BiCKEKTON. — On the Origin of Double Stars. 163 



remain, yet most of the body will dissipate into space, or become an ex- 

 tremely rare nebula. I have shown in my paper " On Causes tending to 

 alter the Eccentricity of Orbits,"* that this will certainly render the orbits 

 more circular ; although of course they may be left in their final state as a 

 very long ellipse. The nebula will doubtless be gradually absorbed by the 

 two stars, and as this absorption will be chiefly at perihelion, it will tend to 

 make the orbit still more circiilar. What will become of the nucleus of the 

 coalesced part itself is a difficult problem ; not improbably it would become 

 associated with one of the two stars in a highly elliptical orbit — making a 

 triple star ; or it may be drawn out into long trains by the unequal attraction 

 its opposite sides would be subject to each time the stars passed near it, 

 and finally be absorbed by the two stars. 



The accompanying diagram (Plate IIa.) is intended to represent the 

 several stages of the formation of such a pair of binaries. A and B 

 represent the two stars at a distance. They describe the hyperbolic orbit 

 represented, and if the bodies were in all respects the same, except that the 

 volume was so small as to all6w them to escape without collision, then they 

 would pass away in a curve, represented by the dotted curve on the other 

 side of the axis. If we suppose them to come into collision, and the middle 

 piece to exercise such an additional attraction that the new orbit is repre- 

 sented by the long semi-ellipse, — then, on then* return to the centre, instead 

 of passing to P' it passes to p, its permanent perihelion, owing to matter 

 passing outside its orbit and lessening the central attraction. On its several 

 passages through perihelion it suffers resistance, and is brought to a'a'. 

 After thus being retarded an mdefinite number of times, it finally takes up 

 the orbit represented by the thick line. 



Of course, the centre of gravity of such a pair may have a motion of its 

 own, and, doubtless, during the actions here illustrated, apsides would 

 rotate ; but, as such rotation is quite unimportant in accounting for the 

 final orbit, the diagram is not complicated with its illustration. 



It is probable that for many years after bu'th the two stars would be 

 variable, as it is evident that the side from which the central mass was 

 struck must be very much the hottest. Struve has discovered that more 

 than a score of double stars are variable, and a very large number more are 

 suspected of variability. The occasional variability of binaries seemed to be 

 so certain on this theory of their origin that I was searching more than a 

 year for evidence of their variability. Finally, my attention was called 

 by the late Dr. Powell to a paper in the "Intellectual Observer " for 1862, 

 in which I obtained this information. Assuming this theory to be right, 

 these variable doubles are recently associated — probably they have not been 



* See Art. XIII, 



