132 ATOMIC WEIGHT DETERMINATIONS. 



the atomic weight the number 855.84 (0 = 100). R. 

 Schneider has shown that the data as given by BerzeUus are 

 discordant, {Poggcnd. Ann., 88, 319,) a fact of small import- 

 ance in view of the succeeding investigation. The higher 

 oxide analyzed by Berzelius was produced by gently heat- 

 ing the ammonium salt. [Poggend. Ann., 22, 1831, 14; 

 Kongl. Vet. Akad. Handl, 1831.) 



Roscoe examined some ammonium vanadate which Ber- 

 zelius had sent Faraday and found that it contained phos- 

 phorus. {Liebifs Ann., S, 6, 1868, 93.) 



H. E. RoscoE : 51.33 (O = 16). 



Roscoe made four experiments on the reduction of vana- 

 dic acid (V2 O5) in carefully purified hydrogen. The acid 

 was prepared from ammonium vanadate. To free this 

 compound from phosphorus and silicic acid it was powdered, 

 decrepitated with sodium in an iron crucible, washed with 

 water and with chlorhydric acid, re-oxidized with nitric 

 acid, chloridized in a current of chlorine, the chloride rec- 

 tified and decomposed with water. The acid so obtained 

 was dried, moistened with sulphuric acid, exposed to the 

 fumes of fluohydric acid for ten days and melted. This pure 

 acid was first heated for several hours in dry air and after- 

 wards in hydrogen. The mean result of four experiments was 

 V = 51.371 ; extreme difference, 0.228. Nine experiments 

 were made on the titration of the chloride by Pelouze's 

 method. Eight experiments were also made on the analy- 

 sis of the chloride with argentic nitrate by the ordinary 

 method. The mean of the seventeen experiments on the 

 chloride gives the contents in chlorine at 61.276 per cent.; 

 extreme difference, 0.69. This composition indicates an 

 atomic weight of 51.29. Roscoe takes CI = 35.457, Ag = 

 107.93. The vanadium chloride was purified by rectifica- 

 tion over sodium in a current of carbon di-oxide. - The re- 

 agents were prepared according to Stas. {Liebig's Ann., S, 

 6, 1868, 86.) 



Roscoe mentions atomic weight determinations by Czud- 

 nowicz as giving 55.35. This chemist, however, did not 

 calculate an atomic weight from his analyses, but used that 

 obtained by Berzelius. {Poggend. Ann., 120, 1863, 17.) 



