For millions of years the winds have blown over the surface of the earth, the rain has fallen on it, the sun heated it by day, the frost cracked it. Consider the winds that have circled the earth. All movements of the air are due in the first place to the sun which heats the atmosphere and causes it to expand. The sun's rays passing through the air do not heat it at once, or directly, but heat the land and the sea, which absorb some of the rays and reflect others and so warm the air in contact with them. But, as will readily be understood, the land and the sea do not absorb and reflect the heat rays in the same way or to the same extent; nor do the sun's rays fall equally or constantly on all portions of the earth's surface. So that from various causes one part of the earth is always being warmed in a different way from other parts, and the air above the earth is being warmed in an immeasurable number of different ways. Even if the earth's surface were all water or all land, we should expect therefore that there would be movements of the air due to unequal heating. If, however, the earth's surface were quite even and uniform, we should expect that there would be a certain evenness and uniformity about the movements of the air. These movements would be due partly to the regular heating and regular cooling of the surface, and partly due to the fact that the earth is spinning round taking the air with it--but not taking it quite evenly. The air does not fit tightly on to the earth. It is rather like a loose, baggy envelope with a tendency to slip as the earth moves round. Furthermore, a point situated on the Equator has much farther to travel in twenty-four hours as the earth spins round than a point situated in the Arctic Circle, where a tape measure placed along one of the parallels of latitude (let us say the eighty-sixth parallel, where Nansen turned back in his search for the Pole) would show the earth's girth there to be, not twenty-four thousand miles, but only so many hundreds. This also would make a difference in the way the air would be whirled round the earth; but we could take this point into consideration, and should be able, if, as we have said, our earth were quite uniform, to say always and at all times of the year in what direction the prevailing wind should blow.