ON THE PERIODICITY OF THE AURORA BOREALIS. 321 



of observations. Either the year 1716 or 1719 is the more probable time. No auroras 

 were observed in Sweden in 1717 or 1719, and only one in 1716 and 1720 ; but nine 

 ■were observed in 1718. Four auroras were observed in Berlin in 1717; but four were 

 also observed in 1718, 1720, and 1722. In Great Britain, two auroras were seen in 

 1717, none in 1718, but nine in 1719. Two auroras were seen at Salem in 1719, the 

 first recorded appearances in New England. 4. In 1723 comes another spot-minimum. 

 In Sweden there is a plausible conformity on the part of the aurora, though the mini- 

 mum comes one or two years earlier. In Great Britain no auroras were observed in 

 1724 ; but more were seen in 1723 than in any previous year since 1719. In Berlin, 

 also, none were seen in 1724, but more were seen in 1725 than at the spot-maximum of 

 1717. In Germany an aurora minimum may be conceded to the year 1723, although 

 three more were seen than in 1720, and four more than in the year of spot-minimum, 

 1717. The twenty-eight auroras observed in that country in 1722 militates strangely 

 against the supposition of an auroral minimum the very next year. The absence of all 

 auroras in America in 1723 proves nothing, as they were equally wanting in 1717 and 

 1727, both years of spot-maxima. 5. I pass on to the spot-maximum of 1727. A 

 large number of auroras is recorded against this year in Italy. But the force of this 

 agreement is weakened by the fact that almost as many were seen in 1730 and possi- 

 bly more in 1724. In France none are found for 1727, and fifteen for 1724. In 

 Germany only thirteen auroras were seen in 1727. and forty-two in 1730. In Berlin 

 the number of auroras was larger than in 1723, but not so large as in 1730 and 1732. 

 In Great Britain the number of auroras was greater than in the years immediately 

 preceding and following, but exactly the same as in 1723. In Sweden eight auroras 

 were observed, or more than in any other year since 1718, but the number continually 

 increased until it reached fifty-two in 1730. In St. Petersburg eleven auroras were 

 seen in 1727, and thirty-six in 1730, with six and twenty-five for the two intermediate 

 years. The observations at Abo and Cola may be left out of the account, as the series 

 are evidently defective. 6. In 1733 occurs the next spot-minimum. In Italy only 

 one aurora was seen this year and the next, but none in 1729. An extraordinary 

 number of auroras are recorded against the years 1732, 1733, and 1734, although 

 1733 is the minimum year of the three, in France, which, if not otherwise explained, 

 flatly contradict the Wolf theory. The Wittemberg observations exhibit an unqualified 

 minimum for 1733. The German observations, in general, point to an equally decisive 

 minimum, although it comes a year later. Three times as many auroras were seen 

 in Holland in 1733 as in 1729, but a great increase appears in the following years. 

 In Berlin there is a goqd maximum in 1732, and nearly one in 1733. In the two 

 vol. x. 41 



