It has been shown in the introductory chapter that a simple change in the sun’s distance would not alone produce a glacial epoch, and that those physicists who confined their attention to purely astronomical effects were perfectly correct in affirming that no increase of eccentricity of the earth’s orbit could account for that epoch. But the important fact was overlooked that although the glacial epoch could not result directly from an increase of eccentricity, it might nevertheless do so indirectly. The glacial epoch, as I hope to show, was not due directly to an increase in the eccentricity of the earth’s orbit, but to a number of physical agents that were brought into operation as a result of an increase. I shall now proceed to give an outline of what these physical agents were, how they were brought into operation, and the way in which they led to the glacial epoch. When the eccentricity is about its superior limit, the combined effect of all those causes to which I allude is to lower to a very great extent the temperature of the hemisphere whose winters occur in aphelion, and to raise to nearly as great an extent the temperature of the opposite hemisphere, where winter of course occurs in perihelion. With the eccentricity at its superior limit and the winter occurring in the aphelion, the earth would be 8,641,870 miles further from the sun during that season than at present. The reduction in the amount of heat received from the sun owing to this increased distance would, upon the principle we have stated in Chapter II., lower the midwinter temperature to an enormous extent. In temperate regions the greater portion of the moisture of the air is at present precipitated in the form of rain, and the very small portion which falls as snow disappears in the course of a few weeks at most. But in the circumstances under consideration, the mean winter temperature would be lowered so much below the freezing-point that what now falls as rain during that season would then fall as snow. This is not all; the winters would then not only be colder than now, but they would also be much longer. At present the winters are nearly eight days shorter than the summers; but with the eccentricity at its superior limit and the winter solstice in aphelion, the length of the winters would exceed that of the summers by no fewer than thirty-six days. The lowering of the temperature and the lengthening of the winter would both tend to the same effect, viz., to increase the amount of snow accumulated during the winter; for, other things being equal, the larger the snow-accumulating period the greater the accumulation. I may remark, however, that the absolute quantity of heat received during winter is not affected by the decrease in the sun’s heat,[28] for the additional length of the season compensates for this decrease. As regards the absolute amount of heat received, increase of the sun’s distance and lengthening of the winter are