256 The Evolution of Binary and Multiple Stars [OH. XT 



We notice in the last two lines that increasing period goes with increasing 

 eccentricity, as it ought to do. The parallelism becomes still more striking 

 when visual binaries are included : Campbell gives the following table : 



Type of Star Number Mean Period Mean e 



31 2-59 days 0'04 



Spectroscopic 13 6'90 0'14 



Binaries 33 73'5" 0'36 



15 20-5 years 0'38 



Visual ( 25 32'8 0'48 



Binaries j 25 1081 0'51 



in which the increase of eccentricity with period is very apparent. 



Returning to the original table, it will be seen that the entries form 

 roughly a slanting diagonal thus: -\. Advancing spectral type goes with 

 increasing period and eccentricity, and these according to the tidal friction 

 theory increase with age. The inference drawn by Campbell and others is 

 that, generally speaking, age and advancing spectral type go together*. The 

 youngest binaries are of types 0, B\ then come types A, F, and finally types 

 G M. This would bring us back to our theoretical conclusion of 259 that 

 fission takes place at about .B-type, but we shall immediately find reasons for 

 modifying very considerably this interpretation of Campbell's table. 



268. Increasing separation of the two components of a binary star, whether 

 under tidal friction or otherwise, requires an increase in the orbital momentum 

 of the system. So long as the system remains free from external disturbance, 

 the total angular momentum of the system must remain constant, so that 

 the increase of orbital momentum is necessarily gained at the expense of 

 the rotational momenta of the constituent stars. When the masses of the 

 two components are very unequal there is a large store of angular momen- 

 tum in the rotation of the more massive one, and separation can proceed 

 very far before this has all been transferred to orbital momentum. But, as 

 Russell has pointed outf, conditions are very different when the components 

 are of approximately equal mass, as is the case with the majority of binary 

 stars ( 2). 



Consider a binary star whose components are of masses M, M'. Allowing 

 for the finite sizes of the components and for their distortion from the spherical 

 shape, the force between them may be supposed to be 



* Campbell, Lick Obs. Bull. 181 (1910), p. 42 ; Stellar Motions, p. 269. Eddington, Stellar 

 Movements, p. 178. 



t Axtrophys. Journ. 31 (1910), p. 185. 



