AND ITS CHANGE WITH LUNAR DISTANCE. 169 



variation. For intermediate values of /3 intermediate states would prevail e.g., if 

 ft = JTT, the amplitude change is reduced to a ratio 1'30, and there is a phase change 

 of 17 degrees between perigee and apogee. The observed amplitude change in the 

 lunar magnetic variation might be a little less than 1'39, though hardly so much 

 below this as 1'30; this would be compatible, on the above view, with a change of 

 phase of something less than 17 degrees. Possibly in this way a part of the observed 

 phase change in the lunar magnetic variation might be explained (a more accurate 

 determination of the amplitude change would decide this point more definitely), but 

 in no circumstances could the whole of it be thus accounted for. Probably the 

 explanation which must be sought elsewhere will account for the whole of the 

 phenomenon under discussion. 



Before leaving the consideration of the atmospheric tide, its actually observed 

 phase, as determined from forty years' barometric observations at Batavia, may be 

 adverted to. The lunar diurnal variation of barometric pressure was found to be 



(TOG28 cos (2t + 05), 



t being reckoned from the local time of the moon's transit. The fact that the tide is 

 in advance of the moon is difficult to understand, and it is conceivable that the 

 unknown cause which thus accelerates the tide also has some connection with the 

 perigee-apogee phase change in the lunar magnetic variations. 



So far no real light on the origin of this phase change has been found, since theoretical 

 reasoning indicates no such change in the atmospheric tide which is supposed to 

 produce the lunar magnetic variations. Whether the phase of the tide does or 

 does not so change is, as already stated, unknown ; if it is found to vary correspondingly 

 with the magnetic variations, the tidal theory of the latter would be strengthened, 

 though the phase change in the tides themselves would offer a problem demanding 

 solution. In the present state of ignorance, however, it is natural to consider 

 whether, if the phase of the atmospheric tide is independent of lunar distance, the 

 phase change in the magnetic variations could be accounted for in any electromagnetic 

 way. Self-induction is the only possible cause which suggests itself, and this, 

 unfortunately, seems as incapable of offering an explanation as tidal friction was 

 found to be ; the effect of self-induction is to produce a phase retardation which 

 is independent of the amplitude of a periodic variation, but which diminishes with the 

 period. In the present case the variation of period would not account in this way for 

 more than 1 degree change of phase. 



What has been said as to the failure of tidal friction to explain the phase change in 

 the lunar magnetic variation applies also to tides in the substance of the earth, 

 so that even if (as VAN BEMMELEN supposed in his memoir already cited) such body 

 tides have a part in the production of the lunar magnetic variation, they do not aid 

 in the explanation of the phase change under discussion. VAN BEMMELEN in an 



VOL. ccxv. -A. z 



