DR. S. CHAPMAN ON THE SOLAR AND LUNAR 
for groups of days all at the same lunar phase, afterwards correcting the phase of the 
resulting Fourier coefficients to the epoch of new moon. In this way a considerable 
similarity between the relative amplitudes and phases of the various components of 
the solar and lunar magnetic variations is revealed. 
The changing phase of the non-semi-diurnal terms in the lunar variation is a result 
of the combination of a lunar semi-diurnal variation (a lunar atmospheric tide) with 
an effect depending on solar time. The latter is here identified with the variation in 
the electrical conductivity of the upper atmosphere, owing to some solar ionizing 
influence. At new moon the two effects are in phase, and the lunar magnetic 
variations resemble the solar ; in the latter case, of course, both the atmospheric 
oscillation and the variable conductivity keep time with the one body, the sun. 
Schuster found that while the main cause of the solar diurnal variations was 
external to the earth, there was also an induced system of earth currents, partly 
neutralizing the vertical force variations. This result is confirmed here, though 
with numerical modifications. The external contribution to the horizontal force 
variations is estimated at about 2'5 times the internal, as against about four in his 
memoir; also whereas no phase difference between the two current systems was 
found, a difference of from 10° to 25° is here indicated. It is shown that the internal 
field could be produced through induction by the outer currents, provided that, 
beneath an upper non-conducting layer of 150 or 200 miles depth, the substance of 
the earth has a uniform specific resistivity of amount 2‘74 . 10 12 C.G.S. A conclusion 
of this kind was arrived at in 1889 by Schuster, and.this investigation only modifies 
his in detail; he made no estimate of the resistivity of the inner core, and suggested 
1000 km. as the depth of the outer layer. 
The lunar diurnal magnetic variations are undoubtedly due solely to a semi-diurnal 
atmospheric oscillation. The relative magnitudes of the various harmonic components 
in the magnetic variation afford information regarding the conductivity of the 
atmospheric layers in which they are produced. It appears that the currents flow 
mainly in the sunlit hemisphere, the conductivity in the dark hemisphere being very 
small. Its diurnal variation can be approximately ascertained, and its maximum 
value numerically estimated ; this proves to be higher than the value originally 
suggested by Schuster. The phenomena of electric wave transmission also suggest 
the existence of such a layer of variable electric conductivity, and, in addition, of a 
still higher layer of constant and uniform conductivity. The magnetic variations 
give no indication of the latter layer. 
The main terms in the solar magnetic variation are so similar, in most respects, to 
those of the lunar variation that they too would appear to be produced mainly by a 
semi-diurnal atmospheric oscillation. The 24-hour components are relatively larger, 
however, than in the lunar magnetic variation, so that, although the phases of the 
24- and 12-hour terms show remarkably close agreement, there is ground for supposing 
that a diurnal atmospheric oscillation may be involved in addition. 
