374 
DIFFUSE MATTER IN SPACE 
excitation into kinetic energy of translation. The transfer continues until 
the temperature is raised to the equilibrium value at which it is balanced 
by the converse transfer by inelastic collisions. But when the free path 
lasts for several days conversion by this method is so slow as to be 
negligible compared with other processes. 
( b ) The light of the stars expels electrons from the metals used in 
photoelectric cells in observatories and it must have a similar effect on 
matter in space*. The electrons are expelled with a velocity given by the 
quantum law ■, T72 7 7 
u V 2 = hv — hv 1 , 
where hv 1 is the energy of ionisation. The elements affected will be chiefly 
those with low ionisation potential round about 5 volts such as Na, K, Rb, 
Ca, Ba, Al, etc. For 5 volts A x (= c/vf) is 2468 Á. The mean velocity of 
expulsion of electrons should correspond to a temperature somewhat higher 
than T Ai , say 2000 °, since all radiation between 0 and A 7 is concerned in 
the ionisation. This is an example of a high effective temperature of the 
radiation in space for a particular purpose ; evidently the absence of the 
proper proportion of radiation with wave-length greater than A 7 is quite 
immaterial!. 
It is possible that the actual conditions are even more extreme. A 
small proportion of the stellar radiation is capable of removing a second 
electron from the above-mentioned elements—at any rate from the 
divalent and trivalent elements. Recovery of an electron must be very 
slow and the dissociation may be rapid enough to keep the element as a 
rule doubly ionised. In that case the first ionisation potential of 5 volts 
will not be concerned and the transfer by ionisation will be operated by 
the still higher frequencies above the second ionisation potential—leading 
to higher values of T K . 
We have thus a continually renewed supply of free electrons with 
speeds appropriate to a temperature of some thousands of degrees. These 
will mix with the atoms and by encounters tend to bring them to their 
own temperature. But the electron will suffer some loss of energy in its 
life-time and it is important to discover whether its initial energy of 
expulsion is at all comparable with its average energy. Ultimately the 
atoms will reach the average, not the initial, temperature of the electrons. 
Within reasonable limits it does not matter how slow is the transfer from 
the electrons to the atoms because we know of no process by which the 
atoms can waste the energy handed over to them. Note that the electrons 
take the lead in this adjustment, because the electron is continually going 
* It should, however, be realised that the circumstances are not precisely the 
same. Less work is required to expel the electrons from a metallic film than from 
isolated atoms of the same element. 
t At the end of this section it is shown that this argument does not contain the 
whole truth of the matter.