8o 



NA TURE 



\_May 24, 1888 



I now propose to use this question of variability in 

 Group II. as a further test of my views. 



The first test we have of the theory is that there should 

 be more variability in this group than in any of the 



Fig. 



11.— Explanation of the variability of bodies of Group II. (1) Maximim variation. The ellipse represents the orbit of the smaller swarm, which 

 revolves round the larger. When the variation is great, the orbit of the revolving swarm is very elliptical, so that at periastron the number of 

 collisions is enormously increased. 



FlG * I2 'T.^ylf, n ^ i0 ?^° f th £. va £ iabi ' ;t y of bodies of Group II (2) Medium variation. There will be a greater number of collisions at periastron than at 

 new twiddle of°the rima Vanatl0n In the hght WlU not be w ^ S reat under the conditions represented, as the revolving swarm never gets very 



others. Others are as follows. (2) When the swarm is j lisions, but (3) when it is fairly condensed, the effect at 

 most spaced, we shall have the least results from col- | periastron passage (if we take the simplest case of a 



