140 PROCEEDINGS OP THE AMERICAN ACADEMY. 



gentic nitrate, hence any possible error must have crept in during the 

 electrolysis. 



The existence of flaws in the working of the silver voltameter is no 

 new idea, and a brief resume of earlier anomalies is necessary in order 

 to indicate our own train of thought. 



Lord Rayleigh and Mrs. Sidgwick found that large cathodes yielded 

 higher results than small ones, and warm solutions yielded higher results 

 than cold ones. In some cases the deviations amounted to 0.1 per cent. 



Schuster and Crossley * state that deposits made in a vacuum are 

 slightly heavier than those produced by the same current in air, and 

 these in turn are heavier than those produced in an atmosphere of 

 oxygen. Myers f verified these statements. Schuster and Crossley 

 showed also that with great current density argentic peroxide may form 

 at the anode, and in some unexplained fashion the result is a diminution 

 of the weight of the silver deposited ; moreover, they pointed out the 

 fact that the discrepancies observed by Lord Rayleigh and Mrs. Sidgwick 

 between large and small bowls disappear when the anodes are of the 

 same size. Apparently "//?e anode gives rise to secondary reactions^ 



Rodger and Watson % observed that on continued use of the electrolyte 

 the deposits grow heavier ; and they also found that, when a very strong 

 slightly acidified solution of argentic nitrate was electrolyzed by a 

 powerful current, the acid was removed, and the deposit was much too 

 heavy. They venture to say that a subsalt of silver is formed " having a 

 silver ion heavier than the arseutic salt." 



Again, Kahle § has found that after boiling the electrolyte with oxide 

 of silver, the deposit is increased 5 parts in 10,000. In a later and very 

 important paper || he calls attention for the first time to the fact that in 

 a dilute electrolyte an acid is formed during the electrolysis. Further- 

 more, he shows again that old solutions give too high results, an error 

 which was sometimes removed by treatment with argentic oxide. 

 Colored spots sometimes appeared upon the silver in old acid solutions, 

 apparently caused by the licjuid descending from the anode ; these did 

 not form in a neutralized solution, or in one which had been allowed to 

 remain in contact with silver. Kahle's hypothetical explanation of these 

 phenomena essentially agrees with Rodger and Watson's. 



* Proc. Roy. Soc, 50, 344 (1892). 



t Wied. Ann., 55, 288 (1895). 



t Phil. Trans., 186 A, 631 (1895). 



§ Brit. Ass. Adv. Sc. Edinb. (1892), p. 148. 



II Wied. Ann., 67, 1 (1899), or Zeitsch. f. Instrkunde, 18, 229, 267 (1898). 



