As an illustration of the scale of the temperature variations which may be due to geographical changes, suppose that the belt between 50° and 70° N. is entirely above the sea. Then we have the following theoretical temperatures; for a point on 60° N. at sea-level: January -30° F.; July 72° F. Data for calculating the effect of ice are rather scanty, but the following probable figures can be given, supposing that the belt in question were entirely ice-covered: January -30° F. (as for land); July 23° F. Supposing that the belt were entirely oceanic, the mean temperature in 60° N. would be: January 29° F.; July 41° F. These figures show how enormously effective the land and sea distribution really is. From Appendix it is easy to calculate the probable temperature distribution resulting from any arrangement of land and water masses. Since the geography of the more recent geological periods is now known in some detail, we have thus a means of restoring past climates and discussing the distribution of animals and plants in the light of this knowledge. Of course it is not pretended that no other possible causes of great climatic variation exist, but no others capable of seriously modifying temperature over long periods are known to have been in operation. As we shall see later, there are solar and other astronomical causes capable of modifying climate slightly for a few decades or even centuries, but these are insignificant compared with the mighty fluctuations of geological time. In applying the results of this “continentality” study to former geological periods the method adopted is that of differences. The present climate is taken as a standard, and the temperatures of, for instance, the Glacial period are calculated by adding to or subtracting from the present temperatures amounts calculated from the change in the land and sea distribution. This has the advantage of conserving the present local peculiarities, such as those due to the presence of the Gulf Drift, but such a procedure would be inapplicable for a totally different land and sea distribution, such as prevailed during the Carboniferous period. That it is applicable for the Quaternary is perhaps best shown by the following comparison of temperatures calculated from the distribution of land, sea and ice with the actual temperatures of the Ice Age as estimated by various authorities (inferred fall):