H2 REPORT — 1886. 



upon the state of tlie sun's surface, an increased area of spotted surface 

 coinciding apparently with increased values of the daily means of hori- 

 zontal force all over the earth. 



While such currents might be supposed to possess, as a whole, no dis- 

 tinct daily variation, yet at the time when the sun heats a tropical region 

 it might be supposed to increase the relative conductivity of that region 

 with respect to that of the atmosphere nearer the pole. It would thus 

 divert to the heated region an unusual proportion of the whole current, 

 so that we should have a maximum of horizontal force near noon in the 

 equatorial, and a minimum at the same time in the polar regions. This 

 is probably the case. 



15. One chief object in giving prominence to this part of the subject 

 is with the view of advocating that the Gaussian method of analysis 

 should not be applied merely to the solar-diurnal variation of the three 

 magnetic elements, but should likewise embrace a consideration of the 

 simultaneous variations in the mean daily values of the elements at various 

 stations. We must, in fine, consider the possibility at least of there being 

 in the upper atmospheric regions, not merely currents which present a 

 marked solar-diarnal variation, but others that have no marked solar- 

 diurnal variation, while yet they may be highly susceptible to changes in 

 the sun. The double method of treating matliematically not merely the 

 solar-diurnal variation, but likewise the simultaneous changes in the 

 mean daily values of the elements, would thus appear to be necessary 

 and sufficient for giving us the required information. 



16. If we turn from the solar-diurnal variations to those caused by the 

 moon, we find in this region likewise an attempt to explain the phenomena 

 by the same working hypothesis. It has been remarked by Dr. Schuster 

 that we Uve at the bottom of the atmospheric ocean, where lunar tides will 

 necesparily be small, and lie imagines that in the upper regions of the 

 atmosphere the motions caused by lunar tides may be very considerable. 

 Such motions would be subject to the same magneto-electric laws as 

 those caused by the sun, and we might therefore expect a lunar semi- 

 diurnal magnetic variation, such as, in fact, we have. The late John 

 Allan Broun has shown that the moon's magnetic effect varies approxi- 

 mately as the inverse cube of the moon's distance from the earth, a con- 

 clusion that would seem to point to some sort of tidal influence as the 

 cause of this effect. 



17. Again, if this tidal iufluence be seated in the tipper atmospheric 

 regions, it should be greater durins: the day (when these regions are 

 heated, and so become good conductors) than during the night. Now, 

 Broun was the first to point out that the semi-diurnal lunar variation at 

 Trevandrum, in India, is .subject to this law, and his results have lately been 

 confirmed, in an independently conceived investigation, by Mr. C. Cham- 

 bers, of Bombay. We might likewise expect that the lunar variation, 

 like the solar one, should be greatest at times of maximum sun-spot 

 frequency, and there is some reason to think that this is the case, 

 although the fact is not yet definitely established. 



18. There seems, therefore, reason to believe that the diurnal varia- 

 tion of any one magnetic element — the declination, for instance — may 

 be due to the joint action of several causes, which we may, perhaps, 

 I'epresent as follows : — 



In the first place, the sun may act in producing atmospheric motions 

 in the upper regions ; this would cause a solar diurnal magnetic effect. 



