necessarily far greater, amounting as it does to 
about 1; \ 0 th of the bromine to be determined. From the two analyses made it would 
appear that the method is liable to a constant positive error equal to about 0‘5 mgrms. of 
bromine. But this would be an unsafe conclusion to draw from so small a number of 
experiments. It would be well to regard their results as simply confirming what the 
trials with pure bromide of silver have brought out. 
After the above experiments, all that remained to be done was to see whether the 
method which I had adopted for separating out the bromine from a sea-water as a mixed 
haloid precipitate, does its work exhaustively, and if not, what proportion of the 
bromine escapes precipitation. 
For this special purpose the artificial sea-water had been prepared, and the mode in 
which it was utilised hardly needs description. In each of a series of test analyses one 
kilogram of the artificial sea-water was mixed with a known weight of the standard 
( tppi « -\iniat < ly deci-normal) solution of hydrobromic acid, containing a weight of bromine 
approximately equal to that in a natural sea-water per 19‘45 grams of chlorine, and the 
