10 
SURVEY OF THE PROBLEM 
With the substitution of radiative for convective equilibrium the constant 
y disappears; its place is taken by the numerical constant f which from 
one point of view can be regarded as the ratio of specific heats of the aether, 
aether having now replaced matter as the agent of transport of heat. 
9 . It remains to fix the appropriate value of p, and it is necessary to 
do this with fair accuracy because p is raised to a high power in many 
important formulae of the theory. We may assume that all chemical 
bonds are dissolved at the high temperature in the stellar interior, so that 
the atoms are isolated. Our first impulse is to adopt for p the average 
atomic weight of the elements which we think likely to be most abundant. 
Since iron is often supposed to be the commonest element and is moreover 
an element of medium weight, a value about 50 is suggested. The author’s 
first investigations* were made on this assumption. It was, however, 
suggested to him independently by Newall, Jeans and Lindemann that 
in stellar conditions the atoms themselves would break up to a consider 
able degree, many of the satellite electrons being detached. 
The atom is often compared to a miniature solar system. Nearly all 
the mass is concentrated in a nucleus carrying positive charge; negative 
electrons of small mass, in number sufficient to balance the positive charge, 
describe orbits round the nucleus. At high temperatures a process known 
as ionisation occurs by which these satellite electrons are successively set 
free and travel about in the material as independent particles. The molecular 
weight p appearing in our formulae (e.g. in (8-1)) is the average mass per 
independent particle. We use the term molecule to denote the particles 
moving independently of one another whether they are combinations of 
atoms or portions of atoms. If the ionisation is carried to the extreme 
limit a remarkable simplification occurs; the molecular weight becomes 
approximately equal to 2 whatever the chemical composition of the material, 
provided only that there is not an excessive proportion of hydrogen. 
The number of satellite electrons is equal to the atomic number Z of 
the element, so that if all of them are set free there will be Z + 1 inde 
pendent particles or molecules. Hence if A is the atomic weight 
y = A/(Z + 1). 
It is a well-known rule that the atomic number is about half the atomic 
weight, so that y is near to 2. Some illustrations are given in Table 1. 
Evidently the uncertainty of chemical composition is a much less 
serious matter when we realise that it is column 4 of the table which 
concerns us instead of column 3. 
In the actual conditions of a star the ionisation is not quite complete, 
and for the heavier elements some of the satellite electrons remain un 
detached. This raises the molecular weight a little. It is now possible to 
* Monthly Notices, 77, p. 16.