THE BINARY STARS. 321 



of our knowledge on two points : — First, the mean angular dis- 

 tance of the two bodies apart ; and secondly, the distance at which 

 they are situated from us. (Their period of revolution it is com- 

 paratively easy to arrive at by direct observation.) If Sirius is less 

 than loo billion miles away, then, of course, an angular separation 

 between him and his couipanion of 7 seconds of arc would not 

 mean so great an actual distance between the pair as 3388 million 

 miles. If Sirius is further off than we think, then, of course, 

 seven seconds of arc separation between the pair would mean a 

 greater number of miles between them. If, on the other hand, 

 our measurement of the angular distance is not quite correct, but 

 is either too great or too small, then the result of our calcula- 

 tion will be modified accordingly. 



There are many other binary stars, the distances of which from 

 the solar system are approximately known, and which it has there- 

 fore been possible to weigh in a similar way to Sirius and his com- 

 panion by careful measurement of the distance of the pair from 

 each other and observation of their periods of revolution. 



The second of the four classes into which I have divided the 

 double stars — namely, those in which no revolution has been 

 detected, although the proper motion of the pair through space is 

 in the same direction and along parallel lines, indicating that most 

 probably some physical connection exists between them — is not 

 nearly so interesting as the first class, or binaries proper, and it 

 has added little to our astronomical knowledge. 



To the third class, or merely optical doubles, we are indebted 

 for our knowledge of the distances at which some of the stars are 

 situated from us. The apparent shift in the position of the nearer 

 of the two with regard to the more distant one, when viewed from 

 opposite sides of the earth's orbit at intervals of six months, being 

 the basis of the very elaborate and difficult calculations which 

 have afforded us all the information we at present possess on this 

 very interesting subject of stellar distances, a description of the 

 methods by which this is accomplished is beyond the scope of this 

 paper, and I will therefore pass to the consideration of our fourth 

 class, or the spectroscopic doubles, which are two stars so close 

 together in comparison to their distance from us that the spectro- 

 scope alone is able to divide them. Until the wonderful power 



