242 



Profs. W. E. Ayrton and John Perry. [Mar. 21, 



depends on the supply of electricity by the battery, and on a less or 

 greater resistance to rapid motion offered by the viscosity (or on some- 

 thing related to the viscosity) of the dielectric. 



If we could eliminate the effect of slow supply by the battery (and 

 this is always eliminated in the charging of glass condensers by voltaic 

 cells) we should still have not an instantaneous but a very rapid 

 charging of the condenser ; the rate of this charging does not at pre- 

 sent appear to be experimentally measurable, but ballistic galva- 

 nometers measure the time integral of the current. We know from 

 the shape of the curve which shows the increase of charge with, time, 

 and from the improbability of any want of continuity, that tne rise of 

 charge is not really instantaneous. After a very short interval when 

 the rate of increase of charge is incomparably smaller than the mean 

 rate during the interval, the total amount of charge (SV) is what has 

 usually been taken into account by physicists. But this charge does, 

 however, increase sometimes for many days to a maximum {(S+s)V}, 

 and in certain experiments with a flint glass condenser (,s) at the end 

 of about ten days was equal to (JS). If, when the charge seems to 

 increase no longer (we must not confound with increase of charge the 

 small flow into the condenser which continues even after a long time, 

 and which measures true conductivity) we connect the coatings of the 

 condenser, there is a rapid but not instantaneous loss of charge (SV), 

 and a further loss (sV) only occurring after a long period. Thus for 

 our condenser (exactly as for the strained viscous prism) the equili- 

 briated state for any given difference of potentials is one in which 

 there is a free charge, that is, a charge almost instantly removable, 

 and an absorbed charge, that is, a charge only slowly produced and 

 slowly removable. It has usually been considered that this absorbed 

 charge is quite distinct from conduction, but we know that when 

 strain is being produced in a viscous substance, and all substances 

 seem to be viscous, some of the energy is converted into heat, there- 

 fore if, as we consider, absorption is the production of strain, it must 

 be accompanied by the generation of heat, and the conversion of 

 electric energy into heat is most suitably termed conduction. We 

 cannot, therefore, have absorption without conduction, and further, 

 we consider all true conduction as an absorption in which the whole 

 quantity of electricity is converted into heat, or into some form of 

 energy not electrical. Again, when strain is produced by mechanical 

 stress, we know that the more rapid is the production of strain the 

 greater will be the amount of heat produced by internal friction. 

 Reasoning by analogy, therefore, we may conclude that as the rate of 

 production of electric strain grows less and less as the interval elaps- 

 ing since charging increases, therefore the rate of conversion of electric 

 energy into heat, that is, conduction, also grows less and less, and, 

 therefore, it is correct to say that the resistance of a dielectric does 



