302 
THE SOURCE OF STELLAR ENERGY 
atmosphere its speed will increase by the electrical acceleration until it 
happens to ionise an atom; that will cause a discontinuous drop followed 
by an increase of speed until the next ionisation occurs. Even if the net 
result is at first an acceleration, the frequency of the ionisations increases 
with the speed so that the brake becomes more powerful and a limiting 
speed is reached. But for fast-moving particles the conditions are different; 
the ionising power increases with speed only up to a certain point and 
then falls. If the speed of the particle passes this critical point it can go 
on increasing indefinitely since the brake offers less and less resistance. 
Wilson suggests that in thunderstorms these runaway particles may 
occur; and picking up practically the whole energy of the potential drop 
(about 10 9 volts) they will surpass in energy anything else that is known*. 
Although it is not clear that anything of this kind could occur in the 
interior of a star, it gives food for reflection. If local fields, such as occasion 
terrestrial thunderstorms, exist in the stellar interior an electron going 
fast enough to get a good start will proceed with ever-diminishing chance 
of capture or deflection as its speed increases under the influence of the 
field; so that its free path is greatly extended and it can pick up almost 
unlimited energy. It is difficult to admit local fields of strong intensity 
in a star owing to the high conductivity of the ionised material. So far as 
we can judge the electron would have to start with very high speed in 
order to gain rather than lose energy. Numerical calculations are not at 
all encouraging. Still if a few high-speed electrons started the liberation 
of subatomic energy, this energy would itself send off other high-speed 
electrons, and in certain circumstances the action instead of dying out 
might be regenerative and maintain or multiply the number of runaway 
particles. If anything of this kind is going on the influence of temperature 
and density becomes incalculable, and other factors, more especially 
rotation which is likely to be concerned in causing local fields, may have 
to be taken into consideration. We leave this suggestion as a conceivable 
alternative, but assume it to have been rejected in the arguments which 
follow. 
Dependence on Temperature and Density. 
211. In the foregoing sections we have indicated that the physicist 
has difficulty in admitting that the rate of liberation of energy can depend 
to an appreciable extent on temperature because stellar temperatures are 
trivial from his point of view. Those who have maintained this attitude 
have, I think, been mainly influenced by the known characteristics of 
radio-activity. Disintegration of radium is a spontaneous event involving 
the atom as an isolated system, so that density is irrelevant. It could be 
* For comparison, the energy set free by annihilation of a proton and electron 
corresponds to 9.10 8 volts.