ON THE PERIODICITY OF THE AURORA BOREALIS. 227 



greater one in 1TS2-3. Olmsted 1 placed this maximum somewhere between the limits 

 of 1760 and 1783, and Loomis decides that its epoch is about 1780. Ussher 2 alludes to 

 the greater frecpiency of the aurora in 1788 as compared with the preceding half century. 



The mean interval between the two grand maxima which I have described is fifty- 

 seven years. The uninterrupted series at St. Petersburg leaves fifty-six years between 

 the two maxima, the broken series at Berlin only fifty-four years, and the Huxham and 

 Dal ton catalogues, combined, only fifty-two years. 



A second minimum comes along about the year 1812. No aurora was seen in Salem 

 between the years 1810 and 1813 inclusive; nor in New Haven between 1808 and 

 1813 ; nor in Great Britain between the years 1809 and 1813. None is registered for 

 St, Petersburg between 1809 and 1823 ; nor at Abo between 1806 and 1816. During 

 the last one hundred and fifty years the aurora has failed utterly of being registered 

 anywhere on the earth, only in the years 1812 and 1813. 



A third grand maximum returns in 1849. The Cambridge observations alone post- 

 pone it to the year 1852, and the New Haven observations hasten it on as early as 1847. 

 In New York it was the year 1850 ; in Providence, 1852 ; in Newberry, 1853. In "Wor- 

 cester the principal maximum was in 1840, but there was another in 1848. The short 

 series at Dunse exhibits a maximum in 1843, and that at Makerstoun in 1845. In 

 Christiania there was a maximum in 1852, but a greater one in 1842. In Jakobskavn 

 the maximum of a few years falls upon 1845. The interval between the second and 

 third maxima is about sixty-two years. The Cambridge observations alone imply an 

 interval of sixty-six years, and those of New Haven either sixty-one or sixty-six years, 

 according to the epoch selected for the earlier maximum. There is no other single 

 series which spans the long period between these two maxima, the aurora not being 

 as regularly observed now as formerly at St. Petersburg, and its presence being inferred 

 from magnetic disturbances rather than from direct observation upon the light itself. 

 I have been surprised to see no auroras included in the observations at Sitka in 1853, 

 although there is mention made of many extraordinary magnetic disturbances. 3 Loomis 

 concludes that the average period is about fifty-eight years. Santini 4 sets it as high as 

 sixty-five years. This was also the period which Olmsted adopted. Observations in the 

 Western continent would seem to indicate that the last grand maximum was superior to 

 its predecessor, while the European observations imply the contrary. The multiplication 

 of observers, which explains the first fact, only renders the other more mysterious. 

 Que.telet, who observed at Brussels, adverts to this peculiarity, saying that between 1778 



1 Smithsonian Contributions, VIII. 38, 9. 2 Trans. Roy. Irish Acad. 1788, II. 191. 



3 Annal. de I'Obserr. Phys. Centrale de Russie, 1853, p. 851, &c. * Wocheuschrift fur Astronomie, Meteorologie, &e. 1862, V. 14. 



