Let us take, for instance, the pressure of the air as indicated by the height of the barometer, and show that the oscillations of the mercury must be due to the sun. Imagine for a moment that the sun were to be extinguished, one consequence of the cessation of the arrival of heat at the earth would be that winds would blow no longer. There could be hardly any movements whatever in the air except such as might arise from atmospheric tides. Perhaps also I ought to add that the internal heat of the earth as manifested by earthquakes or by occasional volcanic outbreaks might produce some local and temporary disturbance of the air. It is, however, quite certain that such influences would have very slender effects on atmospheric pressure. The argument will at all events suffice to show that the fluctuations of the barometer to which we are accustomed are almost entirely attributable in one way or another to the action of the sun. It can similarly be shown that the changes in every meteorological element will depend primarily upon the great luminary. In some cases, of course, the rotation of the earth on its axis is also an important element, and to some minute degree the moon must be reckoned with. But when these influences have been i considered, we have no further concern with the heavens; it is the topographical features of the earth which complete the determination of all meteorological quantities.
I cannot here go into the discussion of a celebrated mathematical theorem which bears the name of Fourier. It seems, however, to demonstrate that any meteorological element, such as the height of the barometer or the temperature, must admit of being expressed in a somewhat similar fashion to the height of the tide. No doubt the
