S90 cosmic ruiLosoru r. [pt. u, 



velocities, will move in very eccentric ellipses. In tlie next 

 place, while they will come from all parts of the space which 

 the mass originally occupied, they will come chiefly from 

 regions remote from the plane in which integration has beeD 

 most marked, — that is, from the poles of the nebula rather 

 than from its equatorial regions. And thirdly, having failed 

 to accompany the retreating mass of the nebula while it 

 was first acquiring a definite direction of rotation, theii 

 own revolutions will be determined chiefly by their irre- 

 gular shapes, and they will be as likely to be retrograde 

 as direct. 



All this is true of comets : they come chiefly from high 

 solar latitudes, along immensely eccentric orbits, and in 

 directions which are indifferently direct or retrograde. And 

 when we add that they are nebulous in constitution, it 

 appears highly probable that they are simply outlying shreds 

 of the nebula from which our planetary system has been 

 developed. As for the irresolvable patches of nebulous 

 matter which are distributed about the poles of the galactic 

 circle, their distance from us is so great that we have not yet 

 ascertained anything trustworthy concerning their motions. 

 But the fact that their position in high galactic latitudes is 

 explicable upon the same general principles which explain 

 the positions of comets, raises a presumption that their 

 ielation to the galaxy as a whole may somewhat resemble 

 that which comets bear to the solar system. Between the 

 possible careers of the nebulae and the comets, there is, 

 however, a mighty difference. The nebula which we see 

 through quadrillions of miles shining by a light of its own 

 must needs be an enormous object — enormous in mass as 

 well as in volume — and its gravitative force must be pro- 

 portionate to its size. While, therefore, its gradual con- 

 traction is likely to be attended by its development into a 

 planetary system, by a process of integration and diffe- 

 rentiation such as we have here described \ on the other 



