Causes of the Aurora Borealis. 243 



is reflected by a material substance, existing in those places where it 

 is apparent, and that the pure atmosphere does not reflect the light. 



The appearance of the phenomenon in Siberia is compared to the 

 unwinding of a piece of " flame-colored taffeta ;" and in Hudson's 

 Bay it was said to roll over and over from one end of an arch to the 

 other ; as though a volume of luminous matter was unfolded, and 

 rolled from one point over to supply that which escaped and van- 

 ished at the other. Parry further compares the waving motions of 

 the horizontal bands of light, to the sinuous impulse given to a long 

 ribbon by a person holding it at one end. The streams and sheets 

 of light waving to and fro ; and the groups which in the Orkneys are 

 called " the merry dancers," moving with inconceivable velocity, 

 show the existence of an aerial medium, surprisingly flexible and at- 

 tenuated, obeying the impulses of currents and other agencies that 

 agitate the atmosphere.* The steady immovable banks of white 

 light result from the calm that prevails in that tract of the air where 

 they repose. 



Although the most frequent displays and brightest splendors of the 

 aurora are within the Arctic circle, it is not there more subjected to 

 periodical laws than in the temperate zone. It varies in every year, 

 no two years being parallel to each other. The differences are anal- 

 ogous to the changes of the weather, but without any dependence 

 upon them. It is probably essential that calms should prevail in those 

 regions of the air where the vapor collects, as strong currents would 

 break up and disperse it. This conjecture is strengthened by the 

 fact, that in the temperate latitudes the appearance of these beautiful 

 visions is a prelude to high winds ; and this peculiarity indicates still 

 further that its character has not been mistaken ; because calms are 

 proverbially precursors of storms ; and within the tropics are infalli- 

 ble harbingers of hurricanes. 



The irregularity of its appearance is doubtless chargeable to those 

 unknown influences which at one time favor, and at another prevent 

 the accumulation of the vapor, for probably like the clouds, it pre- 

 vails only occasionally. Alike variable, and almost as transient, it re- 

 sembles in its mutability the lightning and rainbow ; for although the 

 laws which produce those phenomena are understood, yet the auxil- 



* Analogous to those "small streams and currents of air/' seen by Humboldt, 

 "chasing trains of clouds with unequal velocity and in opposite directions," in the 

 upper regions of the Peak of Teneriflo and the Cordilleras. 



