14 
DR. S. CHAPMAN ON THE SOLAR AND LUNAR 
a new determination of the semi-diurnal part of the solar diurnal variation potential. 
The data used were the a 2 , b 2 coefficients of the annual mean solar diurnal variations 
in the three geographical components of magnetic force from fifteen observatories 
(nine North and six South of the equator). The epoch for most of the stations was 
1901, a year of minimum solar activity; in other cases the data were corrected so 
as to correspond to such a year. In § 10 the main symmetrical annual term in the 
horizontal force potential is compared with the corresponding results from other 
investigations. The agreement in phase is good; the amplitude determined by van 
Bemmelen is a little smaller than in most of the other cases, van Bemmelen’s 
analysis also includes a strong unsymmetrical element Q 2 2 , which is somewhat 
surprising, considering that it relates to the annual mean variation. The present 
results do not seem to suggest much asymmetry between the two hemispheres. 
In attempting to separate the internal and external parts of the field, Dr. VAN 
Bemmelen remarked on the hazardous nature of the task, owing to the “ strong 
irregularities ” in the vertical force data. It appeared, as the result, that the 
internal potential for the solar semi-diurnal variation was equal to, or even slightly 
in excess of, the external potential. This differs so greatly from my own conclusion 
(and also from that of Fritsche) that I have carefully compared his and my vertical 
force data. The figures for the nine Northern stations in common were closely 
similar in the two cases, but the Southern data were far from accordant. Only 
Batavia was common to the two sets of Southern observatories; during 1901 and 
1902 the reorganization of the magnetic work at Batavia and the transfer of the 
instruments to Buitenzorg interrupted the record, and the 1889 Batavian observa¬ 
tions, corrected to 1901, were used. As the correction noted in §3 had not then 
been discovered, however, the a 2 , b 2 coefficients as used have twice their true value. 
As regards the other Southern observatories, the date of three sets used, St. Helena, 
Cape of Good Hope, and Hobarton, was 1843, and the two latter series are very 
discordant. It would seem that too much weight has been given to the six Southern 
sets of vertical force data, and that here lies the explanation of the above discrepancy 
between the two separations of the external and internal variation fields. 
§ 6. Schuster’s Second Memoir (1907). 
The theory of the diurnal magnetic variations originally propounded by Balfour 
Stewart (§ 1) was shown by Schuster, in his first memoir, to be so far correct 
in that the main part of these variations arises from electric currents circulating 
above the earth’s surface. Balfour Stewart’s theory also involved the hypothesis 
that the electromotive forces which impel these currents are supplied by the perma¬ 
nent terrestrial magnetic field acting on masses of conducting air which, in their 
bodily motion, cut through the earth’s lines of magnetic force. In this hjqeothesis 
two important factors were unspecified, viz., the atmospheric motions and the 
atmospheric conductivity. In his second memoir Schuster made definite suggestions 
