153-155] 
RusselVs two Theories 
169 
the calculated mean densities of the giant stars were so small as to suggest that 
the stars must be constituted of gas of such low density that the ordinary gas- 
laws would be approximately obeyed. The densities of dwarf stars proved to 
be so high that they might be either gaseous or liquid or solid, but, if gaseous, 
they were so dense as to necessitate wide deviations from the gas-laws, at any 
rate, on the supposition, on which the whole theory was based, that stellar 
molecules or atoms remained intact and so occupied the same volume of 
space as terrestrial atoms and molecules. It was now easy to see why, in the 
giant stars, increase of temperature and density go together; this was merely 
a consequence of Lane’s law. But the dwarfs, so Russell thought, might be 
Fig. 14. Russell’s first Theory of Stellar Evolution. 
more properly compared to solid bodies, in which a loss of heat does not 
result in contraction and heating, but rather in cooling without much change 
of volume. The point Q in fig. 14 at which the sequence begun to turn Was 
accordingly interpreted as the point at which the star’s density became so 
great that the stellar matter ceased to behave like a gas. 
155. Gradually the growth of knowledge of stellar interiors made this view 
untenable. The concept which I introduced in 1917 of stellar matter consist 
ing of highly-ionised atoms and free electrons left it without any sure 
theoretical foundation, since it became an open question whether the much- 
ionised atoms retained sufficient size to cause the gas-laws to fail. Finally 
the discussions of observational material * by Hertzsprung, by Russell, Adams 
* See above § 117.