582 Prof. 0. W. Richardson and Mr. K. T. Compton on 



The fact that the shift required for coincidence is only 

 1*15 volts instead of 1*4 volts, the true contact-difference of 

 potential between silver and magnesium, is probably due to 

 slight oxidation of the magnesium, whereby the actual 

 contact difference was reduced to 1*15 volts. So here, too, 

 the shift and the contact- difference of potential probably 

 correspond. 



This point is also significant because it disproves the 

 emission of electrons with apparent negative velocities. Since 

 the curves reach their maxima at the ordinate which corre- 

 sponds to zero volts, all the electrons possess velocities greater 

 than zero. The occurrence of electrons with apparent nega- 

 tive velocities, that is to say, of electrons which are only 

 completely liberated when there is an external electric field 

 to draw them away from the surface of the emitting metal, 

 has usually been supposed, in order to account for the fact 

 that the distribution of velocity curves have often been 

 observed to cut the axis of zero volts before reaching their 

 maxima. Such an intersection with the current- axis is pro- 

 bably due either to the neglect of the effect of the contact 

 difference of potential on the position of zero volts, or to 

 something which prevents some of the electrons from reach- 

 ing the receiving electrode unaided. Such an effect might 

 be due to a poor vacuum, to electron reflexion, to the ob- 

 struction offered by a wire screen used in the effort to prevent 

 reflexion, or to some peculiarity in the shape of the apparatus. 

 Fortunately the difficulty seems to have been avoided in the 

 form of apparatus used in this investigation. 



It is now easy to understand why the values of the kinetic 

 energies of the emitted electrons which have been published 

 differ so widely. If the uncorrected position (the dotted 

 line) were to be taken for the origin, it is evident that the 

 different metals would appear to emit electrons with practi- 

 cally equal maximum velocities for a given wave-length, 

 although actually those emitted by the electropositive metals 

 possess the highest velocities. 



In the case of sodium we have not been able to obtain the 

 complete curves showing the distribution of kinetic energy. 

 On account of the rapid photoelectric fatigue exhibited by 

 this substance, we have only been able to determine the value 

 of the maximum kinetic energy corresponding to different 

 frequencies. Different methods were tried to reduce or 

 avoid photoelectric fatigue, but they were only very partially 

 successful. In our first experiments we cut the sodium 

 surface in an atmosphere of hydrogen, carefully dried and 

 purified, in a separate vessel attached to the bulb D ; so that 



