312 Royal Society : — Prof. Roscoe on the Chemical Intensity 



Again, it follows as a consequence of :this hypothesis that the 

 circumstances which most favour the formation of a double star are 

 when the two bodies that come into collision are of nearly equal mass. 

 Such cases must be rare ; but when they do occur, there is a very 

 high probability that the issue will be a double star. This appears 

 to account for the fact that a very remarkable proportion of double 

 stars have constituents of nearly the same magnitude. 



Another consequence is that when the stars are very unequal the 

 companion will, as it plunges over and over again through the atmo- 

 sphere of the primary, be gradually deprived of several of its lighter 

 gases ; so that when it finally gets clear it will not emit the principal 

 spectral lines of a solitary star, but others which emanate from denser 

 gases. This probably accounts for the blue, violet, and green colours 

 which are found in the minute companions of double stars. 



Another consequence is that the orbits of double stars will almost 

 always have a considerable ellipticity. 



Another consequence is that the conditions are likely not unfre- 

 quently to arise which would separate the companion into two or more 

 fragments, and that when this happens the separate pieces will 

 pursue paths which are distinct from one another and not far apart. 

 This seems to account for such systems as y Andromeda?. 



When the same conditions act with unusual violence they would 

 probably break up the companion into numerous fragments ; and it 

 is remarkable that they would at the same time be likely to cause 

 the primary to throw off a number of rings. The fragments and the 

 rings would move all in the same direction and nearly in the same 

 plane, and each fragment would rotate rapidly in the direction in 

 which it revolves in its orbit. When the fragments, as must gene- 

 rally happen, are of inconsiderable mass, their orbits would be almost 

 certain to degrade from ellipses into circles before they got quite 

 clear of the primary. Some would probably be found, when this hap- 

 pens, at the distance of the rings, others within the surface of the 

 primary, none beyond both. Those within the surface of the pri- 

 mary would fall into him and be lost. But one that lay within a 

 ring would gather by its attraction the ring round itself, and so be- 

 come covered with an immense atmosphere with which it would con- 

 tinue to rotate while advancing in its circular orbit. If this rotation 

 were sufficiently swift, the new planet would throw off rings which 

 might afterwards condense into satellites, with this peculiarity, that 

 they would always keep the same face turned towards the planet, and 

 revolve round it in the same direction and nearly in the same plane 

 in which the planet revolves round its sun. 



The speculative element in this hypothesis is so considerable that 

 perhaps we may not prudently yield to it a probability higher than 1. 

 But an hypothesis which carries up so many of the main pheno- 

 mena of nature to a single source, and which only asks us to admit 

 what is not antecedently improbable, that the number of incandescent 

 stars is but a small proportion of all that exist, seems nevertheless to 

 deserve to be stated. 



