The first of his books on the solar system is the small volume bearing that title; but since all three of the books here described are several expositions of the same subject it may be well to treat his views on each topic in connection with the work in which he deals with it most fully. Indeed, “The Solar System” is not a general treatise, but rather a discussion of some striking points, and it is these which one thinks of in connection therewith.

In considering the origin of the planets he had become much interested in the meteors, shooting stars, meteoric streams and comets, all or almost all of which he regarded as parts of the solar system, revolving about the Sun in elliptic orbits, often so eccentric as to appear parabolas.[20] The old idea that comets came from outer space and therefore travelled in hyperbolas can, he points out, be true of few, if any, of them. “Very few, three or four perhaps, hint at hyperbolas. Not one is such beyond question.” Many of them are associated with the meteoric streams with which everyone is familiar at certain seasons of the year. Indeed seventy-six of these associations were then known, and comets sometimes break up into such streams.

Now if the comets are travelling in orbits around the Sun they must be throughout their course within its control, and not within that of some other star; and therefore he computes how far the Sun’s control extends. Taking for this purpose our nearest star, α Centauri, a double with a total mass twice that of the Sun, at a distance of 275,000 astronomical units, in other words that number of times our distance from the Sun, he finds that the point at which its attraction and that of the Sun become equal is 114,000 of these units. This he calls the extent of the Sun’s domain, certainly an area large enough for any, or almost any, comet known.[21]

He then turns to some of the planets,—Mercury to show the effect of tidal action in slowing the rotation of a planet or satellite, and causing it to turn the same face always to its master.[22] This involved a highly interesting comparison of Newton’s theory of the tides, long generally accepted, but not taking enough account of the planet’s rotation, and that of Sir George Darwin based upon the effect of such rotation. The general conceptions are even more different than the results, and the later theory is less concerned with the tides in oceans, which probably affect only our Earth, than with those of a planet in a fluid or viscous condition, which may still continue to some extent after the surface has become partly solidified. He therefore studies the tide raising force, and the tendency to retardation of rotation, by the Sun on the planets, and by these on their satellites while still in a fluid state, tabulating some very striking results.

What he says about Mars is more fully dealt with in his other writings; and the same is true of Saturn’s rings, except for the reference to the calculation by Edward Roche of the limit of possible approach by a fluid satellite to its planet without being disrupted, and for the fact that this limit in Saturn’s case falls just beyond the outer edge of the rings. In discussing Saturn’s satellites he brings out a curious analogy between the order of distribution of these attendants of the three best known major planets and the order of the planets themselves about the Sun. In each case the largest of the bodies so revolving is nearly in the centre of the line, as in the case of Jupiter among the planets; the second largest the next, or not far, beyond, as in the case of Saturn; while there is another maximum farther in, for as the Earth is larger than any planet on either side until Jupiter is reached, so a like order is found in the satellites of Jupiter, Saturn and Uranus. In other words, the size in each case rises with increasing distance, falls off, then rises again to the largest and thence declines. This he believed cannot be an accidental coincidence, but the result of a law of development as yet unexplained.

To the ordinary reader the most novel thing he says about Jupiter relates to its family of comets, for no less than thirty-two of these bodies have their aphelia, or greatest distance from the Sun, near its orbit. Moreover, their ascending nodes—that is the place where their paths if inclined to the plane of the ecliptic pass through it—are close to its orbit. At some time, therefore, in the vast ages of the past they must have passed close to the planet, and if so have had their orbits greatly changed by its attraction. He considers the various effects Jupiter may have upon a comet, and shows—contrary to the opinion of Professor H. A. Newton—that any such body moving by the attraction of the Sun would be going too fast for Jupiter to capture completely. Then he takes up other effects of deflection. The comet’s speed may be accelerated and its direction changed even so much as to drive it out of the solar system; it may be retarded so that its path is contracted and the aphelion drawn nearer to the planet’s orbit. After calculating the possible conditions and analyzing the actual orbits of Jupiter’s family, he comes to the provisional conclusion that these comets have been drawn from the neighborhood. “It is certain,” he says, “that Jupiter has swept his neighborhood.... If we consider the comet aphelia of short-period comets, we shall notice that they are clustered about the path of Jupiter and the path of Saturn, thinning out to a neutral ground between, where there are none. Two-thirds of the way from Jupiter’s orbit to Saturn’s, space is clear of them, the centre of the gap falling at 8.4 astronomical units from the sun....

“Jupiter is not the only planet that has a comet family. All the large planets have the like. Saturn has a family of two, Uranus also of two, Neptune of six; and the spaces between these planets are clear of comet aphelia; the gaps prove the action.

“Nor does the action, apparently, stop there. Plotting the aphelia of all the comets that have been observed, we find, as we go out from the Sun, clusters of them at first, representing, respectively, Jupiter’s, Saturn’s, Uranus’, and Neptune’s family;[23] but the clusters do not stop with Neptune. Beyond that planet is a gap, and then at 49 and 50 astronomical units we find two more aphelia, and then nothing again till we reach 75 units out.

“This can hardly be accident; and if not chance, it means a planet out there as yet unseen by man, but certain sometime to be detected and added to the others. Thus not only are comets a part of our system now recognized, but they act as finger-posts to planets not yet known.”

We shall hear more of this last suggestion hereafter.