DIURNAL VARIATIONS OF TERRESTRIAL MAGNETISM. 
69 
examined without detection of error, and they also agree with those in Schuster’s 
investigation. We must conclude that the connection between the pressure and 
magnetic variations is decidedly less simple than our theory has so far assumed. 
One way in which this can easily be demonstrated may be indicated. If the phase 
of the solar diurnal pressure variation diminishes with height through 90 degrees, its 
phase will agree with that of the lunar diurnal barometric variation at the earth’s 
surface. The former change of phase is partly due to the solar semi-diurnal 
temperature variation in the successive layers of air, and this portion of the change 
will presumably have no counterpart in the lunar barometric variation. The 
remaining part, if any, may be ascribed to friction, and this may also atfect the lunar 
variation to a similar extent. On this hypothesis, the solar semi-diurnal oscillation 
of the atmosphere should be ahead of its lunar counterpart, while the corresponding 
magnetic variation lags behind the lunar magnetic variation by about 43 degrees. 
It would appear, therefore, that the solar pressure variation must diminish in phase, 
Relatively to the lunar variation, by (154° — 65) + 43°, i.e., through 128 degrees 
approximately. Part of this may be ascribed to temperature, but if any considerable 
portion is due to friction the lunar variation must be likewise affected to some extent, 
so that the balance of the above 128 degrees, after the “ temperature” portion of it 
is subtracted, must be a differential friction effect. Therefore either the diminution 
of phase due to temperature or that due to pressure, or both, must be larger than is 
generally imagined. It may be noted that the frictional effects usually referred to in 
this connection are those due to skin friction along the earth’s surface, or the eddy 
friction which has recently been brought into prominence by Major G. I. TaylorU 
True viscosity is generally regarded as so small as to be negligible, but this will 
hardly be the case in regions where the density is extremely small. 
The lunar diurnal pressure variation can hardly be other than of tidal origin. The 
difference of its observed phase from that which the equilibrium theory of the tides 
would predict (sin (2? + 90°) instead of the observed sin (2£+ 65°)) may, perhaps, be 
attributed to friction in the lower strata of the atmosphere. If the phase diminishes 
upwards to the value (108), the total actual retardation will be 127 degrees, or, 
measured from the theoretical tidal value, 152 degrees. It would be interesting to 
know whether there is any possibility of accounting for such large changes of phase 
by skin friction and viscosity. If not, there may be some hope of an explanation by 
a modification of the equation (31) connecting the pressure variation with the 
atmospheric motion. 
Until the phases of the annual harmonics in the magnetic variations are explained, 
those of the seasonal harmonics are not likely to be accounted for, and they will 
therefore not be discussed here. 
* “Eddy Motion in the Atmosphere,” G. I. Taylor, ‘Phil. Trans.,’ A, vol. 215, p. 1 (1915). Of. also 
‘Roy. Soc. Proc.,’ A, vol. 92, p. 196 (1916). 
