AURORA BOREALIS. 115 



variety to the original colours as reflected from the point of 

 incidence ; and further, if due regard is paid to the properties of 

 light, its connexion both with magnetism and electricity may be 

 satisfactorily explained. 



In support of this theory — In the first place, we have the 

 fact, that when the spectator's position on the earth is to the north- 

 ward of the 69° of north latitude, the Aurora is generally observed 

 to bear to the southward, very seldom to the northward of east or 

 west, and never in that direction or in the north, unless the sun is in 

 opposition, or a region of icy or snowy substance is between the 

 spectator and the bearings of the sun ; although the illimiinated 

 clouds which I have mentioned might there, as in a lower latitude, 

 very possibly reflect the rays they have received in every horizontal, 

 as well as vertical direction. 



Secondly. The atmosphere between the sun and the spectator is 

 always clear, whether he is to the southward or northward of the 

 Aurora. When he is looking at the Aurora from a high latitude, 

 towards it in the south, the sun is then at his back to the north ; and 

 if the sky becomes cloudy in that (the northern) direction it is fatal, 

 the Aurora immediately disappears, because the rays are inter- 

 cepted by an impenetrable cloud. In like manner, when the 

 spectator observes the Aurora bearing to the eastward or westward, 

 if a cloud or fog intervenes between his position and the sun, it 

 immediately disappears. 



Tiiirdly. When the spectator sees the Aurora to the northward, 

 he is always to the southward of the icy regions, and at that time 

 the sky in the direction of the sun is always clear ; should a cloud 



q2 



