428 Rate at which Electricity leaks through Liquids. [June 16 7 



Table XXX. 



Experiment 108. Olive oil. Charged to about 100 volts. 



Deflections. Ratios. 



Deflections. Ratios. 



Deflections. Ratios. 



198 



160 lis 

 130 1-20 



73 121 

 m 1>20 



49 1,25 

 4y 1-17 



40 1-17 



34 ' 

 28 121 

 23 122 



Ratios — 



Mean 1*214 



Maximum 1 *25 



Minimum 1'17 



Experiments on Residual Charge. 



Carbon disulphide showed peculiarities in the above experiments. 

 With a short and sudden charge the rate of fall of potential was much 

 quicker at the beginning than at the end of the readings. This is 

 what would be expected if there were electrical absorption, part of 

 the fall being due to true leaking, part to absorption. If the rate of 

 leak is according to Ohm's law, the ratio curve for such a liquid 

 would be inclined to the horizontal at first, but the inclination would 

 diminish with time. This is what is found in the case of carbon 

 disulphide when the condenser is charged for a short time. 



Again, the condenser, with carbon bisulphide between the cylinders, 

 was charged for a time, then quickly discharged, and its inner cylinder 

 connected with the electrometer. The deflection was at first connex- 

 ion zero, then shortly rose to a maximum value, and finally diminished 

 again after some time to zero. If the condenser was charged with 

 opposite sign, the deflection from zero was in the opposite direction. 

 If the condenser was charged first with one sign and then with 

 the other, the deflections from zero were much smaller, but they 

 appeared in the sense expected from residual charge phenomena. 



These effects were greatest just after the CS 2 had been redis- 

 tilled, but at times were totally absent. The still was cleaned, but 

 the effects after fresh distillation were as marked as before. 



Attempts to increase the effects by rendering the liquid less homo- 

 geneous were successful, for heating or cooling the condenser un- 

 equally always exaggerated the deflections. 



The following readings are a sample of numerous experiments : — 



