the trades and the counter trades; why, on one side of it, the prevailing direction of the wind should be polarward, on tho other towards the equator; and we also discover the influences which determine their geographical position; for:—
657. A meteorological law.—An accumulation of atmosphere over one part of the earth's surface implies a depression over some other part, precisely as the piling up of water into a wave above the sea level involves a corresponding depression below; and in meteorology it may be regarded as a general law, that the tendency of all winds on the surface is to blow from the place where the barometer is higher to the place where the barometer is lower. This meteorological law is only a restatement of the dynamical truism about water seeking its own level.
658. The barometer in the calm belts.—The mean height of the barometer in the calm belts of the tropics is greater (Plate I.) than it is in any other latitude. The mean height of the barometer in the equatorial calm belt is less than it is on any other parallel between the tropical and equatorial calm belts. The difference for the calm belt of Cancer is 0.25 inch. This difference is permanent. It is sufficient to put both systems of trade-winds in motion, and to create an indraught of air flowing perpetually towards the equatorial calm belt from the distance of two thousand miles on each side of it.
659. Winds with northing and winds with southing in them.—In like manner, as we go from either tropical calm belt towards the nearest pole, the barometric pressure becomes less and less. The meteorological law just announced requires the prevailing wind on the polar side of these calm belts to be from them and in the direction of the poles; and observations (Plate I.) show that such is the case. Dividing the winds in each hemisphere into winds with northing and winds with southing in them as has been done in Chapter XXI. and Plate XV., actual observation shows (§ 852) that they balance each other in the southern hemisphere between the parallels of 35" and 40°, and in the northern between the parallels of 25° and 60°; that between these parallels the average annual prevalence of winds with northing and of winds with southing in them is the same, the difference (Plate XV.) being so small as to be apparently accidental; that, proceeding from the medial band towards the pole, polar-bound winds become more and more prevalent, and proceeding from it towards the equator, equatorial-bound winds become more and more preva-
