242 The Evolution of Star -Clusters [OH. x 



This scheme gives a fair representation of the J/-stars, but when we pass 

 through K, G, F and A -stars, the representation becomes less good. If we are 

 right in believing the 3/-stars to be those of least mass we can find a reason 

 for this. In the earlier stages of evolution of the system, when the stars were 

 much more closely packed than now, encounters between the various stars 

 were of frequent occurrence. The light J/-stars would be easily and rapidly 

 deflected from their paths by encounters with the heavier K, G, F, A -stars, but 

 these more massive stars would only be slightly affected by encounters with 

 Jf-stars. Thus we may naturally expect some departure from the equipartition 

 law to be observed in the case of these heavier stars ; we expect to find some 

 trace remaining of their original motion along the nebular arms. The actual 

 motion to be expected has been investigated; we found it to consist of a 

 motion of " star-streaming," the directions of star-streaming forming circles 

 round the axis of symmetry of the whole figure. This, according to Charlier, 

 is the type of motion observed, and the amount of star-streaming, which is 

 insignificant for the 3/-stars, becomes progressively greater as we pass through 

 types K, G, F, A. 



251. Thus the observed stellar motions would admit of a simple and 

 natural interpretation if only the .B-stars did not exist. The J?-stars fail to 

 carry on the sequence of the other types K, G, F, A in almost every respect. 

 No very satisfactory explanation of their anomalous behaviour suggests itself, 

 but a conjecture may perhaps be risked. 



Equation (543) suggests that the most massive stars may perhaps have 

 been formed when the primitive nebula had its greatest density *, and therefore 

 in the latest stages of its history. Thus, dividing the stars according to their 

 present spectral types, the stars which are now 5-stars may, statistically, 

 have been created last. They may be the youngest type of stars. This con- 

 ception may go some way towards reconciling Russell's theory of stellar 

 evolution with older views to the relative ages of stars of different spectral 

 types; we have already had occasion to notice ( 201) that it is in no way 

 inconsistent either with Russell's theory or with the results of our own 

 theoretical investigations on stellar evolution. 



Thus we may conjecture that the 5-stars fail even to approximate to law 

 (576) because they have not had time in which to begin to do so ; they were 

 created when the universe was already not far removed from its present state 

 in which collisions and close encounters of stars are very rare. They have 

 been acted on only by the main gravitational attraction of the universe as a 



* The mass depends on C 3 />~^. For homologous contraction p varies as T' J or as C H , so that 

 the masses would all be uniform. But homologous contraction is an ideal process, occurring 

 only if the physical properties of the gas remain unaltered with changes of temperature and 

 density. It is possible that actually T 3 may have varied more rapidly than p, leading to the result 

 stated. 



