416 
Conclusion 
[CH. XVII 
distribution of matter in this plane would be unstable, gravitational 
instability would again cause condensations to form and the shed matter 
would ultimately break up into a series of separate detached bodies. Calcula 
tion shews that these would each have a mass of the same order as the masses 
of the stars. This makes it exceedingly likely that we have here found the 
birthplaces of the stars. It is significant that stars are observed in abundance 
in the outer regions of the spiral nebulae, but none in the inner lenticular 
regions, or in those spherical or elliptical nebulae which, having stopped short 
of the critical lenticular configurations, have shed no matter out into their 
equatorial planes. Thus stars are found precisely in those regions in which 
theory predicts that they should be formed by gravitational instability and 
nowhere else. 
Normally the reign of gravitational instability must end with the birth of 
stars. The masses of the stars are too small for further shrinkage to carry 
them again along sequence ( 6 ), so that if such shrinkage occurs, the stars 
must follow sequence (a) and may finally break up into binary systems. 
In a few exceptional cases, however, gravitational instability may come into 
action again. On rare occasions it may happen that one star passes so near to 
another that it draws out long arms of matter, these being in effect exaggerated 
tides caused by the near proximity of the two stars. When this occurs, the 
matter in these arms is a fit subject for the operation of gravitational instability, 
and calculation shews that the matter would condense into detached bodies, 
each of mass about equal to that of the planets. In this way we conjecture 
that our earth and the other planets were born out of the sun. Such planets 
as do not liquefy or solidify at once may in their turn be caused to eject long 
arms of matter which the operation of gravitational instability will break up 
into detached masses, the satellites of the planets. Our moon forms a rather 
exceptional case, having a mass far more nearly equal to its primary than is 
found anywhere else in the solar system. We have found that this indicates 
that the earth must have been partially liquefied before the moon was born. 
The action of gravitational instability must finally end with the birth of 
satellites. To escape the fate of dissipating away into space these must con 
dense into either liquid or solid form immediately after birth, and when they 
have done this, gravitational instability can obtain no further hold over them. 
We have found, however, that gravitational instability accounts for the birth 
of four generations of astronomical bodies in succession, of nebulae out of 
chaos, of stars out of nebulae, of planets out of stars and of satellites out of 
planets. Our conclusion that these successive generations are born by gravi 
tational instability demands no hypotheses beyond the presence of forces which 
are already known to exist, namely gravitation and gas-pressure, and it survives 
at every step the test of numerical computation. 
387. Another section of cosmogony deals with the time required for these 
processes to occur and, as a consequence, with the total age of the universe.