[22] Comptes-Rendus, July 9, 1838. Taylor’s “Scientific Memoirs,” vol. iv., p. 44 (1846).
[23] The mean temperature of the Atlantic between the tropics and the arctic circle, according to Admiral Fitzroy’s chart, is about 60°. But he assigns far too high a temperature for latitudes above 50°. It is probable that 56° is not far from the truth.
[24] The probable physical cause of this will be considered in the Appendix.
[25] The mean temperature of the equator, according to Dove, is 79°·7, and that of the north pole 2°·3. But as there is, of course, some uncertainty regarding the actual mean temperature of the poles, we may take the difference in round numbers at 80°.
[26] Trans. of Roy. Soc. Edin., vol. xxii., p. 75.
[27] Connaissance des Temps for 1863 (Additions). Lagrange’s determination makes the superior limit 0·07641 (Memoirs of the Berlin Academy for 1782, p. 273). Recently the laborious task of re-investigating the whole subject of the secular variations of the elements of the planetary orbits was undertaken by Mr. Stockwell, of the United States. He has taken into account the disturbing influence of the planet Neptune, the existence of which was not known when Leverrier’s computations were made; and he finds that the eccentricity of the earth’s orbit will always be included within the limits of 0 and 0·0693888. Mr. Stockwell’s elaborate Memoir, extending over no fewer than two hundred pages, will be found in the eighteenth volume of the “Smithsonian Contributions to Knowledge.”
[28] When the eccentricity is at its superior limit, the absolute quantity of heat received by the earth during the year is, however, about one three-hundredth part greater than at present. But this does not affect the question at issue.
[29] Scoresby’s “Arctic Regions,” vol. ii., p. 379. Daniell’s “Meteorology,” vol. ii., p. 123.
[30] Tyndall, “On Heat,” article 364.
[31] Tyndall, “On Heat,” article 364.