Constitution and Evolution of the Stars 105
supplied by gravitational contraction alone, the period
should decrease by about forty seconds per year. The
observations, which cover more than a century, show in-
deed a decrease of period, but at the rate of about a
second in twelve years—five hundred times slower than
the previous theory would demand. Here again we have
evidence that the rate of stellar evolution—in a giant star
this time—is many hundreds of times slower than it
would be if there was not some internal store of energy
to draw upon.
It is certain that no corresponding evolution of heat
from any source occurs within the Earth, and we must
therefore suppose that energy from the “unknown source”
becomes available only at exceedingly high temperatures,
such as prevail inside the stars. But if this is the case,
and a star, in contracting, gets hot enough inside to start
this process going, why does this not make the interior
still hotter, and so cause a still more rapid transforma-
tion of the unknown energy into heat, till the process
ends in an explosion on a colossal scale? I mulled over
this idea for a couple of years before I saw the simple
answer. If heat energy is supplied to the interior of a
giant star, the star will have to expand, and if it ex-
pands, it must grow cooler. The process is the exact re-
verse of that by which contraction makes the star hotter,
and at the same time compels the escape of heat from the
surface into space. Hence, if too much heat is supplied
from the unknown source, the star will expand and cool,
shutting off farther supplies. It is easy to see that we
have here a self-regulating process, which, in the long
run, will automatically adjust the supply of heat in the
interior so that it just makes up the loss due to leakage
toward the surface and radiation into space. In the