Chemical Affinity in terms of Electromotive Force. 275 



After passage of a current of about 

 300 microwebers for two days, so 

 as to evolve some 5 cubic centims. 



Values gradually ri- 



of gaseous hydrogen, and nearly )■ sing from 1*965 to 



saturate the voltameter-liquid with 2'150 volts. 



oxygen and hydrogen in the two 



limbs respectively. 

 After another week's passage of a 1 Values rising from 



current of about 50 microwebers. j 2*150to2'250 volts. 

 Constant maximum attained after al 9.9^7 i± 



fortnight's passage of the current. J 



86. In precisely the same way analogous numbers were 

 obtained with the same voltameter and different currents, or 

 different strengths of acid, and similarly with the other volta- 

 meters. It was found that when the sources of loss of hydro- 

 gen other than diffusion discharge were wholly eliminated, 

 the value of the counter E.M.F. set up with a given current 

 not only remained constant as long as the current did not 

 vary, but also was the same if the current was varied through 

 a considerable length of time (whilst making other observa- 

 tions), and then brought back to the original amount. With 

 feeble currents of strength not exceeding a small number of 

 microwebers per square centimetre of electrode- surface, how- 

 ever, the value of e was found to be very sensitive to causes 

 which would act on the equilibrium of the auras surrounding the 

 electrodes, especially to changes of temperature which would 

 increase the normal effect of diffusion discharge by setting up 

 heat-convection currents ; so that the constant E.M.F. due to 

 the original current was often not absolutely (but only ap- 

 proximately) arrived at until after the lapse of some hours. 

 For this reason too, the counter E.M.F. observed was often less 

 by 1 or even 2 or more per cent, in such cases when determined 

 in the afternoon (in the warmest part of the day) than when 

 determined the first thing in the morning — partly because the 

 actual rise of temperature diminished its value, but more 

 particularly because of the increased effect of heat-convection 

 in affecting the aurse. In making the following determina- 

 tions, therefore, the mode of procedure adopted was as fol- 

 lows : — The voltameter having been saturated with hydrogen 

 and oxygen in the two limbs respectively by a fortnight's 

 passage of a current of some 50 or 100 microwebers, a parti- 

 cular current was adjusted to pass by applying a given bat- 

 tery (of Minotti, De la Rue and Miiller, Leclanche, or other 

 cells, as most convenient, but chiefly the three first), and in- 

 troducing a given resistance into the circuit to measure the 

 current with. The whole was then left to itself till the next 



