MANGANESE. 127 



From BaCiO^ ._. Cr = 53.200, d= .064 



" CfjOg " = 52.482, zb .018 



" Ag^CrO, " ^52.536,^.074 



" Ag^Cr^Oj " = 52.188, ± .109 



" KjCr^O, " =: 52.116, ± .078 



" CrClg " = 51-992, d= .047 



General mean " == 52.453, zb .015 



Or, if O = 16 " = 52.574 



On account of the wide discrepancies between different 

 data, and of the known constant errors vitiating some of the 

 series of experiments, the foregoing general mean can have 

 but little real value. In fact, a careful consideration of all 

 the work represented in it will show that the most accurate 

 estimate of the atomic weight of chromium must be deduced 

 from the experiments of either Berlin, Kessler, or Siewert. 

 Berlin's figures, taken by themselves, and combined, give, 

 if the single analysis of silver anhydrochromate be assigned 

 equal weight with a single analysis in the monochromate 

 series, Cr — 52.389, ± .019; or, if O = 16, Cr = 52.511. 



Siewert's results, both for chromic chloride and the silver 

 anhydrochromate, properly combined, give Cr = 52.009, 

 ± .025. If O ^ 10, this value becomes Cr = 52.129. In 

 brief, the atomic weight of chromium may be nearly 52.5, 

 or it may be 52. Only a revision of all the experiments 

 could enable us to decide positively between these values. 

 But as Siewert has pointed out probable sources of error in 

 Berlin's work, I am inclined to give preference to the lower 

 value. 



MANGANESE. 



Rejecting the early experiments of J. Davy and of Arf- 

 vedson, the first determinations of the atomic weight of 

 manganese which we encounter are those of Turner* and of 

 Berzelius.f Both of these chemists used the same method. 



* Trans. Roy. Soc. Edin., ii, 143. 1831. 

 f Lehrbuch, 5th Ed., 3, 1224. 



