the voyage of h.m.s. challenger. 
1*2 
• : nts are always weighed along with the tube. No cork-joints are used 
-•j.. -rations ; the connection of the tube with the air-gasometer or the 
, is . fleeted by means of a closely (though not hermetically) fitting glass 
• . j.jm-.I int<* the end of the operation-tube, which works perfectly well, as the gas 
: ..me no pressure. The combustion-tube is never heated directly in the gas 
; ar ted from them by an interposed magnesia bath, as recommended by 
> i -mail i <artieles of haloid of silver which unavoidably pass into the decantate 
:■ i mu a filter, ignited within the filter, chlorinated by evaporation with nitric 
■ hloric a' ids, weighed as chloride, and the weight is allowed for in the calcula- 
: : t«-t d bromine. This weight was always so small, that even supposing it had 
• ■ i:- full . inplcincnt of bromide, instead of being mere chloride as assumed in 
elation, the error would have been inappreciable. 
I : >. pa- to the test experiments which were made for determining the valuer 
t •, i ! . 1 -t ttling certain other questions. Pure bromide of potassium solution was 
. baking bromine with solution of the commercial pure salt, and then distilling 
The distilled bromine was converted into zinc salt, the latter decomposed by 
• . iuivalcnt of pure carbonate of potassium, and the carbonate of zinc filtered 
Th- lut i* >n aft r having been conveniently diluted, was analysed by the gravi- 
ne tri : >rm of Volhard s method, with the following result : — 
Found in one gram of solution 
I. II. Mean. 
1 3*099 1 3‘08 1 13’09 nigrms. of bromiue. 
• ■ : n ■ rwd for the synthesis of solutions containing known weights of 
© © 
1 : mid. Part of it was utilised for preparing pure bromide of silver, 
pur- -ilver salt (weighed after fusion) when chlorinated in a boat within 
t ■ " (■" explained above) gave '1499 grin, of chloride — loss of weight 
. 1-lG mgrms. of bromine. Calculated from the weight of the bromide 
' tie- error +0’G7 mgrm. In order now to see whether chloride of 
*d r ii t'-d r< t its weight, 100 c.c. of dccinormal silver solution were 
■ : ilorie acid, the precipitated chloride, in one case («), filtered off, 
, J v if ' add e.invcniently be detached from the filter-paper, chlorinated. 
i ), the precipitate was collected by decantation. Loss of weight in 
rhlurination ( f »f als»ut 1120 mgrms.) — 
(«) (b) 
0 3 mgrm. 0 
' ..rr< jj-.n ling to 0-54 ,, 0 of Bromine. 
1 i, ‘- • • p riments a known weight of bromine (in the shape of our 
p.Y/) w a- mixed with enough of pure chloride of sodium to produce 
