THE COEFFICIENT OF OPACITY 
217 
147. There can be little doubt that the principal process of absorption 
in the stars is the photoelectric effect. A quantum of radiation is absorbed 
and its energy is employed in removing an electron from the atom and 
endowing it with kinetic energy. Other processes of absorption are known, 
but these contribute relatively little to the stellar opacity. Ionisation 
reduces the absorption because it leaves fewer electrons capable of per 
forming the photoelectric process. The decrease may also be described— 
but perhaps less accurately—as a saturation effect; the atoms cannot deal 
efficiently with a very large quantity of radiation which removes their 
electrons faster than they can be replaced. The ionisation is the sign of 
overwork. We said less accurately because ionisation can also be caused 
by collisions of atoms and electrons, so that in certain circumstances the 
decrease of the coefficient might occur without implying great intensity 
of the radiation; but, on the other hand, the principle of detailed balancing 
in thermodynamical equilibrium permits us to deal with the photoelectric 
effect and its converse as if these were the only processes occurring in the 
stellar interior, so that the small absorption coefficient in the stars is 
directly attributable to the intensity of the radiation. 
The practical effect of ionisation on the absorption coefficient will be 
seen from the following numerical results which have been calculated for 
iron at the temperature and density of the centre of Capella*. If X rays 
of similar wave-length but of ordinary laboratory intensity were directed 
on iron in a terrestrial experiment they would be very strongly absorbed. 
The absorption is mainly performed by the L group of 8 electrons. As 
each quantum is absorbed, an L electron is expelled; but in terrestrial 
experiments the L group is completed again by the falling in of an outer 
electron or by capture of a wandering electron before the atom’s turn for 
another absorption. The coefficient of absorption by the L electrons alone 
is found experimentally to be 2950 c.G.s. units. Now let the intensity of 
the X rays be raised to the actual strength in Capella. The L electrons will 
be wrenched away almost immediately they take their places so that they 
are not usually present in the atom. At any moment only 1 out of 1200 
places for L electrons is filled; that is to say, instead of each atom having 
eight L electrons, only 1 atom in 150 has even a solitary L electron. The 
L absorption coefficient has therefore only J 200 if s laboratory value 
and is reduced to 2-5. Actually more absorption in Capella is performed 
by the two K electrons. In the laboratory these do relatively little work, 
because most of the radiation has a frequency too low to operate the 
K mechanism; the laboratory absorption coefficient is 8-3. But in 
Capella about 70 per cent, of the K electrons are in their places at any 
* Monthly Notices, 84, p. 113. Corresponding results for other elements are also 
given.