If but a single planet revolved around the sun, then the orbit of that planet would be an ellipse, and the shape and size, as well as the position of the ellipse, would never alter.  One revolution after another would be traced out, exactly in the same manner, in compliance with the force continuously exerted by the sun.  Suppose, however, that a second planet be introduced into the system.  The sun will exert its attraction on this second planet also, and it will likewise describe an orbit round the central globe.  We can, however, no longer assert that the orbit in which either of the planets moves remains exactly an ellipse.  We may, indeed, assume that the mass of the sun is enormously greater than that of either of the planets.  In this case the attraction of the sun is a force of such preponderating magnitude, that the actual path of each planet remains nearly the same as if the other planet were absent.  But it is impossible for the orbit of each planet not to be affected in some degree by the attraction of the other planet.  The general law of nature asserts that every body in space attracts every other body.  So long as there is only a single planet, it is the single attraction between the sun and that planet which is the sole controlling principle of the movement, and in consequence of it the ellipse is described.  But when a second planet is introduced, each of the two bodies is not only subject to the attraction of the sun, but each one of the planets attracts the other.  It is true that this mutual attraction is but small, but, nevertheless, it produces some effect.  It “disturbs,” as the astronomer says, the elliptic orbit which would otherwise have been pursued.  Hence it follows that in the actual planetary system where there are several planets disturbing each other, it is not true to say that the orbits are absolutely elliptic.

At the same time in any single revolution a planet may for most practical purposes be said to be actually moving in an ellipse.  As, however, time goes on, the ellipse gradually varies.  It alters its shape, it alters its plane, and it alters its position in that plane.  If, therefore, we want to study the movements of the planets, when great intervals of time are concerned, it is necessary to have the means of learning the nature of the movement of the orbit in consequence of the disturbances it has experienced.

We may illustrate the matter by supposing the planet to be running like a railway engine on a track which has been laid in a long elliptic path.  We may suppose that while the planet is coursing along, the shape of the track is gradually altering.  But this alteration may be so slow, that it does not appreciably affect the movement of the engine in a single revolution.  We can also suppose that the plane in which the rails have been laid has a slow oscillation in level, and that the whole orbit is with more or less uniformity moved slowly about in the plane.