Constitution of the Sun and Stars. 311 



Those stars in which the force of gravity is very much less than 

 on the sun appear to form a distinct subclass. The four hydrogen 

 lines are not found in them, and at the same time new spectral lines, 

 arranged in bands each of which is closely ruled and fades off on 

 the less refrangible side, make their appearance. May we not here 

 venture the suspicion that when gravity upon a star is below a certain 

 limit, such conditions prevail as compel the hydrogen, which would 

 otherwise be free, to enter into combination with some other element 

 of low vapour-density, and that the resulting compound emits that 

 spectrum of the First Order, as Pliicker has called it, which we see ? 



To account for the colours of the companions of double stars we are 

 again forced to enter upon speculative ground. If the sky be peopled 

 with countless multitudes of dark stars, which as well as the small 

 number that are visible, move only in virtue of their mutual attrac- 

 tions, it cannot be an absolutely unusual occurrence for two stars to 

 come into collision. Whenever this happens, either the two stars 

 emerge from the frightful conflagration which would ensue as one 

 star, or, if they succeed in disengaging themselves, they will be found 

 after the catastrophe moving in new orbits. If their previous courses 

 had been parabolic, it can be shown that the new relative orbit will 

 be elliptic. Hence they will return to the charge again and again, 

 and at each perihelion passage there will be a fresh modification of 

 the orbit. It is shown that these modifications will in some instances 

 be such that the perihelion distance will be constantly on the increase, 

 so that the stars will, in their successive perihelion passages, climb 

 as it were asunder through one another's atmospheres. And the 

 distance to which they will ultimately withdraw before they separate 

 will of necessity be immense, since their atmospheres must have been 

 dilated to a vast size by the friction to which they have been subjected. 

 As the stars recede from one another the amount of heat which they 

 generate at each perihelion passage is progressively less and less, until 

 at length the atmospheres of the stars shrink in the intervals between 

 two perihelion passages more than they expand when the brush takes 

 place. When this happens the final separation of the two stars is 

 imminent, and a new double star is on the point of being permanently 

 added to the sky. 



The astonishing appearances witnessed last year in T Coronse 

 seem to receive an easy explanation upon this hypothesis. They 

 are exactly what we should expect upon the occurrence of one of the 

 last perihelion passages that take place before two stars which are in 

 the state of transition into a double star finally separate. The outer 

 parts of the atmospheres becoming engaged would raise to incandes- 

 cence the region in which hydrogen only is found, thus transform- 

 ing what had previously been its four dark lines into intensely bright 

 lines. At the same time the strata that lie further down would be 

 very sensibly heated, though not to incandescence — quite enough, 

 however, to lessen temporarily in a very material degree the extent to 

 which they at other times subdue the light of the photospheres. 

 This extent would of necessity have been very great, inasmuch as the 

 enormous dilatation of the atmospheres must greatly enfeeble the force 

 of gravity upon the outer strata of both stars. 



