32 Scientific Proceedings, Royal Dublin Society. 



interrupter giving 50 interruptions per second. Several intermediate values 

 were also tried. 



With this range of resistances, and an electrometer having a sensibility of 

 about 2000 per volt, it was possible to measure accurately currents between 

 5 * 10" 13 ampere and T5 x 10- 8 ampere. By using even smaller capacities (c 2 ) 

 and a slower interrupter it should be possible to measure currents less than 

 10~ 13 amperes. In consequence of the long period of the electrometer needle, 

 it would appear possible that an interrupter which made contact only once 

 every two or three seconds might be employed. 



§ 7. An obvious modification is to convert the method into a " zero " or a 

 " compensation " method. This may readily be done by connecting the 

 contact screw A, not to earth, but to a potential dividing device, of which 

 one terminal is earthed. The condenser, c z , then feeds in an inverse 

 " compensating " current (= -n c 2 v), which tends to prevent the electrometer 

 system charging up under the influence of the current i. This compensating 

 current may be varied by changing either n, or c 2 , but preferably by changing 

 v by means of the potential divider. The electrometer readings V E are now 

 proportional to the difference between the current i being measured and the 

 " compensation " current. The total value of i is obviously 



i = n c 2 ( V E + v). 



The advantages of a method of compensation have frequently been 

 recognized in cases where small variations of a comparatively large current, 

 say in an ionised gas, have to be examined ; but the method here described 

 appears to the writer to be more satisfactory than any yet tried. 



In the practice of the compensation method small values of c 2 and n are 

 chosen, and a comparatively high potential is applied at A by means of the 

 potential divider. A large electrometer deflection may, therefore, be obtained 

 even when the current i is almost balanced by the " compensation" current. 



§ 8. It is hoped that the method outlined above may be found of use by 

 other workers. An important advantage found in practice was the economy of 

 time during a set of observations. This was due to the fact that it was 

 generally unnecessary to earth the electrometer between consecutive observa- 

 tions. Any induction effect, for instance, produced on changing the potential 

 of the ionisation vessel / by a moderate amount died away in a few seconds ; 

 and the electrometer seldom required more than half a minute to reach its 

 new position of rest. In certain classes of work this rapidity should be very 

 advantageous. 



I have pleasure in thanking Professor M'Clelland for the interest he has 

 shown in this work. 



