170 



These results indicate that thi> method is ool sitive and fails 



when much cobalt is present. 



C. The Liebig-Gauhe Method. 



10 cc. portions of CoSO*, 0.0921 molar, with varying amount- of NiSO< 

 were treated with a slight excess of KCX over that required to dissolve the 

 precipitate, and heated and rotated until complete oxidation of the eobalto- 

 eyanide had taken place. They were then diluted to ."() cc. and freshly pre- 

 pared sodium hypobromite added. After the precipitate had floccula 

 it was filtered off, washed, dissolved in dilute HO. r.eutralizt d with ammonia 

 and tested for Xi with dimethylglyoxime. Results below. 



-\iSO4 

 0.C005 molar 



Mg. Xi 



TABLE VIII. 



Ratio Ni : Co Ratio Ni 



HiO 



Results. 



9 CC 1 0.2641 



6 cc j .1760 



4 cc 1172 



3 cc .0880 



2 cc j . 0587 



1 cc I . 0293 



None ' None 



1 



206 



1 



309 



1 



462 



1 



617 



1 



925 



1 



IS. 50 



1 : 568 000 

 1 : 852,000 



Blk. ppt. Ni confirmed 

 Blk. ppt. Ni confirmed 



Blk. ppt. Ni confirmed 

 Blk. ppt. Xi confinm-cl 

 Blk. ppt. No Ni 

 Blk. ppt. No Ni 

 Blk. ppt. No Ni 



This method is shown to be capable of detecting 0.1 mg. nickel in a volume 

 of 50 cc.. l:ut a confirmatory test must in every case lie applied as the ppt. 

 contains Cot OH) 3 . 



Comparing the results of the three methods, the minimum amount of 

 nickel detectable within one hour in a volume of 50 cc. is found to be: 



Silver 0.02 mg. 



Tschugaev-Treadwell 23 mg. 



Liebig-Gauhe 09 mg. 



These figures do not adequately convey the relative merits of the time 

 methods, for it should be noted in addition that the Liebig method requires 

 a confirmatory test to make the result trustworthy: the Treadwell method 

 failed to show the stated minimum amount of nickel when so little as 23J 

 times as much cobalt as nickel was present, Avhile the silver method ap] 



