SURVEY OF THE PROBLEM 
17 
widely in brightness, density and physical condition, but they mostly 
contain about the same amount of material. It is as though nature had 
a standard model before her in forming the stars, and (except for occasional 
lapses of vigilance) would not tolerate much deviation. The extreme 
range (about ^ to 100 times the sun’s mass) does not give a fair idea 
of the general uniformity of mass; our methods of observation tend to 
select the more exceptional individuals—highly luminous stars which 
are especially massive, and double stars in which the original unit is 
subdivided—from many millions of stars of normal size. A mass range 
of 5 : 1 would, I believe, include more than 90 per cent, of the stars. 
We can see in a general way the cause of this uniformity, though the 
details of the explanation are not clear. Gravitation is the force drawing 
matter together, and as it gathers in more and more material tends to- 
build globes of enormous size. We must assume that, opposed to this, 
radiation pressure is the main disruptive force*. It is only when the mass 
has reached 10 33 gm. that this check on the aggregating power of gravita 
tion emerges from insignificance; but with further increase of mass it 
rapidly rises. The increase in the proportion of radiation pressure in 
Table 2 is a measure of the increasing peril to the unity of the star. The 
stability of a fluid mass subject to radiation pressure has not been in 
vestigated ; but since a few stars are known in which the radiation pressure 
is found to amount to 80 or 90 per cent, of the whole, it is presumed that 
it does not of itself cause instability. But it may well render the star more 
liable to be broken up by a small rotation or disturbance of symmetry, 
so that the more massive the star the smaller the chance of survival. As 
plausible figures we might suggest that radiation pressure below 15 per 
cent, will not have an important effect; and gravitation will build up to 
the corresponding mass without much hindrance, unless the star divides 
under very rapid rotation according to the well-known theory; but a 
50 per cent, radiation pressure would be a serious danger, and the corre 
sponding mass could only be reached in circumstances of exceptional 
tranquillity. 
We have not in our minds any definite idea as to the stage in the 
formation of the star at which the mass accumulation is limited by action 
or threat from radiation pressure. At present we can only point to the 
significant fact that stellar masses congregate just at the point where 
radiation pressure is beginning to endanger the safety of the star, and 
larger masses occur in rapidly diminishing numbers as though their con- 
* There is no mathematical proof of this, and the speculation rests on the 
numerical agreement. We know that a star without radiation pressure is stable 
(unless the rotation is very rapid), but the stability of a star with radiation pressure 
has never been investigated. But since we observe that stellar masses cease abruptly 
when radiation pressure becomes important, we venture to forecast the answer. 
E 2 
/