ON TIII-: sii.vr.i: YOI.TAMKTI i: 



ii it me f the impurity ; thus if high values could be obtained with a variety of added 

 impurities, that one which most readily gave striated deposits was the most likely to 

 be the characteristic impurity of the actual solutions used for electrolysis. 



The deposits from the mother-liquors showed an exceedingly marked striatiou, 

 although the silver was dull in appearance. 



H. Influence of Oxide, Carbonate and Chloride. 



The influence of silver oxide on the silver voltameter has been investigated by 

 PATTERSON and GUTHE,* GUTHE, t RICHARDS^ and KAHLE. PATTERSON and GUTIII 

 used a solution saturated with silver oxide, and GUTHE'S comparison of it with the 

 Rayleigh type showed the two to agree. RICHARDS, however, found his form of 

 voltameter to give a deposit lower by 0*1 per cent, when compared with PATTERSON 

 and GUTHE'S type, and KAHLE found the effect of silver oxide was to increase the 

 deposit by 0'05 per cent. From the point of view to which we have referred it was 

 probable that any silver salt which was insoluble in water would, if dissolved in the 

 nitrate solution, give an abnormally heavy deposit. Two experiments made with 

 solutions containing silver oxide gave confirmation to this view. Pure sodium 

 hydroxide was prepared by the action of water-vapour on metallic sodium and was 

 added to a 50 per cent, solution of silver nitrate ; the filtered solution gave a 

 slight brown precipitate on diluting to 400 cub. centims., and on electrolysis it 

 gave for the electrochemical equivalent the value 



1-11852 (50rf), 



0'021 per cent, higher than the normal figure. A later experiment carried out in the 



same way gave the value 



1-11842 (80c). 



These results are not directly comparable with those of other observers owing to the 

 difference in the size of the voltameters, but may be regarded as substantially in 

 agreement with that of KAHLE. 



Very similar results were obtained on adding sodium carbonate, which raised the 

 electrochemical equivalent by about two parts in 10,000 (Observation 536). 



Addition of potassium chloride gave the values 



1-11840 (50c), 1-11847 (806). 



The chloride is freely soluble in concentrated silver nitrate solutions, especially when 

 hot, and is copiously precipitated on dilution ; its effect on the electrochemical 



* PATTERSON and GUTHE, ' Phys. Rev.,' 7, p. 257, 1898. 

 t GUTHK, 'Phys. Rev.,' 19, p. 145, 1904. 

 J RICHARDS and UKIMROD, 'Proc. Am. Ac.,' 37, p. 426, 1902. 

 KAHLE, ' Brit. Assoc. Report,' Section A, 1892. 



