EARTH’S MAGNETISM AT PAVLOVSK AND POLA (1897-1903). 
297 
variations originate in currents flowing mainly above the earth’s surface,* * * § and has 
shown that these may very probably be ascribed to the effect of electromotive forces 
produced by the periodic motions of the atmosphere across the earth’s permanent 
magnetic field, the electrical conductivity necessary for the passage of the currents 
being found in the rarified upper atmosphere, which is assumed to be ionized by some 
solar action such as ultraviolet radiation, t This theory is so valuable and successful in 
many ways that it is desirable to examine carefully the points not yet elucidated by 
it, among which the chief are the ratios of the amplitudes of the several harmonic 
components in the magnetic variation (as compared with the ratios of the 
corresponding terms in the barometric variation), and the seasonal changes in these 
components. If the natural assumption that the mechanism of the lunar and solar 
diurnal magnetic variations is similar is correct, an investigation of the former should 
throw much light on the latter; since the lunar diurnal variation arises from an 
atmospheric oscillation of a single frequency there is good reason to suppose that the 
relative amplitudes of the various components of the lunar magnetic variation 
(computed for the separate lunar phases and reduced to the time of new moon, say) 
will enable the daily variation of the atmospheric electrical conductivity to be 
determined. 
The chief point of novelty and importance in the investigation for which the data 
here presented will be employed lies in the use of the harmonic components of the 
lunar magnetic variation of frequencies other than 2, as well as the main semi-diurnal 
component. In particular, as is explained in the memoir already referred to,J the 
fourth harmonic term has a special significance for our purpose. 
The Method of Computation. 
§ 3. The following is the method of computing the lunar diurnal magnetic 
variation from the published hourly values of the magnetic elements which, after 
much consideration, has appeared to me to be the simplest and best. It resembles 
that used by J. A. Broun in his reduction of the Trevandrum observations of 
declination,§ more than any other method hitherto used. 
The monthly mean solar diurnal variation, as ordinarily calculated from a calendar 
month’s hourly observations, may be considered free from any lunar term, since in this 
period|| the lunar disturbance characteristic of any lunar hour affects in turn each 
* Schuster, ‘Phil. Trans.,’ A, vol. 180, p. 467. 
t Schuster, ‘Phil. Trans.,’ A, vol. 208, p. 163. 
J ‘Phil. Trans.,’ A, vol. 213, p. 306. 
§ ‘Trevandrum Magnetical Observations,’ vol. I., p. 113, 1874. 
|| More strictly, in an exact lunation, but the two are so nearly equal that in view of the relative 
smallness of the lunar variation the difference may be neglected. 
2 Q 
VOL. CCXIV.-A. 
