OF ARTS AND SCIENCES. 17 



silver. Now it is easy to estimate volumetrically -j-i^ of this differ- 

 ence with certainty, "We therefore prepared with great care a button 

 of pure * metallic silver, which we annealed, and rolled out to a thin 

 ribbon. We then weighed out from two to four grammes of bromide 

 of antimony, prepared by sublimation as described above, and dis- 

 solved this salt in an aqueous solution of tartaric acid, which we then 

 transferred to a litre flask, and diluted to about 500 cubic centimetres. 

 We next very accurately weighed out a quantity of silver slightly less 

 than that which calculation showed was required for complete precipi- 

 tation. This silver was dissolved in nitric acid, and the solution hav- 

 ing been evaporated to dryness over a water bath, the silver salt was 

 washed into the flask containing the bromide of antimony. As soon 

 as the supernatant liquid had cleared, the small additional amount of a 

 normal silver solution required to produce complete precipitation was 

 run in from a burette, and measured with the usual precautions. We 

 used no extraneous indicator, because it was important not to intro- 

 duce any possibly new disturbing element into the experiment, and in 

 the titration of bromine with silver, the normal and familiar phenom- 

 ena, which mark the close of the process, furnish a very sharp indica- 

 tion. The details of one of the determinations were as follows : ■ — 



The weight of the bromide of antimony used amounted to 2.5032 

 grammes. To precijiitate the bromine from the solution of this material, 

 2.2-iOl: grammes of silver would be required if Sb = 122.00, and 2.2529 



[* A quantity of silver which had been reduced from chloride and bromide of 

 silver, obtained as a product of previous analytical processes, was dissolved in 

 nitric acid, and precipitated as chloride by hydrochloric acid. The precipitate 

 was first boiled in aqua regia, and then thoroughly washed, after which the 

 chloride was reduced by boiling with caustic soda and inverted sugar, and the 

 precipitate, again washed, having been transferred to a porcelain crucible and 

 dried, was heated to a low red heat in a muffle until the grains were sintered to- 

 gether. The sintered mass was melted on a block of prepared coke before a gas 

 blow-pipe, and while cooling was covered with a reducing flame in order to 

 prevent the occlusion of oxygen gas. The metallic button was next rolled out 

 into a ribbon between steel rollers ; and, after the ribbon had been annealed in 

 a muffle, the surface was etched with dilute nitric acid, and afterwards scoured 

 with sand. The metal thus prepared was preserved under distilled water. 

 The oxygen occluded by the metal thus prepared must have been, if any, 

 exceedingly small in amount; but, even allowing the average quantity found 

 by Dumas in metal which had been melted in the air under ordinary circum- 

 stances, we calculated that this amount would only affect the third decimal 

 place in the atomic weight of antimon}' ; and it seemed therefore unnecessary 

 to take so inappreciable an effect into consideration. Moreover, the great purity 

 of our material was subsequently made evident. 

 VOL. xvii. (n. s. XI.) 2 



