96 THE ATOMIC WEIGHTS. 



Hence, using the atomic weights and probable errors 

 previously found for N and 0, Tl = 203.715, ± .0365. If 

 O = 16, Tl = 204.183. 



Crookes himself, using 61.889 as the molecular weight of 

 the group NO3, gets the value Tl = 203.642; the lowest 

 value in the series being 203.628, and the highest 203.666 ; 

 an extreme variation of 0.038. This is extraordinary ac- 

 curacy for so high an atomic weight, at least as far as 

 Crookes' work is concerned. But its value depends in reality 

 upon the accuracy of other chemists in fixing the atomic 

 weights of N and O; a slight variation in either of the 

 latter constants producing a large variation here. What 

 Crookes really has done has been to fix with almost abso- 

 lute certainty the ratio between Tl and NO3. If the latter 

 group should have the molecular weight 62, in accordance 

 with Front's hypothesis, then Tl = 204.008. In other 

 words, the ratio thus fixed by Crookes is almost exactly 

 represented by two whole numbers, and supports Front's 

 hypothesis in a very decided way. Crookes himself seems 

 to have overlooked this fact, for he regards his results as 

 militating against the hypothesis in question. 



GLUCINUM. 



The atomic weight of glucinum is at present much in 

 doubt; our knowledge of it depending upon the unsettled 

 question whether the oxide is.GlO or GLOg. The formula 

 GIO agrees with Mendelcjeff's law, and is advocated by 

 Reynolds,* Lothar Meyer,t and Brauner.J The symbol 

 GUO;,,on the other hand, is favored by Nilson and Fetters- 

 son, H and by Iiumpidge.§ Ilumpidge, Meyer, and Brauner 



"Phil. Mag., (5,) 3, 38. 1877. Chem. News, 42, 273. 1S80. 



t Ber. der Deutsch. Chem. Gesell., 13, 1780. 1880. Also, 11, 576. 1S79. 



J Phil. Mag., (5,) II. Jan., 1881. 



II Berichtc, 11, 381 and 906. 1879. Also, 13, 2035. iSSo. 



^ Chem. News, 42, 261 . iSSo. 



