ON THE SILVER VOLTAMETER. 553 



was opened. The precision with which this was secured was tested by placing a 

 second chronograph and battery in place of the silver voltameters and noting the 

 difference in the intervals of time recorded by the two instruments. The mean of 

 20 readings indicated a difference of a little less than one-hundredth of a second, 

 which is equivalent to an error of 1 part in 600,000 in the observations made with 

 the voltameters. The time was measured by the standard clock presented to the 

 National Physical Laboratory by Lady GALTON. The rate of the clock was deter- 

 mined by means of signals from Kew and Greenwich. The battery was earthed at 

 such a point, E, that the mean difference of potential between the voltameters and 

 the earth was very nearly zero, but the insulation of all the apparatus from earth was 

 also very carefully attended to. The switches C and CX were on one board and 

 could not be separately operated ; C reversed the current through the standard 

 resistance, and C 7 reversed the connections of the standard cell to the potential points 

 of the resistance. C" is a switch for placing either of two cells S, S' in the potentio- 

 meter circuit ; S' was employed for the adjustment of the current before including the 

 voltameters in the circuit, and S continually afterwards. No secondary potentiometer 

 circuit was used. The resistance coils of RI were of manganin and were immersed in 

 a large bath of paraffin oil. The double mercury trough M was bridged by a copper 

 sliding piece which shunted a portion of the resistance of the trough and allowed ot 

 a fine adjustment. A change in current of 1 part in 1,000,000 was easily detected, 

 and sometimes a current constant to this amount could be maintained for an hour or 

 more. G was a galvanometer of the Broca type of 1000 ohms resistance. 



The circuit was at first closed so as to exclude the voltameters, and remained 

 closed for 1 hour or more before any adjustment for constancy of current was made. 

 On many occasions a determination of current in absolute measure preceded the 

 deposition of silver. This usually occupied 20 minutes ; immediately afterwards R, 

 was diminished by an amount comparable with that of the voltameters and the latter 

 switched into the circuit. With the Rayleigh form of voltameter a current steady to 

 1 part in 100,000 was secured within 20 seconds after closing the circuit ; a slightly 

 longer time was necessary for the Richards type and longer still for the largest of 

 the syphons. 



Owing to the difficulty of maintaining a steady current through the syphon and other 

 modified forms of voltameters, some of the observations are relative only. In these 

 latter cases the standard is the Rayleigh form, but the constancy of this had been 

 well established before any relative observations were made. In order to distinguish 

 between the relative and the absolute values, we have placed an asterisk against all 

 absolute determinations. 



When Lord RAYLEIGH* determined the electrochemical equivalent of silver, the 

 current that passed through the voltameters also passed through the standard current 

 lalance, and was thus directly determined in absolute measure. We also might have 

 * RAYLEIOU and SIDGWICK, 'Phil. Trans.,' 175, p. 411, 1884. 



VOL. ccvn. A. 4 B 



