in a Recalculation of the Atomic Weights. 105 



Ccesium. — Johnson and Allen, Bunsen, and Godeffroy. 

 Cs = 132*583; ±-024. 



Thallium. — Although I have recalculated the results ob- 

 tained by Lamy, Hebberling, and Werther, I need consider 

 here only the experiments of Crookes. His weighings, calcu- 

 lated with the values for and N given above, make 

 Tl = 203*715; ±*037. Crookes himself, using the value 

 N0 8 =61'889, found Tl=203*642 and regards his results 

 as evidence against Prout's hypothesis. His experiments, 

 however, really fix only the ratio between N0 3 and Tl. If 

 M) 3 = 62, then Tl = 204*008 ; that is, the ratio which Crookes 

 has rigorously established is a ratio between two whole num- 

 bers, and is confirmatory of Prout's idea. 



Glucinum. — The results of Awdejew, Klatzo, Weeren, and 

 Debray have high probable errors, and practically vanish from 

 the mean when combined with those of Nilson and Pettersson, 

 The weighings published by the latter chemists give me 

 Gl = 9-085 ± -0055, or Gl = 13-628 ± -008, according to whether 

 the oxide is GIO or G1 2 3 . 



Magnesium. — Scheerer, Svanberg, and Nordenfeldt, Bahr, 

 Jacquelain, Marchand and Scheerer, and Dumas. General 

 mean of all, Mg=24'103 ; ±'004. Some of the series are 

 vitiated by constant errors. The best value is to be deduced 

 from Marchand and Scheerer's work alone, and is Mg= 23*959 ; 

 ±•005. 



Zinc. — Jacquelain, Favre, and Axel Erdmann. General 

 mean, Zn = 65*557 ; +'011. Erdmann's results alone, how- 

 ever, give a more probable value (considered chemically) of 

 Zn = 64*9045; ±-019. 



Cadmium. — Von Hauer, Lenssen, and Dumas. General 

 mean, Cd = 111*770; +'030. 



Mercury. — Part of Turner's data are to be considered, to- 

 gether with results by Erdmann and Marchand, Millon, and 

 Svanberg. General mean, Hg = 199*712 ; + '042. New 

 determinations are much needed. 



Chromium. — Peligot's work was not available for discus- 

 sion. The data studied were by Berlin, Moberg, Lefort, 

 Wildenstein, Kessler, and Siewert. General mean from all, 

 Cr = 52*453; + '015. Berlin's work, considered by itself, 

 gives Cr = 52-389; +'019. Siewert's results give Cr = 52-009; 

 + •025. I regard the last value as freest from constant 

 errors, and use it in subsequent calculations. 



Manganese. — Turner, Berzelius, Dumas, v. Hauer, Schnei- 

 der, Rawack. General mean, Mn= 54*128 ; + *011. Schnei- 

 der's and Rawack's results give a better value of Mn =53*906 ; 

 + •012. 



