ON THE SECULAR PERIODICITY OP THE AUROR 



10 



to communicate with the public by printed books, or through the transactions of 

 academies. And he quotes from Fontenelle a remark to the effect that a man does 

 not often see more than he already knows to exist. Considerations like these aiv, no 

 doubt, entitled to great weight in comparing together the number of auroras observed 

 in ancient and modem times. They account in part for the increasing number of ob- 

 served auroras as we approach the present epoch ; but they fail to explain the alter- 

 nate increase and decrease in the number of auroras from one quarter of a century to 

 another, and especially will they fail if it is found that the same fluctuations which 

 are indicated by Mairan's historical investigations are repeated during the last century, 

 and even under our own eyes. 



I shall next consider how the observations of the last one hundred years affect this 

 question of the periodicity of the aurora. In 1754, Mairan * commented on the fact 

 that the aurora had begun again to diminish in frequency in France, so that he found 

 no instance between November 3, 1740, and February 3, 1750, which would an -wer 

 the purpose of determining the height of an aurora by its parallax. The St. Peters- 

 burg observations declare that Mairan was correct in his forebodings. They famish 

 two hundred and sixteen auroras between 1729 and 1743 inclusive, and only forty-six 

 between 1744 and 1758 inclusive, there being thirty-five in one year, 1730, and none 

 in 1753 and 1754. Dalton t has published two hundred and twenty-seven appearances 

 of the aurora in Kendal and Keswick, between the years 1787 and 1793, of which 

 twenty-nine were observed at both places, and all but ten at Kendal. In Dal ton's 

 catalogue of auroras, observed in Great Britain and Ireland between the years 1793 

 and 1834, fifty-five occurred before 1810, and only seventeen between 1810 and 1826 

 inclusive, though the latter period is longer by one year. Then, again, one hundred 

 and thirteen auroras were observed in the last eight years of Dalton's observations, 

 between 1827 and 1834 inclusive. In this latter period eight are designated n*grm«l 

 and many others as fine. Thirty-two auroras were seen in one year, 1830. But there 



were none in 1807, 1809, 1810, 1811, 1812, 1813, 1815, 1822, 1823, and 1824. No 



wonder that Singer remarks in his Elements of Electricity,! published in 1814, that 

 the aurora was then rarely visible in England. Arago§ stated in September, 1827, 

 that no aurora had been seen before in Paris for twenty years. Bdckmann || observed 



at Carlsruhe twenty-three auroras in 1779, seventeen in 1780, fifteen in 1781, eight in 

 1782, ten in 1783, one in 1789, and no more for the next 12 years. 

 It may be interesting to inquire how the case stands in the western hemisphere, and 



* Page 430. f Meteor. Observ. and Essays, p. U. \ Page 253. 



§ Amer. Journ. Sci., XIV. 107. || Gilbert Ann., VII. 32. 



