This under-current is not the immediate effect of differences of temperature, but a secondary effect induced by the friction of the earth's surface and the continual deflection of the air's eastward motion from a great circle, (in which the air tends to move,) into the small circle of the latitude, in which the air actually does move. The force of this deflection, measured by the centrifugal force of the air as it circulates around the pole, retards the movement from the equator, and finally wholly suspends it; so that the upper air circulates around in the higher latitudes as water may be made to circulate in a pail; and the air is drawn away from the polar regions as this circulatory motion is communicated to it, and tends to accumulate in the middle latitudes, as the circulating water is heaped up around the sides of the pail. Hence, in the middle latitudes there is a greater weight of air than at the poles, and this tends to press the lower air to higher latitudes. Centrifugal force, however, balances this pressure, so long as the lower air moves with the velocity of the upper strata; but as the friction of the earth retards its motion and diminishes its centrifugal force, it gradually yields to the pressure of the air above it, and moves toward the poles. Near the polar circles it is again retarded by its increasing centrifugal force, and it returns through the middle regions of the atmosphere. Thus there are two systems of atmospheric circulation in each hemisphere. The principal one extends from the equator to high middle latitudes and partly overlies the other, which extends from the tropical calms to the polar circles. These two circulations move in opposite directions; like two wheels, when one communicates its motion to the other by the contact of their circumferences. In the middle latitudes the lower current of the principal circulation lies upon the upper current of the secondary circulation, and both move together toward the equator. This principal lower current first touches the earth's surface beyond the tropical calms, and having lost its relative eastward motion and now tending westward, it appears as the trade-wind, very regular and constant; while the upper secondary current returns, without reaching the tropics, as an undercurrent, and in our latitude appears as the prevailing northeastward wind,--a very feeble motion, usually lost in the weather winds and other disturbances, and only appearing distinctly in the general average.