ON THE PERIODICITY OF THE AURORA BOREALIS. 205 



■ 



regard to the frequency of the aurora, although Hansteen \ Kaemtz 2 , and at one time 

 Broun 3 , seem to imply that he did; nor would such a comparison have affected the 

 fate of his theory. As early as 1736, Kraft 4 concluded from 141 auroras, observed at 

 St. Petersburg, that the phenomenon occurred more frequently at the equinoxes than 

 at the solstices. Since that time the evidence has accumulated so as to place this con- 

 clusion beyond doubt, and to show, also, that the summer minimum is the principal one. 5 

 The difference between the summer and winter minima is attributable in part to the 

 short nights and long twilights of summer, as contrasted with the long nights and short 

 twilights of winter. But Muncke calculated that the aurora is a little more frequent 

 in winter than in summer, even after allowing for the interference of daylight with the 

 visibility of the phenomenon. 6 On the other hand, if the effect of sunlight in obliterat- 

 ing the aurora is eliminated, the contrast between the maxima of spring and autumn 

 and the minimum of winter will stand out even bolder than it now does. 



It is not surprising that some observers, especially in high latitudes, have found the 

 aurora Wholly wanting in summer, particularly in June and July. 7 Fortunately for sci- 

 ence, an aurora can he fell, by its influence on the magnetic needle, even when its pres- 

 ence is invisible. It appears, as might have been expected, that the comparative 

 frequency of the auroras actually seen at different seasons of the year varies considerably 

 with the latitude of the observer. I have given for each place, where a series of obser- 

 vations has been made, a table in which the auroras are arranged according to the 

 months of the year ; and the aggregates of the different months may also be expressed 

 by a yearly curve. Moreover, all the observations, without regard to geographical 

 locality, have been collected into a similar table, and represented by a similar curve. 

 By comparing the general numbers and curve with the numbers and curve which refer 

 to any particular place on the earth, whatever is local in the phenomenon will manifest 

 itself. In the general curve, the spring maximum occurs in March. But in Dunse, N. 

 Britain, the maximum appears as early as January, while at Utrecht it is as late as 

 May. The summer minimum is found on the general curve in June, while on the 

 yearly curve drawn for some parts of the earth it comes as early as May or as late as 

 August. The autumn maximum stands, on the average of all the observations for the 



i Mem. do l'Acad. Belgique, XX. 117. 2 Course of Meteorology, p. 458. 



a Edinburgh Phil. Trans., XIX. Pt. 2, p. LXXXI. i Coram. Acad. Scient. Impcr. Petropol. IX. 340. 



6 Sur la Physique du Globe, par Quetelet, pp. 121-23. 6 Gehler. Physikalisches Worterbuch. VII. 148. 



7 Hansteen remarks that the aurora is less frequent at the winter solstice than at the equinoxes, although the winter nights are 

 longer and more favorable for observations. At the summer solstice the twilight is so bright at Christiania that a book may be read 

 at midnight. This light obliterates the fainter light of the aurora ; but the presence of the latter is felt in tho magnetic disturb- 

 ances. — Nouv. Mem. de l'Acad. Brux. XX. 117. 



