152 Proceedings of Indiana Academy of Science 



volume of the ether-alcohol mixture. As little as 0.001 m\r. of cobalt in 

 8 cc. can be detected under these conditions. If only traces of iron 

 are present, it causes less interference when NaSCN is used than when 

 the ammonium salt prescribed by Treadwell is used. The interference 

 of larger quantities of iron is not very satisfactorily overcome by am- 

 monium acetate and tartaric acid. 



2. In acid solution and using a small amount of ethyl instead of 

 amyl alcohol the test is at least eriually delicate. The optimum conditions 

 were found to be concentrations of thiocyanate and of H2S04 or HCl 

 about 1 normal each and addition of ethyl alcohol e:;ual to about 5 per 

 cent by volume of the aqueous layer and no more ether than rer^uired to 

 form a separate layer. In acid solution ferric iron is readily and quickly 

 reduced by shaking with a few crystals of sodium thiosulfate. Even 

 moderately large amounts of iron cause no interference with the test 

 for cobalt. The easy removal of interference by iron, the more general 

 availability of ethyl as compared with amyl alcohol, and that any one 

 of the alkali thiocyanates may be used are features which recommend 

 the test in acid solution as a preferable modification. 



