THEORIES OF MAGNETIC IRREGULARITIES AND VARIATION Magnetic Irregularities.—The difficulties of accounting for the larger irregularities of the earth's field are bound up with our lack of knowledge of the composition of the earth's crust to any considerable depth. Local irregularities can in most cases be assigned to local deposits of magnetic material. An ingenious at tempt to ascertain the general nature of the conditions necessary to provide for the main features of the earth's magnetic abnor malities was made by Wilde"), who constructed a "magnetarium" designed to reproduce the main features of the earth's field and also its secular variation. Wilde's magnetarium consisted of a globe 18 inches in diameter provided with plates of iron attached to its under side beneath the oceans, and in one or two other locations. Within the sphere, and fastened to it, was a spherical shell of wire gauze supporting a current winding with its planes parallel to the equator. Within the gauze shell was mounted a smaller sphere which could be made to revolve about an axis inclined at 23.5° to the axis of the first sphere, and which was wound with wire in planes perpendicular to its axis of rotation. Wilde's views have not passed without criticism. However, a consideration of the depth to which the probable temperatures conditions within the earth would permit iron to exist in magnetizable form, corn bined with the effect of the ocean areas in modifying the tempera ture gradient, and the requirements imposed by the hypothesis of isostasy lend support to reasonableness of the condition sug gested by the magnetarium, as regards their application to the larger irregularities"). The part of Wilde's conclusions having to do with currents circulating within the earth, and with the secular variation must necessarily remain more speculative.
Theory of the Diurnal Variation.—The fundamental idea underlying the cause of the diurnal variation is undoubtedly that originated by Balfour Stewart. According to this view, the motion of the earth's upper atmosphere under the influences of the tides sets up electromotive forces on account of the cutting of the earth lines of force by the moving air. These electromotive forces set up currents in the atmosphere in regions where the conduc tivity is sufficient and these currents, varying as they must with the tides, produce magnetic fields which also partake of the tidal periods. In addition to the direct effect of the atmospheric cur rents, induced in the above manner, there is a secondary effect caused by the induced currents which the time rates of change of the atmospheric currents set up in the substance of the earth.
According to S. these currents induced in the earth contribute about 28 per cent of the total diurnal variation.
The two bodies chiefly concerned in producing the atmospheric motions which are the origin of the diurnal variation are the sun and the moon. If we first confine our attention to the variation of the magnetic elements throughout the lunar day, averaging the results for any (lunar) time of day throughout one or more months, and if we plot the results so obtained against the time, we find that the curves are symmetrical in horizontal intensity north and south of the equator, and anti-symmetrical in declina tion and (outward) vertical intensity. All the curves are purely semi-diurnal in type. These features are accounted for by the assumption that the atmospheric motions are of the lunar tidal type. The barometer indicates semi-diurnal variations in at mospheric pressure indicating the existence of an atmospheric lunar tide, and this must involve a circulation of the air, the flow being mainly horizontal"). It is consequently the earth's vertical component of magnetic intensity which is mainly con cerned in the production of the electric currents in the conducting regions of the atmosphere. An element of lack of harmony be tween theory and observations appears, however, in a numerical comparison of the phase of the magnetic diurnal variation and that of the oscillations in barometric If, instead of plotting the means for one or more lunar months, we plot the data for the individual lunar days, the curves obtained are more complicated than those constructed from the monthly means. In a general way the complication is in the direction of a magnification of the magnetic variations during the daytime. On plotting the magnetic elements throughout the solar day, using for this purpose a number of the quietest days for the month, we obtain curves which have many features in common with those obtained by plotting against time for the lunar day, but the ampli tudes are about eleven times those for the lunar day. This is in harmony with the expectations arising from the barometric varia tion, which, for observations plotted throughout the solar day, indicate amplitudes of tidal motion fifteen times those obtained by plotting the .monthly averages throughout the lunar day. It is probable that the atmospheric motions associated with the sun have their origin in temperature differences rather than in ordinary tidal forces; for, if this were not the case, the lunar variation of barometric pressure should be greater than the solar variation, since the purely tidal influence of the moon is about twice that of the sun.