330 The Ages of the Stars [ch. xn 
calculate the time during which encounters must have acted to produce the 
observed distribution of periods and eccentricities. 
In the case of visual binaries we have found that both periods and 
eccentricities shew a certain degree of conformity to the final steady state law 
which would be attained after encounters had been in progress for an eternity 
of time. The observed partial conformity to this law is found to indicate an 
age of the order of 5 million million years. 
In the case of spectroscopic binaries, the orbits are so small that encounters 
with other stars can get almost no grip on the system, and so leave its period 
and eccentricity almost unchanged, at any rate through periods of millions of 
millions of years. This is in agreement with the observed fact that spectro 
scopic binaries shew very little progressive increase, either of period or of 
eccentricity, with advancing age, and what progressive changes are observed 
can properly be attributed to tidal friction and loss of mass by radiation. Thus 
we cannot estimate the ages of spectroscopic binaries from their orbits. 
The ratio of the masses of the two components of a binary system ought 
to tend towards unity with the passage of time, as a consequence of the more 
massive star changing its mass more rapidly than the less massive. Observation 
reveals such a tendency, and its amount provides a means of estimating the ages 
of binaries. The ages of spectroscopic binaries are in this way found to be 
millions of millions of years, and of the same order as those of visual binaries. 
Observation has disclosed a very marked tendency towards equipartition 
of energy in the translational velocities of the stars. This provides material for 
an alternative estimate of the ages of the stars, which is again found to indicate 
ages of the order of millions of millions of years. 
In Chapter XIV below (§ 349) we shall obtain yet another estimate of the 
ages of the stars by considering the time necessary to break up and scatter 
moving star clusters, and this will be found to confirm the estimates of 
stellar ages made in the present chapter. 
The ages suggested by these various modes of investigation are all in 
substantial agreement, and are such as to indicate that a star lives long enough 
to lose the main part of its original mass by transformation into radiation, as 
has been suggested by various other observational facts. This rules out as in 
adequate all sources of energy except the complete annihilation of matter or 
some subatomic equivalent; nothing else can provide sufficient total radiation 
for the calculated lives of the stars. 
A detailed discussion of the stars of the galactic system has suggested 
that the majority of these were born many millions of millions of years ago, 
that the process of creation was specially active at the time of birth of the 
stars which are now of absolute magnitudes 2 to 5, a period from 2 to 8 
million million years ago, and that since then it has almost ceased.