Ultra- Violet Lie/ Jit from the Mercury Arc. 401 



each unit of light releases an electron from a molecule in the 

 surface upon which the beam is incident. The velocity of 

 the electron depends solely upon the energy in the unit and 

 consequently upon its frequency. 



2. The photoelectric effect may be of the nature of 

 resonance. There may be many different atomic s} r stems 

 in matter each of which can be rendered unstable by light 

 of suitable frequency *. Here, as before, the velocity of 

 the ejected electron is determined by the frequency of the 

 light*. 



3. The electrons emitted from a surface may be in tem- 

 perature equilibrium with the beam of light. The greater 

 the frequency of the light, the higher is its radiation 

 temperature. This also leads to the known experimental 

 result that the grenter the frequency of the light, the greater 

 is the velocity of the electrons emitted. 



The principle of the experiment is as follows. The 

 distribution of velocities of electrons released by light of 

 the shortest available wave-length is measured, then an 

 intense beam of ultra-violet light of longer wave-length is 

 superposed and the distribution of velocities is again obtained. 

 On the first and second views of the photoelectric effect 

 indicated above, the final distribution of velocities would 

 simply be the sum of the distributions due to the components 

 of the beam. On the third view, however, this would not be 

 the case. The temperature of the mixed beam would be 

 higher than that of its long wave-length component and 

 lower than that of its short wave-length component, and the 

 final distribution of velocities would not be obtained by 

 the simple addition of the distributions due to the two 

 components. 



The simplest method of carrying out the experiment is to 

 measure the maximum velocity of the electrons due to a, 

 beam of short wave-length ultra-violet light ; then superpose 

 longer wave-lengths, and examine whether the maximum 

 velocity is the same as before or less. This was done, using 

 the upper portion of the apparatus in fig. 1 to measure the 

 maximum velocity. An apparent diminution in the maxi- 

 mum velocity was found, but this could be explained, to 

 some extent, by considerations of the same kind as those 

 brought up in the discussion of fig. 3. It was therefore 

 necessary to devise a method in which the disturbing effect 

 of reflected light was completely avoided. In the method 



* Sir J. J. Thomson, Phil. Mng. xx. p. 238 (1910). 



