As the sun gains altitude the action of the radial stream is at a greater and greater angle to the circular currents, and attains its maximum at noon, still acting, however, after noon; but seeing that the circular current possesses a force of re-action, that is, that the magnetism of the earth is ever striving to bring these currents to their natural direction, an hour or two after noon, the currents tend again to the equator, and the maximum deflection is passed, and finally ceases a few hours after sunset. Now let us attend to what is going on on the opposite side of the world. The radial stream passing over the polar regions, now produces a contrary effect; the ethereal atmosphere of the great magnet is accumulated on the farthest side from the sun, by the action of the radial stream passing over the polar region, the parallel currents are now bent towards the equator, being at a maximum in places where it is an hour or two past midnight. Before they were concave to the equator, and now they are convex; the magnetic meridian is therefore deflected the contrary way to what it was in the day time, by the same principle of reaction. After the maximum, say at 4 A.M., the deflection gradually ceases, and the magnetic meridian returns to its mean position at 8 or 9 A.M. These times, however, of maximum and minimum, must vary with the time of the year, or with the declination of the sun, with the position of the moon in her orbit, with the perigee of the orbit, and with the place of the ascending node; there are also minor influences which have an effect, which present instrumental means cannot render appreciable.
What says observation? The needle declines from its mean position in the whole northern hemisphere to the westward, from about 8.30 A.M., until 1.30 P.M.; it then gradually returns to its mean position by 10 A.M. After 10 P.M., it passes over to the eastward, and attains its maximum deflection about three or four hours after midnight, and is found again at its mean position about 9 A.M. Now, this is precisely the direction of the deviation of the magnetic meridian, the needle therefore only follows the meridian, or still continues to point to the temporary magnetic pole. And although we have assumed, for the sake of simplicity, that the mean magnetic pole corresponds to the pole of rotation; in truth there are two magnetic poles, neither of which correspond; yet still the general effect will be the same, although the numerical verification will be rendered more difficult.