least 10° Fahr. than that of Northern Asia, where the amount of snowfall is very slight, and rapidly disappears during the short arctic summer. If there be, as some persons believe, a large tract of continental land surrounding the south pole, I should expect to find that the great accumulations of snow and ice are confined to the coast regions. In that case the mean temperature of the region within the antarctic circle would probably be lower than it would be in the supposition, which appears to me more probable, that the lands hitherto seen belong to scattered mountainous islands. If, from any combination of causes, one pole of the earth has ever been brought to a mean temperature much lower than that now experienced, I should expect to find that the phenomena of glaciation would be exhibited towards the equatorial limit of the cold zone, rather than in the portions near the pole. The formation of land-ice depends on the condensation of vapour, and before air-currents could reach the centre of an area of extreme cold the contained vapour would have been condensed. This consideration alone suffices, to my mind, to make the supposition of a polar ice-cap in the highest degree improbable.

Mr. Wallace ("Island Life,” p. 142) cites, as conclusive evidence of the effect of winter in aphelion in producing glaciation, the facts, to which attention was first directed by Darwin, as to the depression of the line of perpetual snow, and the consequent extension of great glaciers, on the west coast of Southern Chili. I have adverted to this subject in the text (p. 229), and I may further remark that if winter in aphelion be the cause of the depression of the snow-line in latitude 41° S., it can scarcely fail to produce some similar effect in latitude 34° S. Yet we find on the southern limit the snow-line much lower, and at the northern much higher, than it has ever been observed in corresponding latitudes in the northern hemisphere, the line being depressed by more than 8oco feet within a distance of only seven degrees of latitude. The explanation, as I have ventured to maintain, is altogether to be found in the extraordinary rainfall of Southern Chili; and to the same cause we must attribute the fact that, in spite of the greater distance of the sun, the winter temperature is higher than in most places in corresponding latitudes in the northern hemisphere. At Ancud in Chiloe, in latitude 41° 46', the temperature of the coldest

month is lower by less than three and a half degrees of Fahrenheit than it is at Coimbra in Portugal, one and a half degree nearer the equator, in the region which receives the full warming effect of the Gulf-stream.

I should have expressed myself ill in the preceding pages if I should be supposed to deny that, in his writings on this subject, Mr. Croll has made an important contribution to the physics of geology. He has, in my humble opinion, been the first to recognize the full importance of one of the agencies which, under possible conditions, may have profoundly affected the climate of the globe during past epochs, although I do not believe that, in the present state of our knowledge, we can safely draw those positive inferences at which he has arrived. Even those who are unable to accept any portion of his theory as to the causes of past changes of climate must feel indebted to his writings for numerous valuable suggestions, and for the removal of many popular opinions which his acute criticism has shown to be untenable.