150 THE ATOMIC WEIGHTS. 



(3.) Ani,^Fe(SOj2-6H20 : FeWO^ : : .0365457. dr .ocx)OOl2 : .028354.9, 



.0000115 

 (4.) AgCl : Ag2W0i : : lOO : 161.645, zb .012 

 (5.) Percentage of CI in WClg, 53.610, ± .015 



From these we get five values for tungsten, as follows : 



From (l) W = 183.703, ±; .041 



" (2) " =183.532,^.156 



" (3) " = 183.923, dz .120 



" (4) " ^183.248,^.069 



" (5) "= 183.639, zfc .109 



General mean " = 183.610, d= .032 



Or, if O = 16, then " = 184.032 



URANIUM. 



It is not the purpose of the present investigation to ex- 

 amine at all systematically such questions as are involved 

 in the discussion whether the atomic weight of uranium is 

 120 or 240. For convenience we may use the formula} based 

 upon the smaller number, and, if eventually the larger value 

 proves to be correct, it will be easy to double the figures 

 which we obtain. Suffice it to say here, that the specific 

 heat of the green oxide, according to Donath,* agrees best 

 with the formula U3O4 and the lower atomic weight. On 

 the other hand, the value 240 fits best into such schemes as 

 that given by Mendelcjeff in his paper on the periodic law. 

 An accurate determination of the specific heat of the metal 

 itself is much needed, for the material with which Rcgnault 

 worked was of uncertain quality ; furthermore, the vapor 

 density of some volatile uranium compounds ought to be 

 ascertained. t Until some such data have been rigidly 



*Ber. d. Deutsch. Chem. Gesell., 12, 742. 1879. 



t The value of 240 for uranium is strongly sustained by tlie recent experiments 

 of Zimmermann upon the vapor density of the tetrachlorid and tetrabromid. For 

 UBr^ the vapor density is 19.46, while theory (U = 240) requires 19.36. For 

 UCl^ the v. d. 13.33 was found. Theory, 13.21. (Bar. der Deutsch. Chem. 

 Gesell., 14, s. 1934. 1881.) 



