346 MEMOIRS OF THE AMERICAN ACADEMY. 



depending upon locality, which seem to have no parallel in the secular periodicity 

 of the aurora* A secondary period of about eighty-eight years, within the larger 

 one, with a range of only T \ of the value of the principal oscillation, has been indi- 

 cated by a close .scrutiny of the observations made at some stations, but the length, 

 epoch, and range of this subordinate oscillation vary with the locality.! The best 

 knowledge of the present day, therefore, furnishes no foundation for the inference, 

 that, because terrestrial magnetism undergoes secular changes, and also the frequency 

 in auroral displays, the two classes of phenomena have a parallelism, not to say some 

 connection with each other. Such a connection UssherJ seems to have hinted at 

 many years ago when he calls attention to the fact of a chasm in the auroral records 

 about 1661 for forty years, compared with the years which preceded and followed, 

 and to that other- fact that during this period of vacancy the declination of the mag- 

 netic needle at Paris became zero. The curve of no variation (or declination) passed 

 through London about 1660, through Paris about 1663, through the Cape of Good 

 Hope between 1605 and 1609. If this event produced any effect whatever upon the 

 recurrence of the aurora, we should expect that effect to antedate by fifty years, at 

 some places, the period of its occurrence at other places. Nothing but negative 

 evidence on this subject can be found. But what we know of the aurora during the 

 last hundred years assures us that the periods of the secular maxima and minima are, 

 to a great degree, independent of geographical locality. If the magnetic poles of the 

 earth are moving, and consequently the magnetic latitudes of places are changing, we 

 can easily imagine that such a change would have some influence on the frequency 

 of auroras. But no motion of the magnetic poles has been discovered, which, in itself 

 or in its effects, conforms, in period, to the well-established secular variations in the 

 aurora. These variations evidently depend upon some cause which operates alike, 

 and nearly simultaneously, over remote continents. De la Rive remarks : " Nous 

 concevons done facilement que le deplacement des glaces polaires, joint a celui du 

 pole magnetique de la terre, ait pu determiner momentanement des conditions peu 

 favorables a l'apparition de l'aurore boreale. Une chose assez remarquable, e'est 

 que cette longue periode sans aurore boreale coincide avec l'epoque ou l'aiguille 

 aimantee etait exactement dirigee dans le meridien terrestre (annee 1662). "§ As 

 equally decided minima in the frequency of auroras occurred in the middle of the last 



* United States Coast Surrey Report, 1859, p. 299. 

 t United States Coast Survey Report, 1858, pp. 193, 194. 

 ± Transactions of the Irish Academy, 1788, H. 191. 

 § Traite d*£lectricite, III. 304. 



