NIOBIUM. 89 



a known weight of nickel. The number is the mean of 

 four experiments; extreme difference, 0.186 for = 16, 

 Au = 196. The nickel was prepared as follows : commer- 

 cial nickel carbonate was dissolved in chlorhydric acid, 

 cobalt was repeatedly precipitated with sodium hypochlo- 

 rite, copper, etc., were removed with hydrogen sulphide, 

 the nickel was precipitated with sodium carbonate, the pre- 

 cipitate dissolved in chlorhydric acid, the chloride sublimed 

 and reduced in a current of hydrogen. [Fresenius' Zeitsch., 

 6, 1867, 22.) 



W. J. Russell : 58.76 (O = 16). 



Determined by the amount of hydrogen set free by solu- 

 tion of nickel in chlorhydric acid. The nickel was that 

 obtained in Russell's former determination of the atomic 

 weight of nickel. {Chem. News, 20, 1869, 20.) 



R. H. Lee: 58.01 (0 = 16). 



Determined by analyses of nickel cyanide salts. They 

 were decomposed in a platinum crucible by heat from above. 

 The carbon separated out was burned oft" first in air and 

 then in oxygen. The metallic oxide was reduced in a cur- 

 rent of hydrogen. The mean of six experiments on the 

 strychnine salt gave Ni = 58.04. The mean of six experi- 

 ments on the brucine salt gave M =: 57.98. The salts were 

 purified by recrystallization. {Berlin. Bericht der Chem, 

 Ges., 4, 1871, 790.) 



NIOBIUM. 



The vapor density of the chloride and of the oxychloride, 

 as determined by Deville and Troost, places the atomic 

 weight at about 94. {Paris Comptes Bend., 66, 1863, 891.) 



H. Rose : 122 (O = 16). 



Rose deduced the atomic weight of niobium from analy- 

 ses of what he supposed to be the tetrachloride, determin- 

 ing the niobium as niobic acid, and the chlorine as argentic 

 chloride. The results, which varied greatly, indicated the 

 value 97.64. [Marignac having proved that the salt is a 

 peutachloride, this number becomes 122.] Marignac showed 



