580 DISCUSSION ON THE 
in its lifetime. We thus have in each nova an evolutionary event 
characteristic of the stars as a whole. Let us examine the event more 
closely. 
The spectrum of a nova shows that the outburst is accompanied by 
the emission of streams or jets of gas, in various directions, at the large 
velocity of some 1,500 kilometers per second. There is little doubt that 
this gas is a portion of the atmosphere driven off by the greatly increased 
radiation pressure. Some years ago when I was engaged in discussing 
the sun’s chromosphere, I found that the hotter stars would have difficulty 
in retaining certain elements such as calcium, since the pressure of the 
outflowing radiation, greedily absorbed by the atoms in their spectral 
lines, would be strong enough not only to balance gravity at the star’s 
surface, but further to repel these atoms in a kind of upward rain. Later 
I examined the probable limiting velocity that such accelerated atoms 
would attain, and I found that it was of the order observed for the gases 
expelled from novae. Thus, one of the characteristics of a nova outburst 
is that the huge increase in brightness blasts off, almost explosively, 
portions of the star’s atmosphere, and that these clouds of gas travel 
outwards through space. We shall see the evolutionary significance of 
this later. 
This driving off of atoms probably has its maximum effect when the 
star is at its brightest. What of the subsequent events as the star fades ? 
The atmospheric layers, or some of them, continue their journey outwards, 
but what of the mother star itself? Let us re-examine such a star a 
few years after the outburst. We find that it has returned to the same 
undistinguished brightness as it possessed before the outburst—but with 
a difference. Its spectrum is now of what is called type O, or Wolf-Rayet, 
indicating a very high surface temperature or surface brightness. From 
this we can make a most important deduction. If the total brightness of 
the star is the same as before the outburst, but the surface brightness— 
the brightness per square mile—is much greater than before, the total 
area of the star must be much smaller than before. And if the area is 
smaller, so must be the radius and the volume. The reduction in radius 
as observed may be as much as 10:1 or even more. Thus, after the 
outburst the star is found to have shrunk. We must resist the temptation 
to describe it as the shadow of its former self. It is smaller in dimensions 
but much brighter per square mile, and as bright as before in the aggregate. 
We have the paradoxical situation that the outer parts expand, forming 
an advancing outward-moving cloud, but the inner parts contract. 
We can only see the shrunken star when the expanding clouds have — 
moved out of the way. During the outburst itself the expanding atmo- 
spheric gases obscure our vision of what is happening. We can only 
conjecture what goes on behind the expanding screen. But there can 
be little doubt as to what is actually happening. The expulsion of the 
atmospheric clouds is a consequence of the brightening, but the shrinking 
of the rest of the star is not just an accompaniment of the brightening— 
it is the actual origin of the brightening. When the star contracts it 
must lose gravitational energy, which is set free as heat and light. — 
Somehow it must get rid of this energy. If it got rid of it slowly there - 
would be little visible effect. Actually we see the star, in its period of 4 
