Permanent winds turn the vane only in a direct sense, or with the sun :

Cyclones or centripetal winds, when they have a progressive motion, produce a rotation which is direct or retrograde, according to the position of the point of observation with regard to the path of the centre:

Finally, by the interference of two permanent currents, whose directions are different, there ensue rotations of both kinds, viz., direct, when, on the west side of the wind-rose, a current is displaced by one which is more northerly; retrograde, when it is displaced by one more southerly. On the east side, on the contrary, the rotation is direct when the displacing current is more southerly than the previous one; retrograde, when it is more northerly. Currents diametrically opposed can check each other and cause a calm, producing the same phenomenon as is observed at the centre of a series of centripetal currents or of a cyclone; viz., winds diametrically opposed, separated by a calm district.

We see, therefore, that the same appearances can arise under totally different conditions, and that it is only by means of a careful investigation of all the phases of a phenomenon that we can pronounce a decision on any one that we have to do with in a particular instance. In this investigation the barometer is one of the chief means at our disposal, and we shall therefore consider its indications with especial attention.

I. Previous to entering on the discussion of the storms themselves, we shall examine more closely the motions of the atmosphere in the different zones.

We find, as has been stated above, three types of these motions:-

but seamen retain the sense of 'with the sun' as (applicable to the coiling or lay of a rope) from left to right. Hence some confusion.-F.

1. PERMANENT WINDS: The Trade-winds of the Torrid Zone.

2. PERIODICAL WINDS: The Monsoons of the Indian Ocean.

3. CHANGEABLE WINDS: The winds of the Temperate and Frigid Zones.

C

18

I. The Permanent Winds.-The Trade Winds.

I. THE UNDER TRADE-WIND.

If any point in a liquid be heated more strongly than the others, currents arise in it, and the colder particles flow from all sides towards this heated point. At this point itself the fluid is at rest, owing to the compensation of the opposite motions. The most obvious illustration of this is offered by the flame of a candle when it is burning steadily.

Were the earth at rest, and the sun fixed over a point on the equator, the air from all parts of the world would flow towards this point, and we should find the conditions of the flame of the candle fulfilled. The earth, however, moves on its axis, and consequently, instead of one heated point, we have a zone of maximum temperature on which the air is at rest. This zone separates the areas of the northern and southern currents of cold air, each of which is thus forced to perform its own circuit.

If we assume that the mean breadth of this zone is invariable throughout the year, and that it moves with the sun through 23 on each side of the equator, we shall find it in June under the tropic of Cancer, in January under that of Capricorn. Consequently, all places in the torrid zone will be for a part of the year in the northern Trade-wind, and for a part in the southern. These periods will be separated by intervals at which there will be no

particular direction of the wind. If the air be supposed to flow towards the parallel of latitude which has the sun in the zenith at the time, the wind will flow from directly opposite quarters, within the yearly period, at all points over which the sun passes twice in the year. The duration of the two currents will be the same at the equator. In the northern hemisphere the northern current will last longer, and vice versâ in the southern hemisphere. The difference in the durations will increase with the distance from the equator. At the tropics there will be only one current interrupted by a calm at the time of solstice. We should therefore obtain at all points in the torrid zone 'Monsoons' or winds depending on the seasons. The rotation of the earth would produce an effect on these, and we should have the following winds at the different seasons, on the respective sides of the equator :

In the preceding investigation we have omitted to take notice of the fact that the magnitude of the masses of air, which are separated by the region of calms, varies very much at different seasons. At the winter solstice the air of the torrid zone is included in the northern Trade-wind circuit, and at the summer solstice in the southern circuit; so that it is only at the equinoxes that the quantities of air in each of these circuits are equal. As the relation between these masses varies with the seasons, it will necessarily interfere with the regular motion of the zone of

Trade-winds northwards and southwards. Another point which it is necessary to remember is, that the vertical Trade-wind circuit does not take place in an area which is uniform in width, but in one which tapers towards the pole. We may consider the space between two meridians as a section of an isosceles triangular prism, whose base has a higher temperature than any other part. Consequently the air which rises at the base will not descend at the vertex, but at some parallel to the base line. From this it is evident that the phenomenon of the Trade-winds will not embrace the whole of a hemisphere, but only a portion of it which is adjacent to the equator, and that the Trade-winds will have interior and exterior limits. It appears, further, that the mean position of the place of meeting of the two currents will not coincide with the equator, but will be thrown up into the northern hemisphere. The reason of this is, that the unequal distribution of dry land on the surface of the earth causes the northern hemisphere to have a higher mean temperature than the southern. The displacement of the zone of Tradewinds at the different seasons, will be influenced more by the conditions of temperature in the torrid zone than by those of the whole earth; inasmuch as the air in districts lying at a distance from the equator is not included in the Trade-wind circuit.

We shall now proceed to examine the observations, and draw our conclusions from them.

In the Distribution of Heat on the Surface of the Earth, I have given the following tables of the mean temperature of the northern portion of the torrid zone for each month in the year.

*Die Verbreitung der Wärme auf der Oberfläche der Erde, &c. Reimer, Berlin, 1852. 4to.