THE SOURCE OF STELLAR ENERGY 
319 
) 
(4) will be more important than (1); the density of interstellar matter may 
be somewhere about this limit. Dark and diffuse nebulae occupy large 
tracts of the sky and must be considerably denser than ordinary inter 
stellar matter so that (3) ranks above (4) and perhaps above (1). We 
take it that the empirical relation of the intensity of the penetrating 
radiation to the position of the Milky Way shows that it is unnecessary 
to consider sources outside our own galactic system. 
Owing to the uncertainty in placing (1) we cannot say whether the 
absence of dependence on the altitude of the sun means that the sun is 
deficient as a source of penetrating radiation compared with the nebulae 
or whether it is accounted for by a smaller integral of odco. The integral 
for the stars is to that for the sun roughly in the proportion of starlight 
to sunlight so that it is very unlikely that the stars contribute much. It 
has been suggested that the younger stars are responsible for this radiation; 
if so, their emission must be exceedingly fierce; moreover, we should only 
receive the radiation from the surface layers which are at low density and 
temperature—conditions not so widely differing from the sun as to suggest 
greatly enhanced activity. 
This then points to the nebulae as the source of the radiation, assuming 
it to be extra-terrestrial. If, further, we consider that its penetrating power 
proves that it is of a frequency too high to come from other than sub 
atomic sources, we must regard it as a sign of subatomic processes (evolu 
tion of elements?) occurring in the nebulae. That is why we have considered 
the phenomenon here as possibly of astronomical significance. All our 
preconceptions tend to regard the nebulae with their exceedingly low 
density and (relatively) low temperature as most unfavourable for sub 
atomic transformations of this intense kind. But our preconceptions have 
certainly had little success in explaining stellar evolution, and we are very 
ready to remodel our ideas if sufficient evidence is forthcoming. If we 
admit that evolution of the elements occurs in diffuse matter it becomes 
easier to understand the occurrence of helium and some other elements 
in the diffuse nebulae, of calcium and sodium in interstellar space, and 
of rather advanced elements in the reversing layers of the youngest 
stars. 
A numerical calculation, whilst uncertain in its details, will give 
an idea of the cosmical magnitude of the phenomenon. Divide the mass 
of the universe into two portions, viz. M 1 which might be the source of 
the radiation and M 2 which cannot be the source. Approximately M 1 is 
the mass of the nebulae and diffuse matter and M 2 the mass of the stars; 
a thin shell from each star should be transferred from M 2 to M 1 but this 
can be neglected. Let e 1 and e 2 be the average rates of liberation of sub 
atomic energy in M 1 and M 2 . Then M 2 e 2 represents the total radiation of 
the stars. We shall find (§ 256) that at an average point in space M 2 e 2