196 ANNUAL, REPORT SMITHSONIAN INSTITUTION, 1923 



Clearly it would be impossible to form such a group if each star 

 were the product of an accidental collision. The only way in which 

 a common motion like this can arise is by associated development 

 from some nebula or other diffuse distribution of matter. The 

 connection is clearly a connection of common origin. Again, prac- 

 tically all the bright stars of Orion form a similar group, having 

 common motion; and, moreover, they have all reached a similar 

 stage of evolution. They are connected with the great Orion nebula, 

 the faint extensions of which fill up nearly the whole constellation. 

 It is obvious that here we have to deal with a single evolutionary 

 development. But another point which militates against a collision 

 theory is the extreme rarity of collisions and close approaches. The 

 distances separating the stars are enormous compared with their 

 own dimensions. Sir Frank Dyson once used the illustration of 

 20 tennis balls, distributed at random throughout the whole interior 

 of the earth, to give a model of the density of distribution of the 

 stars. It has sometimes been objected that we do not know how 

 many extinct stars may be wandering about and colliding. Dyson's 

 20 tennis balls represent only the luminous stars; there may, for 

 all we know, be millions of dark bodies ready to be fired into in- 

 candescence by collision. I think, however, that there is now good 

 evidence, based on the dynamics of stellar motions, that the dark 

 stars can not greatly outnumber the luminous stars — probably not 

 ten times and certainly not a hundred times. (If they were more 

 numerous than that, the average velocities of stars would, owing 

 to the gravitational attraction, be much higher than is observed.) 

 That argument, then, is no longer valid. Taking a very liberal view 

 of the kind of approach that can be held to constitute a collision it is 

 estimated that a star would only suffer collision once in 10 14 years. 



Thus the astronomer is not predisposed to look favorably on a 

 hypothesis of the origin of the solar system which postulates any- 

 thing of the nature of a collision. He has the conception of an 

 orderly development of the stars crystallizing out of the primordial 

 material, and, unless perhaps in exceptional cases, following an 

 undisturbed course of development. We hope for a theory that 

 will show us the star after its first isolation from surrounding 

 material spontaneously developing the system of planets. 



It now appears almost certain that, whether the original matter 

 was gaseous or whether it was composed of meteors, it must at 

 an early stage in the star's history have been completely volatilized 

 into gas. This was while the star was extremely diffuse, and, for 

 example, before the planets separated from it. This means that 

 the material now forming a planet has at one time passed through 

 the furnace, and has cooled down from a gaseous stage. How far 



