258 REPORTS ON THE STATE OF SCIENCE. 



gravitation. But that certainly unwelcome assumption is by no means 

 necessary. A star is presumably formed by the lumping together of the 

 meteoric or gaseous matter in some portion of space. Now if we were 

 to lump together a thousand stars their individual motions would prac- 

 tically cancel and the resultant super-star would be nearly at rest. 

 Similarly in forming a single star the individual motions, produced by 

 gravitation in the matter of which it is composed, might cancel, so that 

 the star would start from rest. 



I find no difficulty in the idea that a star may be born without 

 motion, but it is more difficult to understand how it acquires motion. 

 There is a distinction well known in the theory of attractions : the 

 attraction at a point in a discontinuous medium is regarded as made up 

 of two parts — (1) the corresponding attraction in a smoothed con- 

 tinuous medium, and (2) the part due to the chance distribution of the 

 masses in the immediate neighbourhood of the point considered. Which 

 of these parts is effective in producing a star's velocity? If we fol- 

 lowed the analogy of the kinetic theory of gases we should assign the 

 motions to the second cause. But in this connection we have to con- 

 sider one very important observational fact. In the Taurus cluster, 

 and again in the Ursa Major stream, stars which are well advanced in 

 evolution have evidently preserved throughout their lifetime precisely 

 equal and parallel motions, notwithstanding that they are moving 

 through space occupied by stars not belonging to their system. Is it 

 possible that there have been exceptional circumstances which have 

 enabled these systems to preserve their common motion whereas other 

 similar systems have been broken up? It is hard to think so. If not, 

 it seems that the existence of these clusters leads to the following 

 law : — 



The forces which produce the individual velocities of stars are 

 nearly constant over large volumes of space — i.e., large compared with 

 the distance of a star from its neighbour. 



Thus in the case of the Taurus stream all the stars of the system 

 have experienced the same force, and have therefore acquired the same 

 velocity. 



This law means that it is the central gravitational attraction of the 

 universe, and not the casual interference of neighbouring stars that 

 is effective in changing a star's velocity. It seems to be in accordance 

 with numerical calculation that a neighbouring star can have very 

 little effect. The sun acts on a Centauri, the nearest star, with a 

 force which would take 400 million years to produce a velocity of 1 km. 

 per second. It is true that a close approach of two stars might occur in 

 which the force would be much greater temporarily; but it must be 

 remembered that the change of velocity during the approach would be 

 partly lost as the stars receded ; there would be no net increase of 

 kinetic energy. Some transfer of momentum would, however, take 

 place. The following rough calculation may be of interest: If a 

 huiall star were proceeding with about average velocity and in such a 

 direction that if undisturbed by the sun's attraction it would have 

 passed it at a distance equal to that of the planet Neptune, the result 

 of the close approach would be to change the star's course only 5°. 



