ON THE ELECTROLYTIC METHODS OF QUANTITATIVE ANALYSIS. 301 



obtained indicate that it is not excelled by any of the latter, for the 

 determination of nickel. 



The apparatus employed for most of the determination consisted of the 

 usual cathode basins of about 200 c.c. capacity, roughened on their internal 

 sui'face, with perforated anodes of watch-glass shape. In some experi- 

 ments with small quantities of substance ordinary platinum crucibles were 

 employed in place of the basins ; in these cases a stout platinum wire 

 coiled at the end served as anode. 



The electrical measurements were made by means of Davies' ammeter 

 and voltmeter, described in the third report of the Committee. ' 



When the large basins were employed the solutions were generally 

 made up to a volume of 130-135 c.c. This left plenty of room for further 

 additions, in case these should be found necessary, and the active cathode 

 surface was then approximately 100 sq. cm., so that the ammeter readings 

 corresponded to current density as generally stated. With the crucibles 

 the volume was 18-20 c.c, and the cathode surface was roughly calculated 

 to be 20-25 cm. 



The salts employed for analysis were generally the ammonium double 

 sulphates, specially prepared for the purpose. Sometimes these were 

 directly weighed out for each experiment, but solutions of known strength 

 were also prepared, and measured quantities taken for various determina- 

 tions. For some experiments the pure chlorides were employed. It was 

 considered unnecessary to analyse the material by the other usual methods, 

 as they are not more accurate than the electrolytic process. The results 

 were simply judged relatively to one another, that result being considered 

 best which gave the lowest percentage result, provided, of course, the 

 metal was completely deposited in each case. Great attention was also 

 paid to securing deposits of good general appearance, as a bi'ight, lustrous, 

 and firmly coherent deposit is always much more satisfactory and much 

 less liable to be injuriously afiected. A dark powdery- looking deposit 

 was therefore considered unsatisfactoiy, even although the numerical 

 result came out all right. 



It is usually recommended to test for complete deposition by with- 

 drawing, from time to time, portions of the liquid and adding hydrogen 

 sulphide or potassium thiocarbonate. If this is done, however, the current 

 must be continued for some time after no reaction is observed, otherwise 

 a distinct quantity of metal may be left in solution. It was several times 

 found that the whole volume of liquid after decantation gave a very dis- 

 tinct reaction, although none was visible when a small portion was tested. 

 It was therefore considered more convenient to determine what time was 

 necessary for average experiments conducted under suitable conditions, 

 and to adhere to this generally, unless there were marked indications, 

 for example in the rate of decolourisation, that deposition was not pro- 

 ceeding normally. The decanted liquid was always tested, and any case 

 of incomplete deposition duly noted. 



In the case of nickel the liquid can simply be poured off at the con- 

 clusion of the experiment, and the basin quickly rinsed with distilled 

 water, as the deposit undergoes no apparent deterioration by such treat- 

 ment, provided it is carried out quickly. In the case of cobalt it is not 

 quite safe to work in this way, as the deposit is more liable to tarnish, 

 and it is preferable, if the best results are desired, to employ the usual 



' Transactions, 1895. 



