ON THE ELECTROLYTIC METHODS OF QUANTITxVnVE ANALYSIS. 307 



cases {tartrates or citrates, for example) the i-esultiiig deposits are of very- 

 good external appearance, but frequentlj^ the ninnerieal results are high 

 and the deposits leave a distinct brownish residue when dissolved in dilute 

 acid ; further, the presence of considerable quantities of organic salts may- 

 retard deposition. This apparently does not apply to oxalates. While, on 

 the whole, the use of organic sales, in dealing with ordinary straight- 

 forward depositions, appears to be attended with no real benefit, but in 

 some cases the reverse, it is quite possible that they may be employed 

 with advantage in certain separations. 



The influence of salts of the alkali metals when present in the solution 

 has already been discussed. It is sometimes stated that the presence of 

 magnesium salts is objectionable, but this d -es not seem to be the case to 

 any considerable extent, when only modei-ate quantities are present. The 

 deposits obtained in the experiments tried were somewhat rougher than 

 otherwise, and the results a trifle high. 



The metals of the alkaline earths are of course excluded in presence of 

 sulphate. Neither is it possible to determine nickel by deposition from 

 solutions containing these metals by replacing the ammonium sulphate by 

 the chloride. In that case alkaline liquid attracts carbonic anhydride 

 from tJie air in sufficient quantity to give a distinct deposit of carbonate, 

 so that the results obtained are much too high. 



The examination of the behaviour of other metallic salts brings us to 

 the question of the electrolytic separation of nickel from other metals, 

 which will not be discussed fully here. It may be stated generally, how- 

 ever, that those metals which yield ammoniacal solutions are deposited 

 electrolytically along with nickel. The most important of these are copper 

 and zinc. The former can be removed by electrolysis in acid solution. 

 The presence of even a relatively small quantity of zinc greatly retards 

 the deposition of nickel, but ultimately both metals are completely removed 

 from solution. 



Determination of Cobalt. 



The electrolytic estimation of cobalt has apparently not been the 

 subject of so much investigation as that of nickel ; it is generally stated 

 to be exactly similar in method to that of nickel, no further instructions 

 being given. There is, however, a very considerable difference in the two 

 cases. Good nickel deposits are obtainable with the greatest ease, but 

 the reverse is the case with cobalt, and the best conditions for depositing 

 the one are by no means the best for the other. 



If an experiment with cobalt is conducted under the standard con- 

 ditions given for nickel, the metal is not completely precipitated, and 

 a very poor deposit is obtained. Apparently ammonia has a much greater 

 effect in this case than with nickel, probably due to the formation of 

 stable cobalti-ammonium compounds. On the other hand, the tendency 

 to form cobaltic hydroxide on the anode is not nearly so great, and the 

 quantity of ammonia can be reduced as low as 1-5 grme. Even then, 

 four hours is barely sufficient for complete deposition with moderate 

 quantities of metal, using cold solutions. 



The best determinations obtained even with small quantities Avere 

 very much inferior to those of nickel as regards the jDhysical character 

 of the deposit, and an extended series of experiments was carried out 

 in the hope of securing the metal in better condition. These experi- 

 ments varied considerably as regards the composition and proportions of 



X L' 



