422 Prof. Thomson, On the ionization of Gases 
light wave incident on the metal, but that the corpuscles are 
ejected by the explosion of some of the molecules of the metal 
which have been put into an unstable state by the incidence of 
the light ; if this were the case, the velocity of the corpuscle would 
be determined by the properties of the atom of the metal and not 
by the intensity of the light, which merely acts as a trigger to 
start the explosion. Some experiments made quite recently by 
Dr E. Ladenburg make, however, this last explanation exceedingly 
improbable. Ladenburg has investigated the velocities of cor- 
puscles emitted under the action of ultra-violet light of different 
wave lengths and finds that the velocity varies continuously with 
the frequency ; according to his interpretation of his experi- 
ments the velocity is directly proportional to the frequency. 
Thus, though the velocity of the corpuscles is independent of the 
intensity of the light, it varies in apparently quite a continuous 
way with the quality of the light ; this would be very improbable 
if the corpuscles were expelled by an explosion of the molecule. 
It seems more reasonable to suppose that the velocity is im- 
parted by the light, and yet as we have seen the velocity is 
independent of the intensity of the light. These results can 
however be reconciled by the view stated above that a wave 
of light is not a continuous structure, but that its energy is con- 
centrated in units (the places where the lines of force are 
disturbed) and that the energy in each of these units does not 
diminish as it travels along its line of force. Thus if a unit by 
impinging on a molecule can at any place make it liberate a 
corpuscle, it will do so and start the corpuscle with the same 
velocity whatever may be the distance from the source when it 
strikes the molecule ; thus the velocity of the corpuscles would be 
independent of the intensity of the light. Ladenburg found that 
the velocity of the corpuscle increases with the frequency of the 
light ; this shows that if the view we are discussing is correct the 
energy in the units will increase with the frequency. This seems 
to be just what we should expect, for to take the analogous case 
of Rontgen rays, where thin pulses are analogous to light of short 
wave length, thick pulses to long waves, the thin pulses will have 
the greater energy for two reasons: (1) these pulses are produced 
by the collision of the more rapidly moving cathode particles 
which have more kinetic energy to be converted into radiant 
energy than the slower ones, (2) since the time of impact of the 
quick particles is smaller than that of the slow, a larger pro- 
portion of their kinetic energy will be converted into radiant 
energy. Considerations of a somewhat similar character indicate 
that in the case of light of definite wave length the energy in each 
unit will increase with the frequency of the light. Confirmation 
of this result is afforded not only by the experiments of Ladenburg 
