122 Disturbance of the Needle during an Aurora. 



In examining the column of this table which contains the observa- 

 tions on the horizontal needle out of doors, there appears a remark- 

 able decline of westerly variation, a movement of the north pole of 

 the needle to the east, at some time between the hours of six and nine, 

 P. M. This variation goes on diminishing until between thirty and 

 forty minutes past ten, after which, with the diminished brilliancy of 

 the aurora, the north pole of the needle begins to move westward. 

 This motion was so regular, and its correspondence with the decrease 

 of the brightness in the northern light so complete, that I supposed 

 the phenomenon about to cease, and left my station on top of the house 

 to examine the record of observations already made. On observing 

 at twelve o'clock, I found to ray surprise, that the north pole of the 

 needle had again moved eastward, as shown in the table, and on 

 passing rapidly into the street, the beautiful beams of which I have 

 spoken were seen : this sight I might have missed, having supposed 

 the aurora to be disappearing, but for the warning given by the nee- 

 dle. It is not improbable, from the general course of the phenome- 

 non, as already described, that the beams which I saw first to the 

 west of north, were not the first which had appeared, and that others 

 which preceded them may have been to the east of that point. The 

 variation did not reach as low a point during the brilliant part of the 

 phenomenon as during the former part, the observation at half past 

 ten giving a variation less by 3 J minutes than that for twelve o'clock ; 

 it is possible, however, that the second minimum occurred between 

 half past eleven and twelve o'clock. At one o'clock the disturb- 

 ance of the variation had not ceased, although it had become com- 

 paratively trifling, and the northern light had not entirely disappear- 

 ed. This I have reason to believe was the last part of the phenom- 

 enon. 



In order to present the results just stated in a form which addresses 

 itself more readily to the eye, I have traced (fig. 1, plate I,) a bro- 

 ken line, in which the intervals between the vertical lines represent 

 the intervals between the times of observation, and those between 

 the horizontal lines the differences of variation. In other words, 

 the abscissae of the line correspond to the times of observation, and 

 the ordinates to the amount of variation. From this figure it is at 

 once observable that the westerly variation increased irregularly from 

 SJ, A. M. until 3 P. M., when it attained a maximum ; that the de- 

 crease from this maximum was also irregular, and very rapid between 

 six and nine, P. M. : this includes a portion of the time of the occur- 



