ON THE PERIODICITY OF THE AURORA BOREALIS. 223 



coup de soin, indiquent l'e'tat de 1'atmosphere les jours qui precedent et qui suivent l'apparition d'aurores bore'ales. Or, 

 presque toujours, l'aurore est precede'e de la chute de la pluie et surtout de la neige ; il arrive aussi tres-souvent, mais 

 non tonjours, que cette chute a lieu apre's l'aurore. Enfin, l'apparition des halos luminares, qui precede Ie plus souvent 

 celle des aurores boreales, est une preuve de la presence dans l'atmosphere de ces particules glacees qui forment le reseau 

 illumine par la transmission de l'electricite." ' 



Wartmann has considered the question, whether there is a special period of auroral 

 display in October. 2 The influence of trade-winds and of a low barometer upon the 

 aurora has also been discussed. 3 Broun has examined the recurrence of auroral dis- 

 plays in connection not only with the seasons, but also with the moon's age and mag- 

 netic disturbances. From his own short series of observations he has concluded that 

 these disturbances, whatever and whenever they may be, which result in the develop- 

 ment of the auroral light, agree with the magnetic disturbances of the earth in having 

 a maximum at the opposition of the moon, and a minimum at the conjunction. These 

 magnetic disturbances also conform to another law of the aurora ; viz., they have max- 

 ima at the equinoxes and minima at the solstices, particularly at the summer solstice. 4 



Hansteen 5 has pointed out a yearly period in the changes of the magnetic inclina- 

 tion at Cbristiania ; the two maxima occurring on April 1 and September 30, and the 

 two minima on July 1 and December 31. If the solar spots can be linked with the 

 earth's magnetism, and if the earth's magnetism can be associated with the aurora, may 

 not the aurora and the sun's spots possess some bond of affinity ? The solar spots have 

 been tabulated by Wolf, month by month. The yearly curve resembles the variation 

 in the star (i Lyrse- as found by Argelander. The principal minimum is in July, and 

 a secondary one in November : the principal maximum is in October, and a secondary 

 one in December. This yearly curve corresponds to that which expresses the mag- 

 netic changes as far as the principal maximum and minimum are concerned. But the 

 magnetic curve has a secondary maximum in April and a minimum in January. 

 Both phenomena exhibit clearly an annual periodicity; there are more solar spots 

 from October to March inclusive than from April to September inclusive. 6 



"Wolf 7 expresses the monthly frequency of the solar spots between the years 1S33 

 and 1858 inclusive, by these relative numbers. 



J Traite' de 1'Electricite', ITI. 295. 2 Bibliotheque Universelle, XXX. 206. 



8 Bibliotheque Universelle, XLV. 89, and LXXXIII. See also the calculations of Fritz. Vierteljahrsschrift der Naturforsch. 

 Gesellsch. in Zurich. IX. 124-8. 

 * Edinb. PhU. Trans. XXX. Pt. 2, p. Lxxxi. 



6 Bulletin de l'Acad. Royale de Brux. XXI. 299. Vierteljahrsschrift der Naturforsch. Gesellsch. in Zurich. Jahrgang, II. 123, 

 and 353; V. 36 ; VI. 179. 



c Biblioth. Univ. Archiv. de Nat. et Phys. Sci. Vm. 41. 



7 Vierteljahrsschrift der Naturforsch. Gesellsch. in Zurich. Jahrgang, I. 264; II. 120; V. 29 and 240. 



