TRANSACTIONS OF SECTION A. 81 



were two similar masses of mercury to collide, but with mixed 

 bodies this difference of temperature of different chemical sub- 

 stances would equalise rapidly by conduction and radiation. 

 Then, the hydrogen would be moving more than a dozen times 

 as fast as the mercury : thus in any cosmical collision the light 

 atoms would tend to escape into space and the heavy be re- 

 tained, thus spreading space with wandering molecules of light 

 gas. This fact is of importance in the study of the immortality 

 of the cosmos. All the phenomena connected with this escape 

 of molecules I have grouped under the title of " selective 

 escape." 



17. If the collision of two bodies were as great as a half of 

 each, it is scarcely likely they would escape each other's attrac- 

 tion ; they would whirl round each other, forming a bun-shaped 

 mass of very high temperature. In such a mass any bodies 

 originally associated with them would tend to be entrapped in 

 this nebulous mass, and would revolve about them in the 

 general plane of motion, and would tend to produce a group of 

 bodies similar to our solar system. The details of this action 

 are described in a paper on " Agencies tending to lessen the 

 Eccentricity of Planetary Motion." 



18. Where there was considerable rotation in the case of a 

 nebula composed chiefly of heavy molecules of infusible material, 

 there would be a tendency to aggregate into meteoric dust. 

 This dust would not coalesce entirely ; it would ultimately settle 

 down into a star-cluster. 



19. Amongst groups of large bodies hke nebulae, such as 

 are found at the Galactic poles, impacts would be frequent. 

 Impacts of two nebulae would produce — (1) Twin nebulae; 

 (2) Twin nebulas with an incipient spindle-nebula between 

 them ; (3) Isolated spindle-nebulas ; (4) Spindle- nebulte with 

 an incipient spiral at the centre ; (5) Complete spiral nebulae. 

 All these varieties of nebulae are to be found in abundance at 

 the Galactic poles, and have been drawn by Herscliel and 

 others. 



20. During the impact of two globular nebulae, the two 

 parts colliding would be intensely heated, there would be pres- 

 sure produced, and it will be found on examination that this 

 could find rehef only in a direction perpendicular to the plane 

 of motion : there would thus be a tendency to the formation of 

 an annular nebula, with gauze-like caps of nebulae at the two 

 poles of the ring. There are some such nebuke. 



21. The impact of two universes, such as the Magellanic 

 Clouds, would tend to produce such a configuration ; but the 

 original stars and other bodies would not be likely to come into 

 impact except in the grazing part, but would naturally spread 

 out into an annular mass of a roughly double spiral character. 

 The grazing part would be projected along the axis of the ring, 



6 



