350 ON TAVO NEW CRYSTALLINE COMPOUNDS 



"NATiole amount of mixed sulphides, 0.8423 



Weight of reduction tube, .... 



+ (SbSs + SbOs + S), 

 « « " " + Sb (after reduction), 



" (SbSa + Sb03 + S), . 

 « " Sb, 



0.1922 : 0.1127 = 0.8423 : 0.4939 . • . 0.2024 : 0.1187 = 0.8423 : 0.49397. 



Several similar examples might be given. The filtrate from the mixed sulphide, con- 

 taining nothing but zinc in solution, was collected with the Avashings in a large beaker 

 glass, which was covered with a shallow glass tunnel having the orifice so fine as to 

 allow fluid only to drop. The beaker was heated on a sand-bath, and when the liquid 

 was in ebullition a concentrated solution of carbonate of soda was poured into the tun- 

 nel, from which it dropped gradually into the boiling liquid. The zinc was in this 

 way invariably perfectly precipitated, and the carbonate obtained in a granular condi- 

 tion, which rendered the subsequent filtering and washing exceedingly easy. This 

 very simple mode of precipitating zinc must have occurred to others, but as I have 

 never seen it described, I will add, that it is very much to be preferred to the ordinary 

 mode of precipitating in a flask. It renders the determination of zinc one of the most 

 accurate of Analytical Chemistry ; in proof of which I Avould refer to the analyses of 

 cr}-stals from the alloys of 33, 32.5, and 29.5 per cent of zinc, given in the table of 

 analyses on the next page. The only source of error in this process arises from neg- 

 lecting to wash thoroughly the precipitated carbonate of zinc. 



In all the analyses published in this memoir, the zinc was determined by the method 

 just described ; but, unfortunately, the accuracy of the method indicated above for de- 

 termining antimony was not ascertained until the investigation was half finished. In 

 the earlier analyses the antimony was detennined from a separate portion of crystals by 

 the method commonly used, namely, dissolving in aqua regia, adding tartaric acid be- 

 fore diluting, in order to keep the oxide of antimony in solution, and determining the 

 sulphur in the precipitated sulphide as sulphate of baryta. This process in the hands 

 of the author proved much less generally accurate than the one which has been recom- 

 mended ; for the slight loss of antimony to which the analyst is liable while reducing 

 with hydrogen was found to be more than counterbalanced by the greater length of the 

 last process, and the danger of obtaining a small amount of oxide mixed with the sul- 

 phide. The analyses were all made in my laboratory, under my direction and immediate 



