10 ATOMIC WEIGHT DETERMINATIONS. 



analyses of the ter-chloride, and the penta-chloride {Poggend. 

 Ann., 3, 1825, 443) made by himself by the same method, 

 which give respectively 1512.91 and 1508.6. {Poggend. Ann.^ 

 98, 1856, 455.) 



W. P. Dexter: 122.336 (0 = 16); 764.6 (0 = 100). 



Attempts were made to determine the atomic weight of 

 antimony from its reducing action on the chloride of gold, 

 but no constant result was obtained. Berzelius' method (vide 

 supra) was, therefore, adopted. From the mean of ten irre- 

 proachable experiments Dexter deduces the value 1529.2; 

 extreme difference, 3. The metal was prepared as follows: 

 From antimony tartrate, sodium metantimonate was pre- 

 pared, and antimonic acid separated out with nitric acid. 

 The antimonic acid was reduced with carbon, and melted 

 with another portion of antimonic acid to remove traces of 

 sodium, etc. It was also heated in a current of hydrogen 

 to remove traces of oxide. The investigation was carried 

 out in Bunsen's laboratory, and with his assistance. {Pog- 

 gend. Ann., 100, 1857, 563.) 



J.Dumas: 122 (0 = 16). 



Neither the reduction of cervantite nor of the sulphide, 

 nor the oxidation of metallic antimony gave accordant re- 

 sults. Dumas, therefore, resorted to the analysis of the 

 ter-chloride with argentic nitrate. The chloride was pre- 

 pared by three different methods, and was dissolved in water 

 acidulated with tartaric acid. Seven experiments gave an 

 average of 121.975; extreme difference, 0.69. Ag = 108 ; 

 CI = 35.5. {Ann. de Chim. et de Phys., (3,) 1859, 175.) 



F. Kessler : 122.24 (0 = 16). 



In four experiments crystals of antimony ter-oxide were 

 employed. This oxide had been sublimed in a current of 

 pure, dry carbonic acid. A known weight of the compound 

 was nearly oxidized in a chlorhydric acid solution by a 

 known, slightly insufficient, weight of potassic chlorate. 

 The remainder was titrated with a standard solution of 

 potassic bi-chromate, and countertitrated with ferrous chlo- 

 ride. The mean result was Sb = 122.16. In three experi- 

 ments metallic antimony was employed. It was prepared 

 by reducing the precipitate formed when ammonic hydrate 

 is added to stibium-ammonium tartrate. The metal was 

 oxidized in chlorhydric acid solution by potassic chlorate, 

 (not weighed,) and reduced to antimony ter-chloride by 



