With Dr. Croll's connection with this subject a new era may be said to have been inaugurated. Previously it had been deemed sufficient to point out certain agencies which seemed to be adequate to produce the observed effect, without making any effort to show mathematically that they were adequate; but this physicist contended from the first that cause and effect should be determined in absolute measure, just as in the other branches of physical science. His failure to observe this excellent rule in one case is to be attributed to the same paucity of trustworthy observations which is the occasion of the obscurity enveloping the whole subject.
In Chapter IV. the physical agencies leading to changes of climate are discussed, and an explanation of the present low temperature of the southern hemisphere is offered in Chapter V. As has already been intimated, that hemisphere, which has its winters in aphelion, has a longer winter and a shorter summer than the mean. Now, it is perfectly obvious that this variation in the length of the seasons increases with any increase in the eccentricity of the terrestrial orbit, and similarly diminishes with any diminution of eccentricity; for the eccentricity of the planetary orbits may vary within pretty "wide limits, which have been determined by La Grange, Leverrier, and, more recently as well as more satisfactorily, by Mr. J. N. Stockwell, of Cleveland. The present eccentricity (0·0168) is such that there is a difference of about eight days in the length of summer and winter in either hemisphere, when the solstices coincide with the apsides; and, the winter solstice of the southern hemisphere being now not far from aphelion, that hemisphere has the long winter and short summer. It is admitted, however (even too readily, it would seem), that the present degree of eccentricity is too insignificant to exercise much influence on the climate of the globe; but 210,000 years ago, when, according to Dr. Croll's elaborate computations, the eccentricity was 0·0575, the excess of winter over summer, due to this cause, amounted to 26·7 days; and 850,000 years ago, the eccentricity then being 0·0747, it amounted to 34·7 days; and it is argued that so great a difference in the relative length of the seasons would indirectly, through the intervention of a number of physical agencies, materially affect the earth's climate. These agencies are shown to be such as would be brought into operation by an increase in the length of winter, even if its severity was not increased; and they are mainly dependent on the increased proportion of moisture precipitated as snow instead of rain. Of course, this snow would remain until melted by approaching summer, just as it does in every region where much snow falls.
"There are three separate ways whereby accumulated masses of snow and ice tend to lower the summer temperature": 1. By means of direct radiation [and by direct contact]; 2. By direct reflection back into space of the solar rays; and, 3. By chilling the air and condensing the vapor into thick fogs which intercept the solar rays. This
