There is another phenomenon due to the action of the radial stream, which has given much trouble to the physicist, and which has yet never been explained. This is the horary oscillations of the atmospheric pressure which, in some countries are so regular that the time of day may be ascertained by the height of the barometer. According to Humboldt, the regularity of the ebb and flow in the torrid regions of America, is undisturbed by storms or earthquake. It is supposed that the maxima occur at 9 A.M. and 10½ P.M., and the minima at 4 A.M. and 4¼ P.M. From the morning minimum to the morning maximum is, therefore, five hours; from the evening minimum to the evening maximum is 6¼ hours; from the evening maximum to the morning minimum is 5½ hours, and from the morning maximum to the evening minimum is 7¼ hours. Again, these oscillations are greatest at the equator, and diminish with the increase of latitude. [Illustration: Fig. 23] If we suppose the earth's axis perpendicular to the plane of the vortex, and P the pole in the above figure, and SP the line joining the centre of the earth and sun, M and m will represent the points in the earth's equator where it is midday and midnight respectively. The solar stream penetrates the terral vortex; and strikes the earth's atmosphere along the lines parallel to SP. The direct effect would be to pile up the atmosphere at N and n; and therefore, were the earth at rest, the maximum would be at 6 A.M. and 6 P.M., and the minimum at midday and midnight; but the earth rotating from N towards M, carries along the accumulated atmosphere, being more sluggish in its motions than the producing cause, which cause is still exercised to force it back to N. From this cause the maximum is now found at K. For a like reason the minimum at M would be found at L, but on account of the motion of the earth being now in the same direction as the solar stream, the minimum is found still more in advance at k; so that, according to the theory, the interval between the morning maximum and the evening maximum, should be greater than the interval between the evening maximum and the morning maximum; and so it is, the first being 13½ hours and the last 10½ hours. The morning minimum should also be less marked than the evening minimum, and this also is a fact. The effect also should be greater in the tropics than in high latitudes, which again also obtains; being 1.32 French lines at the equator, and only 0.18 at latitude 70°. Had the earth no obliquity, the effect would be as the squares of the cosines of the latitude; but the ratio is diminished by the inclination of the axis. But there are other variations of the barometer of longer period, apparently depending on the phases of the moon, but which cannot be reconciled to the attracting power of the moon as an atmospheric tide; and Arago concluded that they were due to some _special cause_, of which the nature and mode of action are unknown. Perhaps this theory will