THE ASTRONOMY OF THE INVISIBLE 373 



Of especial interest to us is the fact that the nearest of 

 the stars likewise forms one of these binary systems. Alpha 

 Centauri has a double ; their period of revolution, as estimated 

 by See, is eighty-nine years. They are separated apparently 

 about as widely as Sirius and its companion, and their combined 

 mass is computed at about double that of our sun's. When 

 we put with this Sir David Gill's conclusions, on other grounds, 

 that alpha Centauri is almost the duplicate of our sun in tem- 

 perature, density, and the like, we may conclude that the com- 

 panion is of about the same density and only a little less 

 luminous. 



Of the six binary systems of which both the orbit and parallax 

 are known, and from which therefore their mutual distances 

 may be estimated, there is only one in which the space of separa- 

 tion seems greater than the distance of Neptune from the sun. 

 In one of the nearest, and therefore most certainly observable, 

 of the six, that of Procyon, the distance of its companion is 

 apparently not* more than ten times the earth's distance from 

 the sun that is, about the distance of Saturn. 



When now we compare these apparently well-fixed distances 

 with the periods of their revolution, we gain some idea of their 

 speed. It is not very different from that of the planets. The 

 period of Saturn is twenty-nine and a half years ; that of the 

 companion of Procyon is forty years. The companion of Sirius, 

 at about the distance of Uranus, completes its circuit in fifty- 

 two years, where Uranus takes over eighty. It is therefore 

 moving about twice as fast. The companion of alpha Centauri, 

 at about the same distance, moves at nearly the same speed 

 as the planet. 



But it is from the spectroscope that the most interesting 

 discoveries of binary systems have come. It is easy to per- 

 ceive the mode. If the spectroscope can disclose the radial 

 motions of the stars, it would follow that it ought equally to 

 disclose the periods of revolution of a binary system, provided 

 their motion is sufficiently rapid. If the system were at rest, 

 compared with our own, or moving more or less at right angles 

 to the line of sight, the backward and forward motion the two 

 stars present to an observer would be represented in their 

 spectra by a backward and forward motion of the bright and 

 dark lines. If the system were moving within the line of sight, 

 the mutual revolutions of its suns would be revealed by a constant 



