44 BULLETIN OF THE 



(abstract.) 



The two generally recognized circumstances of an annual pe- 

 riodicity in the number and brilliancy of auroral displays, and of 

 the great elevation — of several hundred miles — frequently attained 

 by their arches and streamers, an elevation to which our atrao 

 sphere can hardly be supposed to extend, having led to a question 

 whether extra-terrestrial matter in the form of cosmical dust or 

 gaseous rings (similar to the August and November rings of me- 

 teors) might not be concerned in the phenomenon, by being peri- 

 odically grazed by our planet or its atmosphere — causing the 

 discharge of electrical brushes — the recent elaborate Memoir of 

 Prof. Lovering on Auroras was consulted with much interest, to 

 see what light, if any, might be thrown upon the speculation by 

 his carefully tabulated results. 



As the summing up of a large number of local observations, 

 two independent. tables, classified by months, are given; the first 

 table being derived from catalogues reaching back some 575 years 

 from 1864, and embracing an aggregate of 9885 observations ; 

 and the second or supplementary table covering about 50 years, 

 reaching down to 1868, and embracing an aggregate of 2497 

 observations (amounting together to 12,382 observations) : both 

 show when presented graphically the same characteristics, namely, 

 two very notable maxima of frequency, in March and October, 

 and one very remarkable minimum in June — the second minimum 

 in December being much less marked. By plotting a third curve 

 giving the summation of the two curves, this is still more clearly 

 presented : the total maxima of March and October being re- 

 spectively 1436 and 1341 ; and the two minima of June and 

 December being respectively 455 and 1090. 



Mairan in 1754, from a tabulated catalogue of 1441 auroras, 

 estimated the number occurring at our perihelion (in the early 

 part of January), as being about seven times that at our aphelion 

 (in the early part of July). Prof. Lovering, however, from his 

 much larger collection, has pointed out that the two minima agree 

 very nearly with our solstices, and the two maxima with our equi- 

 noxes. The great minimum of June occurs when our nights are 

 shortest ; but after making due allowance for this, there is still a 

 very marked deficiency observable. 



Plotting in a corresponding manner the monthly frequency of 

 meteoric displays, in three separate curves, representing, first, an 

 aggregate of 1358 observations collected by Biot from the Chinese 

 annals (reaching back considerably beyond the Christian era) ; 

 secondly a table of Arago, comprising an aggregate of 813 ob- 

 servations of noted fire-balls ; and thirdly, a table of Baumhaur, 

 of aerolites and fire-balls observed previous to 1845, amounting 

 to an aggregate of 767, we see that while there is a rough cor- 

 respondency in these three curves, they show apparently no rela- 

 tion whatever to the curves of aurora frequency. 



