SIR DAVID GILL, ON THE SIDEREAL UNIVERSE. 177 



if each of these stars has a proper motion of its own. In order 

 to simplify our ideas about the question, let us imagine that all 

 the stare are moving at random. Let me define what I mean 

 by " at random." Imagine a swarm of bees and the queen bee at 

 rest upon a bougli of a tree, and that she is surrounded by her 

 swarm of bees flying about her in all directions. If the queen bee 

 looks outwards in any direction she will see as many bees flying 

 away from her as towards her, as many to right as to left, as 

 many up as down. If that were not the case the swarm would 

 not retain its general globular form. If the stars moved like 

 the bees we might call these proper motions " random-motions," 

 that is to say, motions which on the whole have no sytematic 

 tendency in any particular direction. 



Now, to find out how the sun is moving, I will adopt Mr. 

 Eddino^ton's method of illustratiuGj and discussinu; it, which will 

 be probably the easiest for you to understand. We have 

 ascertained the apparent direction of the motion of each of a 

 large number of stars, of which sufficiently numerous old and 

 recent observations exist. 



A certain area in a map of the heavens or on a celestial globe 

 is taken and a list prepared of all the stars in that area of which 

 the motions have been determined. These motions are then 

 sorted out in groups according to their directions. We then 

 count the number of stars moving in a particular direction, and 

 lay down as in fig. 2 (p. 190) a solid line from the centre towards 

 0°, representing in length and direction the number of the stars 

 observed to be moving in that direction. Similarly other solid 

 lines, from the centre towards 15°, 30°, etc., are drawn propor- 

 tional in length to the number of stars movino; in the direction 

 of each of these lines respectively. Finally a curve is drawn 

 through the ends of the lines radiating from the centre, which 

 proves to be a rough ellipse. If all these lines were of 

 equal length like the dotted lines we should assume that the 

 stellar motions are not affected by any one common cause and 

 do not show a tendency to any particular drift, but as a matter 

 of fact that is not the case, and this figure means, therefore, 

 that the star-drift as a whole is in the direction of 0°, and that 

 the sun is moving with respect to the stars in the opposite 

 direction. Now if we select a number of such regions sym- 

 metrically distributed over the sky, and form similar figures to 

 the above from the motions of the stars in those regions, 

 we find that the figures so formed would when placed in 

 their corresponding positions upon a celestial globe have their 

 longer axes directed nearly to one point, as shown in fig. 3, and 



