DIURNAL VARIATIONS OF TERRESTRIAL MAGNETISM. 
3 
earth’s surface. The result demonstrated the accuracy of Balfour Stewart’s 
conclusion that the origin must be external. 
At the close of his first paper Schuster suggested that the convective atmospheric 
motions indicated by the diurnal barometric changes are those which are responsible, 
in the manner proposed by the above theory, for the daily magnetic changes. In his 
second memoir this hypothesis was carefully examined, using the data of his former 
investigation as the basis of discussion. The general conclusion was favourable to 
the theory, although attention was drawn to several features of the phenomenon 
which remained difficult to explain. 
Among other works on the subject which have appeared since the publication of 
Schuster’s earlier memoir, those by Fritsche,* G. W. Walker,! and van Bemmelen| 
may be noted here. The two former authors confined themselves to the solar diurnal 
magnetic variations, but van Bemmelen broke fresh ground by applying harmonic 
analysis also to the lunar diurnal variations. The importance of the latter was 
fully recognized by both Balfour Stewart and Schuster, who, in his second 
memoir (p. 181), urged the desirability of further study of them. 
Some of the above, and other, writers reached conclusions adverse to the Stewart- 
Schuster theory, partly owing to the fact that the results of their analyses of 
the observational data differed from Schuster’s. The present paper embodies an 
attempt to resolve the points in dispute, and to remove other obscurities in the 
theory. New analyses are made both of the solar and lunar diurnal magnetic 
variations, so that the chief facts relating to each may be discussed together. It is 
taken as axiomatic, in view of the general resemblance between the two phenomena, 
that in the main the same theory and similar mechanisms must apply to each. 
Various modifications of Schuster’s hypotheses and results are found to be necessary, 
but the essential points of the theory are confirmed by this investigation. 
The discussion of the third and fourth, as well as the 24- and 12-hour, harmonics 
in the magnetic variations is one of the more novel features of this paper : 
previous discussions have generally been confined to the diurnal and semi-diurnal 
components, and doubts have been cast on the value of the higher frequency terms, 
which I hope the present investigation will remove. In the case of the lunar 
variations, only the semi-diurnal term has hitherto been used; this, however, was 
because, while other harmonics are present, their phases vary through a multiple of 
2ir throughout each lunar month, so that they disappear from the ordinary mean 
monthly variation calculated as it has been in the past. In an earlier memoir § 
I have shown how all four harmonics can be determined by computing the variations 
* Fritsche, St. Petersburg, 1903, and Riga, 1905 and 1913 (these papers were apparently privately 
printed and circulated). 
t G. W. Walker, ‘ Roy. Soc. Proc.,’ A, vol. 89, p. 379, 1913. 
| van Bemmelen, 1 Meteorologische Zeitschrift,’ 5, p. 218, 1912; 12, p. 589, 1913. 
§ ‘Phil. Trans.,’ A, vol. 213, p. 279, 1913; and A, vol. 214, p. 295, 1914. Also see ‘Phil. Trans., 
A, vol. 215, p. 161, 1915. 
