Mr Campbell, Discontinuities in Light Emission. 523 



the resistance and measuring the potential difference between its 

 ends. The quantity jp, the logarithmic decrement of the charge 

 on the electrode system, was found by observ^ations on the decay 

 of the deflection of the electrometer needle : the capacity was then 



found from the relation G= -^s- • Measurements made in this way 



Kp 



showed that the capacity of the electrode system increased notably 



with a decrease of R, while it was expected that G would be almost 



independent of R. The only change made in the apparatus when 



R was altered consisted in a change of the distance apart of the 



electrodes immersed in the high resistance fluid, and the calculated 



capacity of these electrodes at their nearest approach was only 



15 cms., less than -^ of the whole capacity. Moreover, it was 



found that a variation in the resistance made no difference in 



the ratio of the capacities of the electrode system with different 



potentials on the electrometer needle : it would seem then that 



the change of capacity must lie in the electrometer and not in the 



rest of the system. 



When the resistance was sufficiently high, or when additional 

 capacities were inserted, measurements could be made of the 

 capacity by other methods. In all cases these agreed with the 

 capacities measured by the decrement method : but in those cases 

 when the measured capacity appeared abnormal, when measured 

 by the decrement method, no other method of measurement of 

 sufficient accuracy could be devised, whereby the results might be 

 checked. 



§ 10. It will be observed that, if the theory of the measure- 

 ments which has been given is correct, a knowledge of the number 

 of electrons liberated by a single light disturbance can be found 

 from the relation 



ft) / eis' 



where i is the total photoelectric current. 



In making use of this relation, only those observations made 

 with resistances greater than 0'15 have been used : it has been 

 assumed, for the reasons given at the end of the last paragraph, 

 that the values for smaller resistances are abnormal. The mean 



value of [&) + — j is then found to be 4-1, if the value of e, the 



charge on an electron, is 465 x lO"'". 



The resulting value of on depends on the assumptions made in 

 calculating rf. It will be seen that the value is inconsistent with 

 the second assumption made in § 7, when discussing the value of 

 this quantity, for on that assumption (o could not be greater 



