162 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1921. 



parts by fission both parts will continue to shrink, so that either or 

 both may in turn again break up, and a triple or quadruple system 

 be formed. Russell finds that in a multiple system which has been 

 formed in this way the distance between the stars formed by subse- 

 quent fissions can not be more than a small fraction, at most about 

 one-fifth, of the distance between the pair generated by the original 

 fission. A mere glance at a catalogue of multiple stars will show 

 that this condition is fulfilled by the majority of observed systems. 

 On account of foreshortening the apparent separations will not 

 always appear to conform to the rule, but Russell has shown, as the 

 result of a careful statistical discussion, that the exceptions agree, 

 both in kind and in number, with what might be expected from 

 foreshortening. 



We have now traced out the life-history of a rotating and shrink- 

 ing mass from beginning to end, from its start as a gaseous mass 

 of very low density, through its assumption of a lenticular shape 

 and its first break-up as a spiral nebula, through its subsequent con- 

 densation into separate stars, to their final fissions into binary and 

 multiple systems. The picture has been distressingly incomplete, 

 and it can not be denied that the story is beset by many difficulties 

 and uncertainties. The mathematical investigation is far from per- 

 fect ; gaps in theory have frequently been bridged by nothing more 

 substantial than conjecture; in many cases there has been room 

 for grave doubt as to the identification of observed formations with 

 those predicted by theory ; in one instance at least a formation pre- 

 dicted by theory, the ellipsoidal star, is practically unknown to the 

 observing astronomer. But, after allowing for all imperfections, 

 we have a tolerably complete knowledge, so far as the main out- 

 lines are concerned, of the whole chain of configurations which will 

 be assumed in turn by the rotating shrinking mass of Laplace, and 

 on this chain there does not appear to be any room for the solar 

 system. 



Apart from this, there are weighty reasons for thinking that our 

 system has not been formed as the result of a rotational break-up. 

 The angular momentum of a system remains constant during a 

 process of breaking up, and, as was pointed out by Babinet in 1861, 

 even if the whole angular momentum of the solar system were now 

 concentrated in the sun it would still have less than a quarter of 

 the angular momentum requisite for breaking up at its present 

 density. Except in the improbable event of the solar system, since 

 fission, having been robbed by a passing star of by far the greater 

 part of its angular momentum, its rotation can never have been 

 sufficient to cause a break-up. Clearly there is a case for examin- 

 ing whether some other agency can not produce a system such as 

 ours. 



