these lines cut the earth's surface. The dotted continuations represent how the lines would run if the atmosphere took the place of the solid. earth. The curved lines starting from B and C to E and F, denote the lines as they occur on the earth's surface or appear on a plane chart, when the contours curve round a central area A, where

In

the pressure in this case is about 29.7 inches. considering the general circulation, A may be taken to represent the North or South Pole, in which case the diagram represents something like what actually takes place. When we are dealing with particular motions, A would correspond with the centre of a cyclone or travelling disturbance.

Returning to our story, it is plain from these maps, that the circulation of the atmosphere comprises a great deal more than a mere system of trade winds, blowing towards the equator.

Any theory that pretended to explain the entire system, would have to account for the prevailing high pressures about latitude 30° N. and S., and the poleward trend of the wind on the polar sides of these atmospheric sierras.

Ferrel, in his first paper in 1856, not only shewed that Hadley's theory was mathematically

incorrect, but that Maury's fascinating scheme, put forward in his Physical Geography of the Sea, erred both in fact and the laws of physics.

Reversing the usual procedure by which laws are induced from observations and starting with a few fundamental principles, such as the law of deflection already noticed, he shewed that the high pressure belt about 30° and the system of poleward winds on the polar sides of it were necessary consequences of these principles.

Ferrel's final ideal chart of atmospheric circulation on the earth is represented by fig. (18)

increases what would otherwise be an insignificant where the average motions near the surface are represented in plan in the shaded circle, and in vertical section at the border, and though his explanation is a complicated piece of mathematical reasoning, the following represents the pith of it in simple language.

Assuming that the air over the equatorial zone is heated above that near the pole, it expands near the equator and contracts over the pole. Consequently the uplifted air over the equator tends to flow downhill as it were towards the poles and a corresponding flow near the surface takes place towards the equator. If the earth did not rotate on its axis the upper air would flow direct to the pole then descend and return towards the equator along the surface. Since however the earth does rotate, the upper air is deflected in the northern hemisphere to the right (increasingly as it travels polewards) so much that were it not for the downward slope towards the pole, it would eventually deviate towards the equator. Consequently the pressure to the left of its path, i. e. towards the pole, is decreased and the pressure to the right is increased. This slope to what is actually observed. By the time this upper air has reached latitude 30° which divides the hemisphere into two equal areas* (though it is only a third of the actual distance on a meridian) it has descended to the surface and overpowers any tendency towards contrary motion in the air there, and the entire atmos

* This most important fact is one of those things which is not as a rule taught at school, though it is of immense signifi

cance.

phere tends to move bodily eastwards from thence to the pole.

Meanwhile the air near the surface between latitude 30° and the equator, moving towards the latter, deviates towards the west and heaps up pressure to its right and lowers the pressure to its left in the same way. Consequently on all accounts there is a tendency on the part of the air to heap up and increase in pressure about latitude 30° and to be reduced in density or pressure near the poles and the equator. Also since the air reaches a terminus at the poles and equator there will be calms at the surface at both these points. Moreover since on either side of latitude 30° or more correctly 35° the air moves along the surface in contrary directions, there will be an absence of prevailing winds over this region. These calms are known to exist, and owing to their proximity to the tropics used to be called the calms of Cancer and Capricorn.

The preceding explanation may be better realised if we take a vertical section of the atmosphere along a meridian as it actually exists, and draw sections of the general planes of equal barometrical pressure as they exist by observation on the average at different levels between the equator and the poles as in fig. (19). The poles are at N. and S. and the equator is at Q.

The line A B C D E F represents a section of the general isobar of 30 inches at sea-level and shews two cols or hills at latitude 30°. Then as we ascend these cols gradually disappear owing to the absence of the surface trades and therefore of the side pressure they create which keeps the opposite pressures exerted by the winds on their polar sides from pressing the air into one high

pressure belt over the equator. As we ascend, therefore, these cols gradually coalesce into one

central hill from which the air descends as a westerly upper current (blowing partly from the south in the northern hemisphere and from the north in the southern), into the two polar valleys on either side.

The depth of the atmospheric valleys of pressure below the top level of pressure at the surface of the earth converted into feet of air is found to be about 262 feet at the equator, about 320 feet at the north pole, and 640 feet at the south pole.

At a height of 30,000 feet above the surface the north polar valley is 2,800 feet and the south polar valley 3,100 feet below the mean level. The height at which the equatorial valley disappears varies from 8,000 to 12,000 feet. Above this level there is a downward slope all the way to the arctic and antarctic circles and possibly to the poles themselves.