107 
95-98] Atomic Weight of Stellar Matter 
Even a moderate decrease in temperature throughout the star lessens this 
factor enormously. Reducing the value of T uniformly to half its value, 
divides this factor by (2) 7 ' 5 , which is equal to about 181. We can now divide 
our calculated values of N*/A by 181, and find that, from being impossibly 
high, they have now become far too small to be at all plausible. Thus 
reducing T by something less than a half must lead to reasonable values of 
the atomic weights. 
The pressure in a star’s interior is fixed, as regards order of magnitude at 
least, by the circumstance that it has to support all the layers above against 
gravity. If the stellar matter obeys the laws of a perfect gas, this, in con 
junction with the fixed mean density of the star, leaves no opportunity for 
substantial adjustment of the temperature so long as the gas-laws are obeyed. 
Imagine that the temperature is artificially reduced to a certain uniform 
fraction 6 of its value throughout one of the gaseous stars we have been 
considering, each particle of the star retaining its position, so that the density 
and the star’s gravitational field remained unaltered. The gas-pressure is 
now reduced by a factor 6 and the radiation-pressure by a factor 6 4 , so 
that the total pressure is inadequate to support the star against its own 
gravitation. The star would start to collapse except that a new factor may 
immediately come into play. Let us suppose, to take a definite illustration, 
that the original temperature was so high as to keep the majority of the 
atoms ionised right down to their nuclei. The diminished temperature cannot 
maintain this high degree of ionisation, so that a certain number of atoms 
start to reform as far as their jK'-rings. Now at such densities as we have 
found would prevail at the centres of gaseous stars, if-ring atoms cannot be 
treated as mere points, so that the gas-laws will not be strictly obeyed. The 
new gas-pressure will accordingly be greater than 6 times the old pressure, 
and may even exceed the old pressure if the deviations from the gas-laws are 
sufficiently large. If it should happen that the total new pressure is exactly 
equal to the total old pressure throughout the star, the star will be in equi 
librium again. The pressure and density will have remained unaltered by 
the change, but the values of N*/A which are necessary to maintain radiative 
equilibrium are reduced by a factor of 6 7 ' 5 . 
It will of course be understood that this is a purely fictitious case; an 
actual star could not undergo the changes we have described without a good 
deal of internal readjustment. Let us nevertheless continue to use this purely 
fictitious case to examine what values of 6 would be necessary to pass from 
an ideal gaseous star to an actual star. 
We shall find later that the atomic numbers of stellar atoms are probably 
in the neighbourhood of N — 95, and so just higher than those of uranium for 
which N — 92. The value of /x for fully broken-up uranium is 2‘56; for 
stellar atoms, which are not quite fully broken up, it will be about 2‘6. The