ON THE SILVEB VOLTAMETER. 573 



The differences from the mean value are within the limits of error, and there is, 

 therefore, no certain change in the mass of silver deposited per coulomb from electro- 

 lytes containing from T5 to 50 parts of silver nitrate in 100 parts of solution. 



\\'<- tliink it necessary, however, to point out that a current of O'l ampere was used 

 to electrolyse the T5 per cent, solution ; when strong currents were used the silver 

 was deposited as long, needle-shaped crystals, and on one occasion it was precipitated 

 in a spongy form. It is insufficient, therefore, to state the range in the concentrations 

 of the electrolyte without also specifying the quantity of the electrolyte, the extent 

 of the kathode surface, and the current to be used. 



Variation of Current Density. 



We believe that one of the objections to the silver voltameter is that the ordinary 

 size of voltameter possibly allows only of currents of the magnitude of 1 ampere 

 to be measured. It was of some importance, therefore, to decide whether or not 

 currents of the order of half an ampere and others of the order of 10 amperes 

 deposited exactly the same mass of silver per coulomb in our form of the Rayleigh 

 voltameter. This might have been tested by evaluating the currents, noting the 

 times, and determining the masses, but a much simpler and more accurate way was a 

 comparison of the masses of silver deposited, similar to the calibration of a box of 

 weights. This latter method was adopted by us. In our first experiment we 

 compared the masses of the bowls A and- B and of C and D. B and C were then 

 placed in parallel and A in series with them, and a current of 1 ampere passed 

 through A for 2 '4 hours. A mass of silver, weighing about 10 grammes, was thus 

 deposited in A. The bowl D was then substituted for A and the same current passed 

 through it for the same time. At the conclusion of the experiment there were about 

 10 grammes of silver in each bowl, but the silver in A and D had been deposited 

 with a current of 1 ampere, and that in B and C with a current of half an ampere. 

 A was again compared with B, and C with D. The difference (A B) (C D) should 

 be the same as before if the change in current had no effect. Similar observations 

 were made for currents in the main circuit of 2, 4, and 8 atnperea The masses of 

 silver deposited and the difference (A B) (C D) before and after the depositions 

 are given in Table IX., p. 574. 



We conclude from these results that the voltameter employed by us can be used 

 for the determination of currents as great as 8 amperes and as small as 0'5 ampere, 

 and that these currents will deposit the same mass of silver per coulomb. With a 

 current of 4 amperes the electrolyte was warm to the hand after the experiment, and 

 with 8 amperes the temperature rose from 18 C. to 35 C. The deposits with high 

 current densities are firm and of a matt surface, while with very low current densities 

 the silver is loose and the deposit striated. 



We have made one observation which connects the deposits obtained with 1 ampere 



