probability that it would return by the figures for the years 1531, 1607 and 1682 being so much alike: he inferred that it must be the same comet running regularly round a track. Then he noticed that the returns to the Sun were not quite regular: there was an interval of more than seventy-six years between the first pair and of less than seventy-five years between the second pair. This was against him unless he could give a reason for the difference. With great acuteness he assigned the reason in general terms: he said that when the comet was far away from the Sun, loitering slowly along, any of the planets which happened to come by might attract it out of its course a little, upsetting the regularity of the return. It will do us no harm to look at his actual words—
The motion of Saturn is so disturbed by the other planets, and especially by Jupiter, that his periodic time is uncertain, to the extent of several days. How much more liable to such perturbations is a comet which recedes to a distance nearly four times greater than Saturn, and a slight increase in whose velocity could change its orbit from an ellipse to a parabola?… I may, therefore, with confidence predict its return in the year 1758. If this prediction be fulfilled, there is no reason to doubt that the other comets will return.
Now, the meteor swarm which I consider responsible for sunspots can be pulled out of its course in the same kind of way as Halley's comet: and if we look for occasions when it was so attracted we
