DOUBLE STARS 161 



shown that it must be, at any rate approximately, circular. 

 There is no indication of the relative masses, and there- 

 fore it is not possible to determine the position of the 

 centre of mass upon the line AC, but the general nature 

 of the motion is represented in fig. 12, in which the mass 

 of the star A is assumed to be twice that of the com- 

 panion C y so that, if O is the centre of mass, OC is twice 

 OA. If then we assume that the solar system lies in 

 the direction OE y the eclipse will begin when A and C 

 approach the line E'OE from opposite sides, and it will be 

 deepest when their centres lie upon it. Hence according 



FIG. 1 1. System of Algol as presented to the Earth. 



to the eclipse theory Algol must be in motion, and it be- 

 comes a question whether its motion may not be detected. 

 If a telescope of sufficient power were possible the star 

 would appear to swing up and down, from K' to K and 

 back, in a period coincident with that of the light varia- 

 tion, but the detection of this is quite impossible. The 

 star should, however, possess another motion relative to 

 the observer. It should, after eclipse, be approaching 

 him, as it traversed the half of its orbit, PQR, while for 

 the half-period before an eclipse it should be receding 

 along the half RSP, the rates of approach and recession 

 reaching their greatest values at Q and .5* respectively. 

 These movements have actually been detected, and their 



