trend inward toward the center to such an extent that the hurricane can not be considered as a system of circular rotations, but of spiral inflowing ascending and overflowing currents of air, or the ideal cyclonic vortex movement. The same conclusion was soon formed with regard to the waterspout and tornado, the only differences being as to the question what are the angles of inflow, ascent and overflow? Since 1871 a still more careful analysis of the United States daily weather maps has shown that it is necessary to consider the fact that the winds on the west side are colder and drier than those on the east side of the storm center. Thus it follows that in the northern hemisphere we have cold dry northwest winds swirling around the central low pressure and running under the more moist southerly winds, sometimes even going so fast as to overflow these for short distances while pushing them aside. In this respect a hurricane storm lies intermediate as to its mechanism between the thunder-storm and the waterspout. In the waterspout we have a relatively small mass of air, no great differences of temperature, a rapid ascension with a rapid horizontal rotation. In the thunderstorm we have a simple horizontal overturning; cooler or drier air descends from overhead and warmer or moister air ascends from below. In every style of storm and in every form of atmospheric circulation there is and must be overturning with overflow and underflow.
The problems of simple overturning have been studied by Margules from the thermo-dynamic point of view. Perhaps I can illustrate these problems by means of this glass box with vertical divisions. I will place this dividing blanket in the middle. On the left-hand side we have a mass of air cooled by this adjacent ice; on the other side is an equal volume of air at the temperature of the room. If I quickly remove the blanket the cold air settles down, flowing beneath the warm moist air and covering the bottom, while the warm air is raised to the top. The differences of density have caused an interchange of energy due to the action of the force of gravity quite independent of radiation or conduction of heat. The descending cold air has expended some of the potential energy of its position in elevating the warm air. If the warm air was anywhere near the point of saturation, then its small loss of temperature due to its rising and expanding may have produced a haze in our little experiment, but in nature it often produces a great thunder cloud, and every gradation of cloud from that form down to the thinnest stratum.
These overturnings are perpetually going on in every room of our buildings and of course in the atmosphere. If a cloudy mass descends it is warmed by compression as it descends; evaporates all its cloudy particles, and in its further descent eventually has its own temperature raised. In general we have very low temperatures at great altitudes in the atmosphere, but if such air were brought down rapidly to the