OF ARTS AND SCIENCES. 15 



few grammes at a time in the apparatus we used, and as the several 

 products were not mixed, each of the portions analyzed was the yield 

 of a separate sublimation, and the agreement of the results is therefore 

 in itself a proof of the constancy of the preparation. In the first set of 

 analyses, the bromine was determined gravimetrically by precipitation 

 with nitrate of silver, as before described. The bromide of antimony 

 was first dissolved in a concentrated solution of tartaric acid, using 

 about five grammes of the crystallized acid to a gramme of the bromide, 

 and the solution was then diluted with from 250 to 500 cubic centi- 

 metres of water. To this solution was slowly added, under constant 

 agitation, a solution of nitrate of silver in slight excess of the amount 

 required for complete precipitation. The amount of crystallized 

 nitrate of silver required was always carefully weighed out, allowing 

 one decigramme in excess of theory for every litre of the solution of 

 bromiile of antimony made as above described. The nitrate of silver 

 was then dissolved in a considerable volume of water, and the bromide 

 of silver was precipitated from a cold solution, care being taken to 

 prevent the formation of lumps which invariably result if the solu- 

 tions are warm or concentrated, and which greatly interfere with the 

 washing of the precipitate. The precipitate was washed by decanta- 

 tion five times, pouring on in each case a volume of lukewarm dis- 

 tilled water equal to that of the original solution, and after the pre- 

 cipitate had settled, drawing off the wash-water with the inverse filter 

 (see these Proceedings, Vol. XII. page 124.) Lastly, the precipitate 

 was transferred to a porcelain crucible, and dried at from 120° to 

 130°. 



The bromide of silver weighed was always tested, sometimes by 

 dissolving the material in strong aqua ammonia, and at other times by 

 heating it to the melting point. Had there been the least occlusion 

 of silver emetic, or any other possible product, there would have been 

 an insoluble residue or a loss of weight ; and, since the bromide of 

 silver always gave a perfectly clear solution, and the loss of weight on 

 melting never exceeded a few tenths of a milligramme, we were as- 

 sured that our product was perfectly pure. 



Of the five determinations whose results are given below, the first 

 three were made under my direction by Mr. G. De N. Hough ; the 

 last two were made with my own hands. 



