412 



Profs. John Perry and W. E. Ayrton. [May 27 r 



curves for residual charge in condensers of different dielectrics at 

 different temperatures, and so we may say that we have measured the 

 power of producing residual charge phenomena in the substances 

 experimented upon in different states, and this power, which we call 

 our " residual- charge-function," ought, according to Dr. Hopkinson, 

 to be constant for the same substance at the same temperature. Thus, 

 if vt is the residual charge at the time t after insulation, and if V was 

 the original long continued charge, vt+Y is our residual- charge- 

 function for the time t, which we may designate by Y t . Of course, it 

 is evident that if F'^ is our function for a different dielectric, then 

 F^-f-F'^ is not necessarily the same as F^-f- F7 ; in fact, the residual- 

 charge-function is only completely specified when a table is given of 

 its values at different times after insulation. We have used the same 

 name and symbol in the experiments we have been simultaneously 

 making during the last five years on strained wires and beams, and 

 on voltameters and magnets, and of course there is always the difficulty 

 experienced by Dr. Hopkinson with his more complicated function 

 " ^rt" which renders it necessary to give a table of values instead of 

 only one value of the function. Thus, for two particular kinds of 

 glass, Dr. Hopkinson finds — 



v< 5 >- B =3-5, 



f'(5)-B 

 y,(60)-B „ 



showing how with the same two glasses the ratio is affected by varying* 

 the time from -5 to 60. Again, also in his investigation of the influ- 

 ence of temperature on residual charge in the same glass, he finds 

 that at 175° F. the values of ^(1) —13 and "^(5)— 13 are respectively 

 0'38 and 0'034, with a ratio one to the other of 11 ; whereas the values 

 of these functions at 108° F. are 0'155 and 0*05 respectively with a 

 ratio one to the other of 31 . This method of measurement is subject, 

 therefore, to exactly the same objections as the easier method of which 

 we have spoken ; in fact, we are not aware that hitherto any method 

 has shown itself to be better than actually giving the curves of rise of 

 potential after insulation, or the curves of change of current during 

 charge and discharge of a condenser. This is the method which we 

 have found the easiest for the discussion of such phenomena, and we 

 may direct attention to such curves for voltameters in a paper com- 

 municated by us to the Society of Telegraph Engineers, and which 

 appeared in their Journal, vol. v, 1877, Nos. 15, 16, p. 391. Curve 

 EFGG' there shown, in fig. 4, p. 396, is especially interesting, as we 

 see from it that if a voltameter with dilute sulphuric acid be charged 

 with one Minotto's cell for a certain time, then discharged for a 



