86 Mr. J. A. Broun on Terrestrial Magnetism. 



An examiuation of the first edition of Mairan's work induced me 

 to refuse tlic credit to that philosopher. M. Secchi, who has 

 since examined the same edition as myself (1733), has repeated 

 IMairan's claim. A careful perusal of the second edition of 

 Mairau^'s work induces me still to believe that Mairan did not 

 announce this fact distinctly in his first edition ; but in the 

 supplement to the second edition (1754), Mairan includes the 

 observations of Celsius (1733), Frobesius (1739), Delisle (1738) 

 and others, all of which show this law distinctly. Mairan then 

 refers to it in the following manner : — " On a pu remarquer en 

 jctant Ics yeux sur les sommes pour les mois de la Table com- 

 posee No. 77, et de la plupart de celles qui la composent, que 

 la frequence des ap])aritions de Faurore boreale autour des equi- 

 noxes ou environ un mois avant et apres, est aussi grande et 

 quelquefois plus grande qu^autour du perihelie ; ce qui paruit dit- 

 on, infirmer les inductions que nous avons tirees de la frequence 

 du phenomene autour du perihelie." (De Mairan, Traite de 

 rAurore Boreale, 2nd edit. 1754, p. 532.) 



The law, therefore, is as old as the time of Mairan, whether 

 its clear enunciation be due to that philosopher or not. The 

 law has been verified by every moderately long series of obser- 

 vations since then, and latterly by M. Hansteen and by myself. 

 It coincides with the annual law of magnetic disturbance found 

 by me. 



According to these results, I think we might perhaps be 

 allowed to suppose that for every law of magnetic disturbance 

 there is a corresponding law for the aurora borealis. 



In this case we might conclude that the number of aurorse 

 seen in summer in Europe is much less than it ought to be, not 

 only because of the strong twilight, but also because the evening 

 before midnight is the epoch of minimum disturbance, and there- 

 fore probably the epoch of minimum intensity and visibility of 

 the aurora. 



Further, the decennial law of disturbance should also have its 

 corresponding law of aurora; and if we could prove that previ- 

 ously the decennial law of frequency of aurora did not exist, we 

 might ])erhaps conclude that the decennial law of disturbance is 

 not continuous. On examining the numbers of aurorse collected 

 by ]\Iairan from all sources, I find that we can draw no conclu- 

 sion till the year 1699; before that time two or three yearly 

 were all that had been registered. In 1699, however, we find a 

 large number (40) ; 1719, again, is in the middle of a group of 

 years, when an increase commences, which, however, continues 

 till 1732. If we take the means of every three years to elimi- 

 nate accidental variations, there is evidence of nothing but a 

 gradual increase of frequency from one in 1710, to ninety iu 



