16 
DR. S. CHAPMAN ON THE SOLAR AND LUNAR 
that assuming the variation of conductivity to follow the law 1 +cos a>, # the ratio 2'6 
would be increased to 47 without drawing at all on Qg 1 . 
As regards the seasonal change in the magnetic variations, it was stated that the 
large increase in summer could not be explained completely by the above variation of 
conductivity, and a cumulative seasonal change was suggested as a possibility, in 
addition to the variation with «. The weight of this difficulty, however, was chiefly 
thrown upon the uncertainties in the atmospheric motions. In this paper the problem 
is simplified by the evidence afforded by the lunar diurnal variations, which indicate 
how largely a semi-diurnal oscillation is able to account for the 24-hour magnetic 
variations, owing to a much more marked variation of conductivity, between day and 
night, than that represented by the formula 1 + cos «. 
Schuster estimated the order of the electric conductivity required by the theory, 
and discussed how far the high value thus found was physically possible or probable. 
He concluded that it was a possible value, which might perhaps be accounted for by 
ascribing the conductivity to the ionizing action of ultra-violet radiation from the 
sun. But it was remarked that the absorption of such radiation in the solar 
atmosphere might render this suggestion invalid. 
The theoretical calculations of the paper dealt mainly with that part of the 
permanent magnetic field of the earth which is symmetrical about the geographical 
axis. It was pointed out, however, that the obliquity of the magnetic axis should 
result in the production of magnetic variations not depending solely on local time, and 
a search for these terms was suggested as a promising line of further work. 
Part II.— A New Analysis of the Solar Diurnal Magnetic Variation. 
§ 7. Description of the Data. 
The data used in this investigation consist of the Fourier coefficients a n , h n in the 
harmonic formula 
( 1 ) 2 ( a n cos nt + h n sin nt) 
for the solar diurnal variations in the North, West and vertical components of 
magnetic force. Results from twenty-one observatories are utilised, of which fifteen 
are North and six South of the equator, between latitudes ±61 degrees. The average 
number of stations represented in any particular section of the final results is 
slightly less than twenty, however, since data in every element were not available 
from quite all the selected observatories at the chosen epochs. 
The stations were selected so as to obtain as wide a distribution in longitude as 
the available records allowed ( cf . § 27). Particulars of their names and positions are 
given in Table A. For convenience in the subsequent numerical analysis they have 
been divided into nine groups, as indicated. 
* According to this law the conductivity evidently varies from a maximum at the point directly beneath 
the sun to zero at the antipodal point. 
