22 ATOMIC WEIGHT DETERMINATIONS. 



this gives bromine = 79.924 with an extreme difference of 

 0.046.] In vacuo this becomes, according to Stas, 79.945. 

 In four experiments potassium bromate was decomposed by 

 heat, and the potassic bromide weighed. [For K = 39.137 

 these experiments give bromine at 80.11 with an extreme 

 difference of 0.56. These latter are evidently much less 

 accurate than the preceding, and I have therefore averaged 

 the first and second series in. vacuo.'] The KBr was prepared 

 by heating bromate purified by recrystallization. [Berzelius' 

 Lehrbuch der Chemie, 5th ed.,3, 1194; BibL Univ., 4-^, 1843, 

 357.) 



W. Wallace : 79.74 (O = 16). 



Determined by analysis of arsenic ter-bromide, by titra- 

 tion with argentic nitrate, according to the method of 

 Pelouze, (see arsenic, Pelouze's determination.) Three ex- 

 periments were made, giving a mean of 79.738; extreme 

 difference, 0.051. As = 75; Ag = 107.97. The arsenic 

 and bromine were directly combined, and the compound 

 was purified by fractional distihation and recrystallization. 

 {Phil. Mag., (4,) 18, 1859, 279.) 



J. Dumas : 80 (O = 16). 



Determined by three experiments on the conversion of 

 argentum bromide into chloride in a current of dry chlorine. 

 The mean is 80.03; the extreme difference is 0.18. Silver 

 is taken at 108, and chlorine at 35.5. The argentum bro- 

 mide was prepared with bromine free from iodine, and was 

 purified from chlorine by digestion with argentum bromide. 

 [Annal. de Chemie et de Physique, (3,) 55, 1859, 162.) 



J. S. Stas : 79.952 (O = 16). 



Four complete syntheses (the weight of each of the con- 

 stituents, and of the compound being determined) were 

 made of argentum bromide, a known weight of silver being 

 converted into sulphate, and precipitated with a known 

 weight of bromine which had been converted into hydro- 

 bromic acid. The mean result was that 100 Ag =: 74.0805 

 Br; with an extreme ditierence of 0.004. Two analyses of 

 argentic bromate, made by reducing the salt in suspension 

 with sulphurous acid, gave for the molecular weight of the 

 bromide 187.84, and 187.90, mean 187.87. A comparison 

 of these data gives Br = 79.940. [This, I think, must be a 

 misprint for 79.949.] Fourteen experiments were made on 

 the equivalence of KBr and Ag by Pelouze's method, (see 



