i68 THE EVOLUTION OF 



generally necessary to assume that, in addition to their 

 differing in brightness, the stars are unequal in size. It 

 appears that it is only, as happened to be the case with 

 the approximate values that we assumed, when the sum of 

 the two minima is equal to the maximum, that the stars 

 must be of the same size. To account for the actual 

 minima of ft Lyrae which are 34 and 64 per cent, respec- 

 tively of the maximum light, it is, in fact, necessary to 

 regard the surface of B as f| of that of A , and its surface 

 brightness the light radiated by each unit of area of its 

 surface as of that of A. The reader will find it an 

 interesting exercise to deduce the two minima values upon 

 these assumptions. 



So far we have found a perfect explanation of the 

 maxima and minima of stars of the ft Lyrae type, and 

 we may now proceed a stage further. According to the 

 eclipse theory, the light should acquire its maximum 

 value as soon as the rise of B clears that of A in the 

 course of its revolution round it, and it should remain at 

 this value until B in its turn begins to be concealed 

 behind A. It is clear that the greater the distance 

 between the stars, the longer the time for which this 

 maximum phase should last. In the actual case, how- 

 ever, there is no constant maximum at all, the light 

 begins to fade immediately after reaching its maximum, 

 and B must therefore begin to be eclipsed by A imme- 

 diately it has finished eclipsing it. The only condition 

 under which this can happen is when B is in actual 

 contact with A , gliding round close to its surface, so that 

 we seem to be presented with a case similar to that of 

 the Earth and Moon in the first stage of their separate 

 existence, a conclusion that is very suggestive. 



Having reduced the system of ft Lyrae to that of two 

 stars in contact and in mutual revolution, we can now 



