PART II. POLAR MAGNETIC PHENOMENA AND TERREL^A EXPERIMENTS. CHAP III. 525 



In those cases for which a coincidence exists, the storms, as usually found, occur somewhat later 

 The average lag 1.6 days gives / / equal to 21, where i, and 1 are respectively the longitudes 

 of sun-spot and central meridian at the time of the maximum of storminess. The lag here found is 

 only half as large as that found by Ricco (') for a number of very great storms. 



As regards the principal maxima, those of September, October and November coincide with quite 

 large and distinct groups of sun-spots. After that a marked maximum of storminess reappears quite 

 regularly at the end of December; but the sun-spots have disappeared. Nor do the great principal 

 maxima of February and March coincide with sun-spots. Not until April does there seem to be an 

 apparent coincidence. 



Regarding the connection between sun-spots and storminess, it seems improbable that the sun-spots 

 can be the direct cause of the magnetic storms; for the sun-spots appear to be rather irregular in their 

 occurrence and with a somewhat different period of revolution than the source of electric radiation. If, 

 then, the source were formed in any way by sun-spots, we should hardly find the variation of stormi- 

 ness so regular as it was actually found to be during the period of our observations. The results 

 suggest that sun-spots and magnetic storms are both of them manifestations of the same primary cause. 



The storminess seems to go on whether there are sun-spots or not. But also from our point of 

 view we shall expect to find that the passage of sun-spots is accompanied by magnetic storms; for the 

 existence of sun-spots is to be considered as a visible sign of a great activity of the primary source. 

 The effect will undoubtedly in a number of cases be the same as if the sun-spots themselves were 

 sending out pencils of electric radiation. The strong magnetic fields near the sun-spots show that violent 

 currents of electricity are actually operating in the sun-spots, and these currents may only be another 

 effect of the same electric activity which produces the magnetic storms and aurora. 



As we saw, the existence of one well-defined monthly maximum would require that there were 

 one single complex of sources which was greatly predominant with regard to emission of electric 

 radiation. It must, however, by no means be regarded as a matter of necessity that the same source 

 should always maintain its predominance; but it is quite possible that the intensity of one source may 

 diminish, and that of another increase so as to take the lead for a certain number of revolutions of the 

 sun, until a new one is called into play to become the principal source. 



In fact we saw that the results of our observations were best explained by dividing the principal 

 maxima into two groups, and in view of the previous considerations these two groups correspond to two 

 different complexes of sources. The first group has its last principal maximum at about January 28 and 

 the second one its first principal maximum at about February 7, consequently the difference in helio- 

 graphic longitude of the two sources should be about 120. 



This change in position of the source must be taken into account when we are dealing with the 

 determination of the monthly period of storminess. In fact, if the period were deduced in the ordinary 

 way from material covering a great many years, the shifting of the source would have the effect of 

 masking the "real" period, or the period deduced from a very long interval of time might be quite 

 different from that here found from the intervals between successive maxima. It is really no wonder 

 then that various authorities have found a different monthly period. 



(') Nature 82, p. 8. 1909. 



Birkeland. The Norwegian Aurora Polaris Expedition, 1903-1903. 67 



