seasonal causes much more slowly than the lower air, and a mile above the earth the daily change of temperature, apart from the passage of "warm and cold waves," is less than one degree. At a height of six miles above the earth a temperature much below zero constantly prevails, while, at ten miles, 80∞ below zero has been recorded in a balloon--this is approximately the temperature prevailing winter and summer above pole and equator. These facts are expressed graphically in Plate III., Temperature at Different Latitudes and Altitudes, which represents half of a section of the earth, from the north pole to the equator, with the superincumbent atmosphere. [Illustration: PLATE III.--TEMPERATURE AT DIFFERENT LATITUDES AND ALTITUDES.] Perhaps it should be explained, that whereas the curvature of the earth with respect to the height of the atmosphere in the previous diagram was not exaggerated, in the present diagram the height of the atmosphere over the radius of the earth is enormously increased. At the north pole the mean annual temperature is about 0∞ Fahrenheit, and at the equator it is about 80∞. It is seen that the atmospheric layer having a temperature of 50∞ (here represented in section by a line) touches the earth at 45∞ latitude, but is about two miles above the equator. In the same way the line of freezing (32∞) leaves the earth's surface at 58∞ latitude and rises to about three and a half miles over the equator; the line of 0∞ rises from the pole to about seven miles at the equator. This is familiarly illustrated by the fact that only the highest mountains in the tropics are snow-capped, while within the Arctic circle the snow-line descends nearly to sea-level. The lines in the diagram show the mean annual temperatures, but the isothermal surfaces rise in summer and sink in winter, the change of altitude being greatest in northern regions and near the ground. Frequently there is an inversion of temperature, that is to say, it is warmer above than below. Notably, in Siberia, where the winter temperature is 60∞ below zero, there can be no immediate decrease of temperature with height, and it is probable that there is a warmer layer of air interposed between the very cold earth and the still colder upper air, so that the temperature first rises rapidly with elevation and then falls slowly to the limits of the atmosphere. In temperate latitudes it often happens, with a high barometric pressure, in winter that the mountain stations enjoy a long period of still and relatively warm weather, as compared to that experienced in the valleys. But the subject of inversions of temperature will be discussed at length in considering the results of the balloon and kite observations.