296 
DR. S. CHAPMAN ON THE LUNAR DIURNAL VARIATION OF THE 
and TrevandrumA A few of its main points are summarized below for the better 
understanding of the object of the present paper. 
§ 2. The magnetic elements undergo regular variations with the period of a lunar 
day, just as they do in the course of a solar day ; the latter variations are much the 
larger of the two. The lunar diurnal changes are the simpler in character, however, 
taking the form, when averaged over a whole lunation, of a purely semi-diurnal wave. 
But if the variation is computed from a number of days all at the same lunar phase 
(i.e., when there is a definite angular distance between the sun and moon), the solar 
diurnal variation having first been eliminated from the observations, harmonic 
components of other frequencies in the lunar day are also found to be present. The 
epoch of the component of frequency 2 (the semi-diurnal term) remains constant 
throughout the lunation, but the epochs of the terms of frequency 1, (2), 3, 4, ..., n 
change regularly during the lunation by amounts —2w, (0), 27r, 4?r, ..., 2 (n— 2) w 
respectively. Moreover, the epochs of the various components generally seem to be 
nearly or quite the same at the time of new moon, when the sun and moon are on the 
same meridian. This law of phase change was discovered and verified by a study of 
the first four harmonic components ; it explains the disappearance of all components of 
frequency other than 2 from the variation computed from a whole lunation. The 
phenomenon clearly suggests a solar action modifying a regular semi-diurnal lunar 
variation. In the paper referred to, I have shown that the law according to which 
the epochs of the several harmonic components change during a lunation is in accord¬ 
ance with the theoretical consequences of such a solar action. While the proof of 
this theorem might easily be stated in terms which imply no special hypothesis as to 
the manner in which the action takes place, it was actually given in terms of a parti¬ 
cular theory, viz., that the effect is due to a variation, periodic in a solar day, of the 
electrical conductivity of the medium in which flow the currents to whose magnetic 
IDotential is attributed the variable field put in evidence by the observations. The 
dependence of the electrical conductivity upon solar time is accounted for if we 
suppose that all or part of this conductivity is due to some ionizing influence from the 
sun. So far as regards the facts thus mentioned, this assumption fits in naturally 
with a view that the lunar magnetic variations arise from the lunar tide in the earth’s 
atmosphere! in the same way as, according to Schuster’s well-known theory, the solar 
diurnal magnetic changes arise from the ordinary daily atmospheric motions^ which 
are indicated by the barometer. 
In two important memoirs Schuster has proved that the solar-diurnal magnetic 
* ‘Phil. Trans.,’A, vol. 213, pp. 279-321. This paper contains fairly full references to the previous 
work on the subject. 
f This has been detected and measured in the barometric records of some Tropical observatories, i.e., 
St. Helena and Batavia. It is, as was to be expected, purely semi-diurnal. Sabine, ‘Phil. Trans., 1847 : 
‘Batavian Observations,’ 28, 1905. 
j Mainly of solar-thermal rigin. 
