94 Yearly Variation of Magnetic Storms and Aurora. 



frequency of magnetic disturbances and aurorse, and those of the 

 greatest inclination towards the earth of the north and south solar polar 

 regions is clearly indicated. 



It is interesting to inquire in what, way this yearly inequality of 

 terrestrial magnetic phenomena is influenced when the sun's polar 

 regions are, for different groups of years, in an undisturbed and 

 disturbed condition. 



It would be expected that the oscillation of more disturbed solar 

 polar regions towards and away from the earth would tend to increase 

 the difference between the frequency of magnetic disturbance at the 

 equinoxes and solstices, while this difference for those years when the 

 less disturbed solar polar regions are in action, should be somewhat 

 reduced. That this is actually the case is brought out by the figures 

 which Mr. Ellis has given in the publication of which mention has 

 already been made. 



Since the greatest magnetic storms are closely associated in point of 

 time with prominence disturbances in the polar regions of the sun, to 

 make the necessary comparison, therefore, the years in which " great " 

 magnetic storms occurred should be grouped together and the yearly 

 inequality determined, and another group of years in which " great " 

 magnetic storms were less frequent formed and the yearly inequality 

 also determined. Fortunately a computation already made can be 

 utilised for this comparison, for Mr. Ellis has determined the number 

 of days of greater frequency (near sunspot maximum), and lesser 

 frequency (near sunspot minimum), of magnetic disturbance, both 

 groups practically including the conditions required. Thus he has 



formed groups of the years 184851, 185861, 186972, 1882 85, 



1892 95, which include, at any rate for the last three groups, the 

 years where prominences were in high latitudes and another series of 

 groups of years, 185457, 186568, 187679, 188790, which are 

 years when prominences were less frequent in these regions.* 



The interesting conclusion to which Mr. Ellis arrived was that " the 

 excess of the equinoctial frequency over the solstitial frequency is 

 greater, the greater the degree of disturbance." 



This result thus helps to endorse the suggestion made in a previous 



* The fact that continuous observation of solar prominences was only com- 

 L in 1870 accounts for our lack of knowledge of the frequency of this 

 )f phenomena before that date. Since, however, during the last three sun- 

 . has been observed that polar prominences are most frequent just a 

 i sunspot minimum and up to and at the epoch of the following sunspot 

 mm, it may be concluded that their appearance previous to the year 1870 

 : the same times in relation to the sunspot cycle. Ellis's groups of years 

 lou. to that date, namely, 1848-51 and 1858-61, may on these ground, be 

 WCh P u lar Pr minenees wer <? P re ^ whilst the groups 

 a8 eP Ch8 When P lar P^^nces were not 



