286 
IONISATION, DIFFUSION, ROTATION 
Whilst the secondary east and west currents are of immediate ob 
servational importance, the primary currents in meridian planes are of 
great interest. In the first place, they will keep the material of the star 
stirred. The stirring will be slow—far too slow to change the equilibrium 
from radiative to convective—but rapid compared with the diffusion of 
the elements discussed in § 196. Thus any tendency of the heavy elements 
to concentrate at the centre will be frustrated. We shall have occasion 
to use this argument in discussing subatomic energy : owing to the stirring, 
when a star divides into two components each component will have the 
same chemical composition, and there is no chance for the larger component 
to take more than its share of the heavy elements. This conclusion is 
subject to one reservation: Bjerknes has pointed out that a circulation 
of this kind tends to become stratified, so that instead of one circulation 
between the centre and outside we may have two or three layers of circu 
lation. Each layer will then be thoroughly mixed, but there will be little 
interchange between consecutive layers. 
Another point is that the core of the circulation will probably be a 
pair of vortices, each vortex being of the form of an anchor ring about the 
polar axis. This is the postulate of a theory proposed by V. Bjerknes to 
account for many of the magnetic and periodic properties of sunspots. 
At any rate, the recognition of an internal circulation of this kind offers 
a hope of explaining many details of the surface phenomena of the sun 
which would be difficult to account for in an entirely static star. 
It has been urged by B. Gerasimovic* that what I have here called the 
secondary currents are really the principal currents. The currents in 
meridian planes would, he considers, not be permanent, whereas a dis 
tribution of circulation about the axis of rotation can have secular stability 
under the dissipative force of viscosity. Admitting that there is a condition 
of circulation about the axis which does not tend to alter through viscosity, 
and that there is a condition of circulation about the axis which satisfies 
von Zeipel’s condition (generalised to take account of varying co), it is 
very unlikely that these two conditions would coincide. It is just as im 
probable as that (198*1) would be satisfied in a static state. Therefore I 
think there is no question of the circulation maintaining itself without 
motive power; the violation of von Zeipel’s condition gives an unequal 
heating which supplies the motive power for the currents required to 
restore the condition; and the important point is that the motive power 
primarily causes currents in the meridian planef. 
* Observatory, 48, p. 148. 
f I think that this shows that the circulation in meridian planes will be at any 
rate sufficient to keep the light and heavy elements from separating. Gerasimovic 
may be right in holding that this circulation remains small compared with the 
east and west circulation (owing to the greater friction); in that case the result 
would not be so favourable to Bjerknes’s theory.