Manchester Memoirs, Vol. xlvii. (1903), No. VX. 7 



times approximating pretty closely to Ni.On, but, 

 especially after boiling for some time, it occasionally falls 

 belozv the sesquioxide. 



W. A Method for the Volumetric Determination of Cobalt. 



Although it is evident, from the experiments described 

 in the first part of this paper, that the black oxide of 

 cobalt which is precipitated by the carbonates of barium 

 and calcium in presence of bromine is not the sesqui- 

 oxide, yet it is plainly fairly constant in composition. It 

 is therefore possible to use it as a means for the volumetric 

 determination of coJDalt by the method of Bunsen already 

 referred to. I now proceed to describe how, so far as my 

 own experiments suggest, the process is best carried out. 



The solution to be operated upon must be moderately 

 dilute, and it must be as nearly neutral as possible. It 

 is best that no other metals than cobalt and nickel should 

 be present in the solution. Nickel does not interfere, but 

 iron, lead, and manganese would, under the conditions of 

 the experiment, precipitate higher oxides capable of 

 liberating iodine ; aluminium and chromium are both 

 precipitated as hydroxides by the carbonate employed, 

 and, as this reaction would liberate carbonic acid, the 

 precipitation of the cobalt would be prevented. It is 

 remarkable that zinc also interferes materially with the 

 reaction. A minute quantity of that metal sensibly 

 retards the precipitation of the cobalt, and a considerable 

 amount nearly stops it altogether. 



A moderate amount of precipitated calcium carbonate 

 (previously made into a thin paste with water) is added to 

 the solution, and excess of bromine water. The total 



