On I7apiil Periodic Variations of Terrestrial Magnetism. 5|^ 



nearly SAV-NE or rever.-e during tlio day time and N\V-SE or 

 reverse during niglit. The supposed system of currents can not 

 be directly identified with that given by Schuster or Bezold, since 

 in the latter case, « passes the value ^ twice during '24: hours, and 



the corresponding curve of « witli the hour as abscissa shows no 

 ai:>parent resemblance either with Fig. IG or with Fig. 10. Nor 

 is the system of currents deduced l)y Birkeland'^ from the 

 disturbances of longer durations observed during his memorable 

 auroral expedition similar to the supposed one. Remembering, 

 however, that the daily variation of the terrestrial magnetic 

 field can be represented in the first rough approximation l.^y a 

 system of currents having its axis considerably inclined to the 

 earth's axis, we may regard provisionally the total current system 

 given by Schuster and Bezold consisting of tw^o parts, and that 

 the one part which is the princij^al and represents a system of 

 zonal currents, shows a more conspicuous regular fluctuation than 

 the remaining secondary part more or less converging toward the 

 pole (see § 28). As a matter of fact, Fig. 19 represents only 

 the averafje distribution of the most fi-equent azimuth in different 

 hours. In actual cases, the points corresponding to different 

 trains of waves are so remarkably scattered that the adoption 

 of the mean value evaluated in the usual manner seems scarcely 

 justified for seeking the most frequent value. It is very probable 

 that among these widely scattered points, there are inany which 

 actually correspond to tlie fluctuations of the part of the current 

 belonging to the higher harmonics. At any rate, our conjecture 

 seems to be justified in the first approximation, though the 

 position of the pole of the circular current can not be determined 

 with certitude. 



20. As to the actual modes of fluctuation of the atmospheric 

 current causing the magnetic waves concerned several possibilities 

 are suggested at the same time. Firstly, we may consider the 

 total intensity as well as the distribution of the current as 

 constant, but oscillating as a whole about its mean position, 



1) Birkeland, loc. cit, PI. X. 



