ength. It is not at all likely that the "migrations" are accurately in the same time: all I mean is that they may not differ so much as the periods of our planets, for instance, the innermost of which (Mercury) takes only eighty-eight days to go round the Sun, while the outermost (Neptune) takes 160 years. It would take too long to explain the reason for and against the "migrations" being all nearly equal in point of time: perhaps you will allow me to state it without further explanation. And now let me add that this does not mean that the swings are of the same length as regards distance. We can pull a pendulum far out from its position of rest, and then when we let it go it makes a big swing: or we can pull it out only a little way and then its swing will be small; but the time of swing will remain the same. This was Galileo's great discovery about the pendulum. In the same way I fancy there are stars with big swings and stars which take only little swings and so never get far from the centre. On the whole the big bright stars will swing out farthest, and the little faint ones will swing quietly; but all much in the same time.
The question arises whether there is likely to be a collision occasionally between the stars migrating in opposite directions. This would doubtless make a considerable "flare-up," and it is tempting to suppose that the "new stars" which we see suddenly appear in the sky represent a collision of this kind. The chief difficulty is that stars are so small compared with the enormous distances between them that a collision between two must be extremely