Viewed as a whole, the general circulation of the air according to Ferrel, may be considered as consisting of two huge atmospheric whirls, or, as they are technically termed cyclones, with the poles as centres, in which the air rotates in each hemisphere in the direction of axial rotation. On the equatorial side of each of these whirls, a belt occurs in which the motion of the air is contrary to that of axial rotation. These are the trade wind belts. Between these two areas the air is heaped up into two zones of high pressure, reaching their highest values in the northern hemisphere about latitude 35° and in the southern about latitude 30°.

Since this system was established by Ferrel, Dr. Werner Siemens, von Helmholtz, Herr Möller, Professor Oberbeck, Dr. Sprung and others have developed the theory by the aid of more modern refined methods and closer reasoning, but their conclusions are substantially the same as those deduced by Ferrel, and the above sketch represents as far as can be attempted in a work like the present the modern theory of the general circulation of the atmosphere.

The most noticeable permanent modification from the ideal condition of things is afforded by the exceedingly low pressure round the south pole and the strong north-west winds which prevail south of the Atlantic, Indian, and Pacific Oceans, and which enabled Australian clippers in the old days to make passages of fabulous rapidity. This is due to the fact that the southern hemisphere is chiefly covered by water which, by exerting less friction on the air than land, allows its motions to occur with greater freedom. In consequence of this the Antarctic barometric depression is

more developed and more symmetrical than the northern. For example the pressure on the latitude corresponding to our Antipodes is permanentlyths of an inch below what we experience, while the wind velocity is three or four times as great.

The seasonal changes and migrations as the sun moves north and south are scarcely noticeable in the southern hemisphere for the same reason. In July the pressure over the tropical * belt as we may term it, is slightly increased, and the belt lies a little further south than in January. In the northern hemisphere on the other hand, the seasonal changes are far more conspicuous.

The high pressure nuclei which in July lie on the eastern sides of the Pacific and Atlantic oceans have by January shifted on to the continents of America and Asia, while the low pressures, which in July occupied the middle parts of North America and the centre of Africo-Asiatic continent, (the centre lying almost exactly over the Persian Gulf which is the geographical land centre) in January lie over the North Pacific and North Atlantic. Meanwhile the equatorial low pressure belt, or barometric equator as it may be termed, which in January is confined between its ideal equatorial limits, in July runs up into the northern continents and in Africo-Asia in particular, may be said to lie entirely over the land surface, where it causes the Monsoon as in figure (20). The relation of these extensive migrations to the effect of seasonal changes in solar heat on

*The term tropical is here used to signify on or near the tropics or turning points and not to the entire space between them as is usually the case.

the air lying over land and water surfaces is obvious.

The result of these large transfers of air and air pressure north and south, and between the oceans and the continents, is to cause what is termed the annual variation of the barometric pressure at any single place on the earth. This annual variation reaching its extremes generally

in January and July, will be found to be largest in the centres of continents such as Asia where the barometric shift is greatest and least on the coasts which are the axes of the annual

pressure sea-saw between the oceans and the continents.

In England the change is small, amounting to about 0.12 inches between January and July.

In India, it increases from 0.26 inches in Bengal to 0.62 inches in the Punjab, while over Siberia and central Asia it reaches about 1 inch.

The mean pressure over the whole earth is 29.89 inches. In the northern hemisphere the mean pressure for January is 29.99, and for July 29.87. For the southern hemisphere the pressures in the same months are 29.79 and 29.91. From this it is evident that there is a much greater difference between the quantity of air over the two hemispheres in the northern winter in January than in the southern winter in July.

This difference in favour of the northern hemisphere really means that owing to the greater cooling and contraction over the northern land area in the winter 32,000,000 tons of air have shifted over to supply the defect. There is no protective tariff placed upon this valuable import from the southern hemisphere.

A seasonal shift of the general wind system of the lower strata occurs like that in the barometric pressures, as the sun shifts north and south.

The shift of the sun in latitude is 47°, but the wind systems only shift from 5° to 8° on the northern, and from 3° to 4° on the southern side of the equator. The accompanying diagram fig. (21) gives an idea of the effect of the shift.

The cen

tral belt (sub-equatorial) represents the district. which is alternately subject to the doldrums or equatorial calms, formerly the bane of sailors, and the attendant bordering trades as they oscillate north and south.

The width of this belt varies from 350 miles in the Atlantic to 200 in the Pacific and lies on the north side of the equator all through the year, owing to the fact that the system of circulation in the southern hemisphere and round the South

pole, owing to smaller friction, is so much more powerful than that in the northern that the pressure belts on that side are all pushed northwards.

The sub-tropical belts, as the calms of Cancer and Capricorn are now termed, are similarly the alternate arena of westerlies, trades, and intervening calms.

So far, we have mostly considered the general circulation with respect to the motion of the atmosphere, near the surface. The Upper current