DIURNAL VARIATIONS OF TERRESTRIAL MAGNETISM 
59 
From Table J it appears that the amplitudes of the three components Q ;j 2 , Q,/ j , QA 
are smaller at the solstices than at the equinoxes, while Q 2 * is approximately constant. 
If the lunar diurnal atmospheric oscillation resembles the corresponding solar diurnal 
variation, the amplitude is greater at the equinoxes than at the solstices; this 
probably explains most of the variation in Qy ; our calculation just made would 
appear to indicate that, so far as the electrical conductivity is concerned, Q 3 2 is 
constant throughout the year. Besides the variation in the amplitude of the 
atmospheric motion QA, which should affect equally all the magnetic variations Q/ 
to QA, it would appear that the non-semi-diurnal components should be further 
reduced at the solstices, because of the conductivity effect. Here the theory and 
observation are only in very rough agreement, since Q 2 4 shows a relative increase at 
the solstices, while the observed decreases of Q 4 3 and Q 5 4 are much more than the 
theory would predict. 
Similar discrepancies are met with when we examine the “ seasonal magnetic 
components Q„“. These are represented in Table 0 by the terms involving odd 
powers of sin S, which vanish at the equinoxes and change sign between summer and 
winter. The following values of their relative amplitudes are calculated in the same 
way as the numbers in Table Q :— 
Qi 1 . Qr 1 . QA Qr 2 . QA Qr 3 - QA Qr 4 . 
(76) —4'2 -24’3 47 0 0’46 0 U‘0012 0 
The remarkable feature here is the excess of Qr 1 over Qf. The amplitude in 
Table J is that of Q,! In any case, however, the above numbers, which should 
roughly agree with those given in Table J for the solstitial inequality (since the 
numbers in Table Q are approximately equal to the equinoctial data), are much 
smaller than the observed values. 
Before considering this point further, we shall turn to the solar diurnal magnetic 
variation data of Table F. In this case we have no such certain means as before of 
determining the period or periods of the atmospheric oscillations to which the 
magnetic variations are due. We may notice, however, the similarity of phase (e/) 
between the four solar harmonics Q ,l )!+1 in Table F, which is at least as close as 
that shown in the lunar Table J. So far as this goes, there is a strong suggestion 
that the solar diurnal magnetic variations also arise from a single atmospheric 
oscillation. 
For comparison of the amplitudes of the lunar and solar magnetic variations 
reference will be made to either the potentials derived from the horizontal force 
variations above (Tables C, G) or to those of the external fields Eff* (Tables F, J). 
The discussion in Part IV. has shown that the potentials from the vertical force 
variations lead to fairly similar sub-divisions of A,/, B m ” between the external and 
internal fields. 
i 2 
