218 Prof. A. W. Bickerton on Cosmic Evolution. 



The two stars that struck have been heated where they 

 were sheared, find they are separating at a speed of 

 hundreds of miles a second. Hence the spectrum of our 

 nova is made up of a continuous spectrum, with broad, 

 bright, indistinct bands produced by the expanding gas, and 

 on this band are superimposed two other lines, bright or dark, 

 dependent on the position from which we view the lake of 

 fire produced by the impact. 



Obviously the tangential retardation will cause rotation, 

 and the cut stars may alternately show their light and dark 

 faces. Thus two variable stars are produced at once ; 

 generally this variability will tend to die out more quickly in 

 one than in the other, yet there are many such pairs still 

 existing. 



It is certain that such pairing is not the result of chance. 

 Whatever the explanation offered to account for variable 

 stars must account also for the existence of pairs. (The 

 accompanying diagram represents such a series of phenomena. 

 With bodies of solar density the time taken to produce the 

 changes shown in the series is less than two hours. The mass 

 of the bodies makes no difference in the time, as with bodies 

 of equal density the velocity acquired by gravitation is pro- 

 portional to the diameter.) 



The middle body attracts and retards the escaping stars, 

 and may wed them into a pair. 



Then, were no other agency to come into play, the pair 

 would return to impact again, but long before they attain 

 aphelion distance the central mass (consisting as it does of 

 gas above the critical velocity) will have fled into space. 



Hence the only force that attracts the stars back again is 

 their own mass, and consequently, instead of colliding, the 

 stars move in the ordinary double-star orbit. Double stars, 

 when first connected, would be variable, and would be 

 associated with nebula? ; this is actually the case, and any 

 satisfactory account of double stars must explain these facts. 



If the two stars had had a considerable proper motion they 

 would not have been orbitally connected, and they would 

 constantly increase their distance from each other. 



This is doubtless the condition of the unassociated variable 

 stars that are in pairs, and it is possible that their increase in 

 distance could be observed. 



Supposing too much had been cut off and coalesced, and 

 the attraction were consequently too great for the heat to give 

 every molecule more than its critical velocity; on attaining 

 equality of temperature the light atoms would rob the others 

 of their energy and escape. 



