62 
DR. S. CHAPMAN ON THE SOLAR AND LUNAR 
diurnal magnetic variations, the defect in the solar variations would seem to require 
us to assume some counter-balancing seasonal magnetic harmonics in the latter, due 
to atmospheric oscillations of type other than Q 2 2 . These can hardly be present in 
the lunar diurnal atmospheric movements, though no observational evidence is 
available. As regards the solar diurnal motions, the semi-diurnal barometric oscilla¬ 
tion is strikingly symmetrical about the equator throughout the whole year. The 
24-hour surface variation, though less definite and well-determined, also seems free 
from unsymmetrical components of type such as Qd. The symmetrical oscillation Qd, 
which has been suggested as a co-factor with Qd in the production of the annual 
magnetic harmonics, would also contribute seasonal harmonics. The amplitudes of 
the first two of these, comparable with (76) as are the two sets of numbers in (77), 
are approximately as follows :— 
Qd- Qfi- 
18 3 
The first of these would tend to neutralize the corresponding harmonic due to Qd, 
while in the second case there would be re-inforcement. The question of the exact 
origin of the seasonal variations must remain unsolved for the present, both their 
amplitudes and their phases (in view of the negative signs prefixed before the 
theoretical values of Qd) being difficult to explain. But as regards the presence of 
a 24-hour oscillation in the upper atmosphere, the calculations of § 26 indicate that 
there are possibilities of its local production by heating effects in the conducting 
layer, even if the surface variation of the same period does not persist into the upper 
atmosphere. 
§ 24. The Absolute Values of the Amplitudes and of the Electrical Conductivity in 
the Upper Atmosphere. 
We have provisionally concluded that both the solar and lunar diurnal magnetic 
variations are, in the main, due to semi-diurnal atmospheric oscillations, roughly of 
type Q_d, in conjunction with a variable electrical conductivity which may be approxi¬ 
mately represented by the formula (73). We will therefore now confine ourselves to 
the principal magnetic harmonics, Qd, in discussing the absolute magnitudes of the 
several variables involved in the theory. 
It is clear that if, as we suppose, the solar and lunar magnetic variations are 
similarly produced, the ratio of the amplitudes of Qd in the two cases will be equal 
to the ratio of the amplitudes of the corresponding semi-diurnal atmospheric 
oscillations. The former ratio was found in § 23 to be about 10. This is smaller 
than the ratio of the solar and lunar semi-diurnal barometric variations which, at 
Batavia for instance, is (l'00/0'063), or approximately, 16. It has already been 
noted, however, that (writing Sp for the pressure variation at a height where the 
pressure is p) 8p/p diminishes somewhat, with height (§ 19) ; Hann has explained 
