DIURNAL VARIATIONS OF TERRESTRIAL MAGNETISM. 
13 
attempt was made to reduce the data to a common standard, so that the consequent 
drawback is less than in Fritsche’s investigations. Besides using various published 
data, new reductions were made for several stations, and the paper is valuable on 
this account as well as for its main purpose. For the present paper it was unfortu¬ 
nately necessary to re-calculate the lunar variation for some of these stations so 
recently dealt with by van Bemmelen, since the harmonics of varying phase could 
not otherwise be determined. The circumstance does, however, render possible an 
interesting comparison (§ 14) between the results of the two sets of computations. 
Although his data referred separately to the summer and winter half-years, van 
Bemmelen discussed only the mean annual values, neglecting the seasonal variations. 
He concluded that the horizontal force variations had a potential, which he deter¬ 
mined (after trial of Schuster’s method) by the method of least squares; unlike 
Fritsche, however, he used the separate values of a 2 and b 2 from each station instead 
of combining them into groups, and apparently, also, only the West force data were 
used. The vertical force data were similarly treated, and a separation of the internal 
and external parts of the lunar magnetic variation field was then effected. 
In a correcting paper of 1913 this calculation was revised, since the “least 
squares ” method of determining the potential seemed to give too much weight to 
some rather irregular data of early epoch from the three Southernmost stations—the 
Cape of Good Hope, Melbourne, and HobartonA Schuster’s method was returned 
to as enabling more discrimination to be exercised between the various data in the 
o 
course of the work. The resulting analysis was perhaps somewhat over-elaborate, 
but the principal harmonic, the one dealt with also in this paper, agrees moderately 
well with the result here obtained (cf. § 14). The original calculation had made it 
appear that the external variation field was actually less than the internal field ; the 
revised paper reversed this conclusion, although the inner field was given as more 
nearly approaching the outer field, in magnitude, than the new analysis of this paper 
would suggest, van Bemmelen, indeed, as the result of his calculations, still 
contemplated the possibility of a primary inner as well as a primary outer field. 
In his first paper he had also attempted to bring the lunar semi-diurnal variation 
into relation with the lunar semi-diurnal barometric variation (as observed at 
Batavia), just as Schuster had done for the solar diurnal variations in his 1907 
memoir. The discrepancy between Schuster’s and Fritsche’s analyses of the 
solar diurnal magnetic variation, which were both discussed by van Bemmelen, 
rendered the conclusions somewhat indefinite, and they must, in any case, have been 
superseded after the revised calculation of the lunar diurnal variation potential. In 
his second paper van Bemmelen avoided the ambiguity just alluded to, by making 
* Certain mistakes of sign had also been made in the first investigation, which were corrected. In the 
original paper the Bombay data were given as of thrice their true value, apparently through a numerical 
slip in the reductions, but on account of their discrepancy with other results they were excluded from the 
discussion. 
