590 MR. F. E. SMITH, MR. T. MATHER, AND DR. T. M. LOWRY 



The impurity present in the mother-liquor evidently consists, then, to a considerable 

 extent of substances which are insoluble in water, but soluble in silver nitrate solu- 

 tions of moderate concentration. These impurities are partially precipitated on 

 diluting the solutions, but a certain amount remains in solution. The property of 

 dissolving in silver nitrate appears to be possessed by a large range of silver salts 

 which are insoluble in water, and many of these possess the property of increasing the 

 weight of the deposit obtained on electrolysing the solution. The removal of silver at 

 the kathode leads, as is well known, to the formation of a film of dilute solution which 

 (owing to its small density) flows upwards towards the surface of the solution. The 

 film immediately in contact with the kathode may be regarded as almost pure water, 

 and the dilution from 15 per cent, to nearly per cent, (by removal of silver nitrate 

 instead of by addition of water) probably causes a precipitation of impurity similar to 

 that which results from the dilution from 50 per cent, to 15 per cent. In this 

 direction we believe an explanation may be sought of the abnormally high equivalents 

 obtained from impure silver nitrate solutions. 



In the course of the later work several centigrammes of the precipitate formed on 

 diluting the silver nitrate mother-liquors were collected and fractionated as 



follows : 



gramme. 



Soluble in dilute nitric acid (Ag 2 O, &c.) . . 0'0231 



ammonia (AgCl, &c.) 0'0664 



Undissolved (Ag 2 S, &c.) 0-0274 



0-1169 



From the first fraction there was recovered 0'0113 gramme AgCl, equivalent to 

 0-0091 Ag 2 ; from the second fraction 0*0567 gramme AgCl. It appears, therefore, 

 that about half the residue consisted of silver chloride, whilst the remaining part 

 contained sufficient oxide and sulphide to give the precipitate a black colour, and so 

 disguise to a large extent the chloride which was its chief constituent. 



G. Stnation of the Deposit. 



It was noticed very early in the investigation that high values for the electro- 

 chemical equivalent were almost invariably accompanied by a characteristic striation 

 of the deposit, whilst normal equivalents were almost always obtained from unstriated 

 deposits with currents of 1 ampere ; it was in fact possible to guess roughly what the 

 weight of the deposit would be by noting the appearance of the silver deposited in 

 the bowl. The impurity which causes the high values is evidently characterised by 

 the property of producing marked striations, and this property was for some time the 

 only qualitative test for the presence or absence of the substance in the silver 

 solutions. It was also considered to be of importance in seeking to determine the 



