230 ON THE PERIODICITY OF THE AURORA BOREALIS. 



aurora about the year 1830, of which he could recall no parallel since the days of his 

 boyhood or since 1793, adds: — 



" Von solclicn Perioden habe icn geglaubt seit dem Jahre 502 vor Christi Geburt bis auf unsere Zeit 24 nachweisen 

 zu konnen, Ton Tvelehen besonders die neunte, von 541 bis 603, die zwblfte, von 823 bis 887, die zweiundzwanzigste, 

 von 1517 bis 1588, und die vierundzwanzigste, von 1707 bis 1788 sieh durch ungewbhnlich starke und haufige Nordlich- 

 ter anszeichneten." l 



Hansteen's periods of auroral display vary accordingly from sixty-two years to eighty- 

 one years. It does not seem to me, however, that such of the periods as he has speci- 

 fied are those which group the phenomena together most naturally ; and in the case of 

 the twenty-second period, which he makes to begin in 1707 and end in 1788 and cul- 

 minate about 1752, it may be objected that the observations indicate one maximum 

 about 1730 and another about 1790, and a closer approach to a minimum than to a 

 maximum about 1752. The observations, however, made by Celsius in Sweden, to 

 which, no doubt, Hansteen attached great weight in forming his opinion, are less de- 

 cisive on this point than those made at St. Petersburg or Cambridge. 



A very cursory glance at the curves which represent the frequency of the aurora 

 from year to year will detect the influence of other periodical fluctuations shorter than 

 the secular one and superadded to it. In the year 1803, Hitter of Jena maintained 

 that the aurora was subject to a periodical variation once in nine or ten years, and that 

 this short period of the aurora was associated with the period of 18.6 years in the 

 moon's motion, therefore with the larger nutation-period of the earth's axis, so much 

 so that the minimum for the aurora corresponded to the largestand smal lest obliquity 

 of the ecliptic, and the maximum of the aurora to each intermediate mean obliquity 

 when the longitude of the ascending node of the moon- was 90° or 180'. From all 

 which Ritter concluded that the connection of the moon with the aurora was unques- 

 tionable. He predicted such a maximum for the year 180G and another for 1816. 2 



I have given in Table LIII. the years of maximum, minimum, and mean obliquities, 

 and also the years of maximum and minimum frequency in the aurora. The latter are 

 printed in capitals, where there is a plausible agreement with Eitter's theory. Al- 

 though the coincidences between the facts and the theory are in some parts remark- 

 able, the discrepancies are numerous and large : too numerous and too large to be 

 reconciled with the truth of the theory. It may be stated, for the benefit of any one 

 who wishes to pursue the comparison further, that the years 1G02, 1611, 1620, 1630, 



1 Pogg. Annalen der Physik und Chemie, XXII. 536, 7. 2 Gilbert, Annalen der Physik, XV. 210, 220, and XVI. 226. 



