440 Prof. A. Anderson and Mr. T. Keane on the 



connected to a battery of cells, was placed above A, the air 

 between A and C being ionized by means of a small quantity 

 of uranium oxide placed on A. The number of cells in con- 

 nexion with C was always greater than that required to 

 produce a saturation current in this air space. When a 

 residual charge was to be measured A was at first earthed 

 and B charged up to a definite potential for a definite time. 

 B was then earthed and A insulated immediately afterwards. 

 A special key was constructed to make the interval of time 

 between the earthing of B and the insulating of A as short 

 as possible. The electrometer will now charge up, but will 

 be continuously discharged by the ionized air between C and 

 A, and when the needle comes back to zero, the time which 

 has elapsed from the instant at which A was insulated will 

 be a measure of the residual charge, provided that, when 

 ihe needle comes back to its resting point, all the charge has 

 disappeared. This can easily be tested by finding whether 

 the rate of motion of the needle after passing the resting 

 point is the same as when no charge had been given to the 

 condenser. It was generally found that a small charge was 

 left when the needle had come back to zero, and for this 

 reason it was always allowed to move through 330 divisions 

 of the scale after having come to zero, and the excess of the 

 time required for these 330 divisions over that required to 

 describe the same interval on the scale when no charge had 

 been given to the condenser, was added to the time to the 

 resting point. The rate of motion of the needle for the last 

 40 of these 330 divisions was always taken, and if it was 

 found to be less than the rate when the condenser was with- 

 out charge, a smaller quantity of uranium oxide was used 

 until it was quite certain that all the residual charge had 

 been registered. It will be seen that the method is a very 

 convenient one for measuring either a primary or residual 

 charge, and can be quite readily applied to determine the 

 relation between the amount of residual charge and the time 

 of charging for any given applied electrostatic pressure. 

 Curves showing this relation were obtained for different 

 pressures, and show its known asymptotic character. 



The curve shown in fig. 2 gives the ratio qjQ, that is the 

 ratio of the residual charge of the condenser to the primary 

 charge for different values of the applied pressure, the time 

 of charging being always greater than that required to make 

 the residual charge closely approach its asymptotic value. It 

 was thought that, inasmuch as the amount of the residual 

 charge depends on the surface resistance, there might be a 

 difference for plates of different metals, but no such difference 



