Chemical Affinity in terms of Electromotive Force. 339 



deflection ; after the experiment was concluded the scale was 

 again similarly standardized. The difference between the two 

 standardizings was always very small or nil, but frequently was 

 just perceptible, owing to leakage of electrometer charge ; 

 the mean of the two standardizings was taken to represent a 

 potential diference of 



1-457 {l-(*-15)x 0-0006} volt, 



where t is the Centigrade temperature, and 0*0006 the altera- 

 tion in E.M.F. per 1° C. (equal to 0*06 per cent, per 1°). As 

 in the whole series of experiments taken together the average 

 temperature of the cell was very close to 15°, the cell was 

 assumed to have the constant average value 1*457 volt 

 throughout. Throughout the experiment the potential diffe- 

 rence between the electrodes was determined at as nearly as 

 possible equal intervals of time by reversal, the quadrants 

 being connected with the mercury-cups of the voltameter 

 through the reversing-gear. As a general rule, the indicated 

 potential-difference rose steadily throughout the experiment 

 to the extent of from 3 to 6 per cent, of its initial value 

 (owing to polarization of the electrodes); but in some cases 

 this effect was just about compensated by a gradual diminu- 

 tion in current-strength, so that the potential difference re- 

 mained nearly the same or slightly fell*. In the experiment 

 the calorimetric details of which have just been given, the 

 following values were obtained : — Scale- 



divisions. 

 Value of Clark cell before experiment . . 80*65 

 „ „ after „ . . 80*55 



Mean 80*60 



Average of 23 pairs of readings of potential difference be- 



* The fluctuations in the current and in the potential difference set up 

 between the electrodes necessarily cause corresponding fluctuations in the 

 amount of decomposition per second and the quantity of heat developed 

 per second^ and hence are allowed for, and (if not too great in extent) 

 produce no effect at all on the end result. In just the same way, it is 

 immaterial whether the electrodes are perfectly parallel and equal in size, 

 whether the current passes more between their edges than between their 

 centres, of what material they are made, and so on ; for what is actually 

 done is to measure the difference of potential between the mercury-cups 

 at the ends of the thick copper rods connected with the electrodes, and also 

 the quantity of electricity passing, and hence the total work done between 

 the cups. The resistance of the cups and thick copper rods being inappreci- 

 able, all this work must be done inside the voltameter between the elec- 

 trodes, either as heat or chemical decomposition • the former being mea- 

 sured directly, the amount of the work representing the latter is known 

 by difference. 



