104 F. W. Very — A Cosmic Cycle. 



from the outer shell which has never undergone the greatest 

 pressures, is still rich in unaltered elements, and contains an 

 excess of elements of small atomic weight and low density. 

 In the majority of stars of composite type, the component 

 formed from the outer shell is the smaller, as if the separation 

 had been that of a comparatively thin layer thrown off by cen- 

 trifugal force. Where, as in a large number of optical bina- 

 ries, the luminous intensities are nearly equal and the spectra 

 similar, it is probable that the separation has been due to a 

 central explosion, dividing the original body into two nearly 

 equal portions of similar compositions. 



There does not appear to be any distinction of orbital eccen- 

 tricity in respect to spectral type. If eccentricity were gradu- 

 ally developed from an initial orbit, circular in every case, 

 there ought to be some progressive change of eccentricity with 

 change of type. Since there is none, I infer that eccentricity 

 of stellar orbits is the immediate result of forces involved in a 

 sudden disruption, and has been only slightly modified by tidal 

 evolution. 



Dr. See says of the eccentricity that " this element, which 

 depends wholly on micrometrical measures, and is independent 

 of the parallaxes and relative masses of the stars, gives the sole 

 clue to the evolution of the stellar systems, and will some day 

 enable us. to lay a secure foundation for scientific cosmogony."* 

 While I cannot make eccentricity the sole clue, it is certainly 

 a very important one, and it seems to me that its indications 

 support the view advanced here. The radial velocities pro- 

 duced by an explosion will be added to those given by tidal 

 interaction of the components of a binary system. The mean 

 orbital eccentricity of about -J, which Dr. See finds for stellar 

 binaries, need not require so long a time for its development 

 as we should infer from the theory of tidal evolution taken 

 alone. 



The wide gap separating stellar and planetary orbital eccen- 

 tricities (0"482 and 0*039 respectively) implies a different 

 origin in the two cases. It is suggested here that the differ- 

 ence is attributable to the mode of division ; that in separa- 

 tions occurring through excessive rotation, the division is into 

 very unequal parts ; that such rapid rotations imply a consid- 

 erable degree of preliminary concentration, with imperfect 

 fluidity which is not favorable to equable division, and on the 

 whole are more likely to prevail at later than at earlier stages ; 

 that at later stages explosions, while assisting division, become 

 less frequent and less powerful, owing to the elimination of 

 especially explosive substances. On the other hand, division 

 by central explosion may be expected to give more nearly 

 equal components, and ought to be more likely to occur at a 

 *T. J. J. See, Evolution of the Stellar Systems, vol. i, p. 252, 1896. 



