In a paper read before the Geological Society, December 15, 1830,[897] Sir John Herschel threw out the idea that the perplexing changes of climate revealed by the geological record might be explained through certain slow fluctuations in the eccentricity of the earth's orbit, produced by the disturbing action of the other planets. Shortly afterwards, however, he abandoned the position as untenable;[898] and it was left to the late Dr. James Croll, in 1864[899] and subsequent years, to reoccupy and fortify it. Within restricted limits (as Lagrange and, more certainly and definitely, Leverrier proved), the path pursued by our planet round the sun alternately contracts, in the course of ages, into a moderate ellipse, and expands almost to a circle, the major axis, and consequently the mean distance, remaining invariable. Even at present, when the eccentricity approaches a minimum, the sun is nearer to us in January than in July by above three million miles, and some 850,000 years ago this difference was more than four times as great. Dr. Croll brought together[900] a mass of evidence to support the view, that, at epochs of considerable eccentricity, the hemisphere of which the winter, occurring at aphelion, was both intensified and prolonged, must have undergone extensive glaciation; while the opposite hemisphere, with a short, mild winter, and long, cool summer, enjoyed an approach to perennial spring. These conditions were exactly reversed at the end of 10,500 years, through the shifting of the perihelion combined with the precession of the equinoxes, the frozen hemisphere blooming into a luxuriant garden as its seasons came round to occur at the opposite sites of the terrestrial orbit, and the vernal hemisphere subsiding simultaneously into ice-bound rigour.[901] Thus a plausible explanation was offered of the anomalous alternations of glacial and semi-tropical periods, attested, on incontrovertible geological evidence, as having succeeded each other in times past over what are now temperate regions. They succeeded each other, it is true, with much less frequency and regularity than the theory demanded; but the discrepancy was overlooked or smoothed away. The most recent glacial epoch was placed by Dr. Croll about 200,000 years ago, when the eccentricity of the earth's orbit was 3·4 times as great as it is now. At present a faint representation of such a state of things is afforded by the southern hemisphere. One condition of glaciation in the coincidence of winter with the maximum of remoteness from the sun, is present; the other--a high eccentricity--is deficient. Yet the ring of ice-bound territory hemming in the southern pole is well known to be far more extensive than the corresponding region in the north.