Insulating Properties of Indiarubber. 517 



a result comparable with that at the higher voltage, the 

 resistance at 51 volts after one hour's charge is 8 per cent, 

 higher than that after the same electrification at 197 volts. 



12. The further experiments on the cable and the paraffin 

 condenser have shown that the charging current may be 

 represented by an equation of the form 



c-c = Kt-*, 



where c = charging current, K and x are constants, and 

 ^ = time in seconds since electrification commenced. When 

 t is very great, t~ x =0, and the current becomes constant and 

 equal to c . This constant current c may therefore be a 

 current due to true conduction, while the variable part, 

 Kr~ x , may represent the energy which is being slowly absorbed 

 by the dielectric, and which is given up on short-circuiting 

 the condenser. On this hypothesis, we may call the ratio 

 of the charging pressure to c the true resistance of the 

 dielectric. From the results given in Table V., it will be 

 seen that this resistance, though as high as 9520 megohms 

 on Dec. 14th, 1889, when the cable had been in water for 

 17 days, had fallen to 3310 megohms on April 10th, when 

 it had been immersed four and a half months. From the 

 experiments of March 28th & 30th, it appears that the 

 discharge current is given by equations of the form 



■= KE 6fcP 



where B = charging pressure, £ = time in seconds since the 

 commencement of discharge, and K, K x are constants. The 

 exponent x varies with the time of charge and is given by an 

 equation of the form: x— X=K 2 T~ S , where T=time of 

 charge in seconds, and X, K 2) and z are constants. If the 

 time of charge be very great, the term K 2 T~~~ becomes zero 

 and t r = X, that is, X is the value of the exponent when the 

 charging is continued until the current becomes constant. 

 The above equations have been found to hold for the discbarge 

 from T = about 35 seconds, until the current falls to 2 x 10~ s 

 amperes, which is the smallest current readable with any 

 accuracy on the galvanometer. It does not seem probable 

 that these equations would hold for the first half minute of 

 discharge, as, if so, the curves of discharge would cross each 

 other when T = from 8 to 12 seconds. It appears more 

 probable that if readings could be obtained of the discharge- 

 currents from a few seconds after discbarge, the fiVe straight 

 lines shown in fig. 8 would, instead of passing through one 

 point, tend to coalesce into one line as that point is reached. 

 Phil. Mag. S. 6. Vol. 2. No. 11. Nov. 1901. 2 M 



