414 
Conclusion 
[CH. XVII 
continues to follow sequence (a), or at any rate a sequence which differs from it 
only in non-essentials, until, when a certain critical degree of central condensa 
tion is reached, it suddenly swings over and follows sequence ( b ), and this same 
sequence is followed by all masses having a greater degree of concentration 
than that at which the cross-over occurs. Thus, as regards essential features, 
there are only the two sequences (a) and ( b) to be considered, and a rotating 
mass will follow the one or the other, according to the degree to which its 
mass is concentrated near its centre. 
The first sequence is that through which an ordinary star breaks up to 
form a binary system. The fact that a star follows this sequence rather than 
sequence ( b ) proves that it cannot have any enormous central condensation 
of mass, detailed analysis shewing that it cannot be in a purely gaseous 
state. In its central regions at least the atoms must be jammed together so 
that the matter approximates to the liquid rather than to the gaseous state; 
this in turn shews that some at least of its atoms must have rings of electrons 
left in orbital motion around them. 
Fig. 63. The sequence of configurations of masses rotating under their own gravitation: 
(«) Liquid masses and masses of nearly uniform density. 
(fc) Gaseous masses and masses whose density is highly concentrated in their central regions. 
The second sequence of configurations ( 6 ) is that exhibited by the spiral 
and other nebulae whose masses are enormously greater than those of ordinary 
stars. When Dr Hubble set out to classify the observed nebular forms he 
tried to disregard all theoretical predictions, but nevertheless found himself 
compelled to classify the normal nebulae as forming precisely the linear 
sequence ( 6 ) predicted by theory (see Plates IX, X and XIII). 
We see then that a large proportion of the configurations of astronomical 
bodies can be explained as the configurations of rotating masses. Binary 
stars and elliptical and spiral nebulae admit of such an explanation, while 
we have conjectured that the rotation which just precedes fission may explain 
the characteristic behaviour of Cepheid variables and of some long-period 
variables. But rotation has not yet been able to explain the characteristic 
spiral shape of the arms of spiral nebulae, and definitely fails to explain the 
distinctive formation observed in the solar system.