124 . ATOMIC WEIGHT DETERMINATIONS. 



cent ; mean 74.46 and Ti at 303.686. According to Gmelin- 

 Kraut, these analyses recalculated for Stas's values give Ti 

 = 48.28. The chloride was prepared by the action of chlo- 

 rine on a mixture of titanic acid and carbon, and was recti- 

 fied four or five times over potassium and mercury. It was 

 clear and developed no chlorine on decomposition with 

 water. {Poggend. Ann., 15, 1829, 145.) 



C. G. Mosander: 4,7.33 (0 = 16); 295.81 (0 = 100). 



Mosander determined the oxygen contents of titanic acid 

 at from 39.83 to 40.82 per cent. ; mean 40.427. Mosander 

 never described the method of analysis. [The oxygen con- 

 tents was probably determined from the chloride, for the 

 above data give Ti =: 294.7, while Berzelius records the de- 

 termination as having given 295.81.] {Poggend. Ann., 19, 

 1830, 212, and Berzelms' Lehrbuch, 3, 1211.) 



J. Pierre : 50.36 (O = 16). 



Determined by three experiments on the titration of the 

 chloride with argentic nitrate by Pelouze's method. Pierre 

 does not give the values taken for CI and Ag. He calcu- 

 lates the atomic weight of Ti at 314.69. [If Ag = 107.93, 

 01 = 35.457; his data give Ti = 314.75 (0 = 100); 50.36 

 (O =: 16), with an extreme diflerence in the latter case of 

 0.08.] He made two other determinations giving lower 

 results, but it was found that the chloride employed was 

 slightly decomposed by contact with air. The chloride 

 was prepared from artificial titanic acid which was free 

 from iron, and was further purified by fractional distilla- 

 tion. {Annal. de Chim. et de Phys., [^,)20, 1847, 257.) 



A. Demoly : 56.51^ (0 = 16). 



Determined by experiments on the tetrachloride. The 

 salt was decomposed with water, the titanic acid precipi- 

 tated by ammonic hydrate, and the chlorine precipitated in 

 the filtrate, after the excess of ammonic hydrate had been 

 volatilized and the solution acidified. Both precipitates 

 were weighed. Demoly calculates the atomic weight of Ti 

 at 350, without mentioning what values he accepted for 

 silver and chlorine. [If Ag = 107.93, 01 = 35.457 ; the 

 atomic weight, calculated frortl the argentic chloride, is 

 353.2 (0 = 100) ; or 56.512 (O = 16), with an extreme differ- 

 ence in the three experiments of 0.88 for O = 16.] The 

 chloride was prepared from rutile by preliminary conver- 



