154 THE EVOLUTION OF 



and that ultimately the mass would divide into two parts, 

 the parts rotating as a whole, and necessarily round the 

 original axis passing through the centre of mass of the 

 two. If this should be so, we should be presented with 

 a condition closely analogous to the one that Darwin has 

 shown to have been that of the Earth and Moon at an 

 early stage of their history. 



In tracing the development of the fluid mass, we have 



Axis of rotation. 



FIG. 8. Poincare's pear-shaped figure of equilibrium. 



supposed its various changes to have been caused by a 

 continuously applied twist, which would, if the mass had 

 been rigid and therefore incapable of undergoing change 

 of form, have produced a continual increase in its rate of 

 rotation. If, however, we assume an initial rotation, and 

 if we suppose the fluid to shrink, as it would if it were 

 gaseous through the radiation of its heat into space, the 

 same series of changes would occur without the application 

 of impressed force. Left to itself the spheroid would 

 shrink, and its rotation would in consequence be acceler- 

 ated. The polar flattening would therefore increase, and 



