Causes of the Aurora Borealis. 245 



rection, and they fall so nearly horizontal to the earth as to illuminate 

 the higher heavens far above and beyond the pole ; consequently re- 

 fracted light from the northern quarter, could not be visible on the 

 globe. But the refractive power of the atmosphere increases from 

 the equator to the pole, and in high latitudes produces a twilight of 

 several hours, morning and evening. Even after the sun has set in 

 lat. 14P and 75° for the long night, a beautiful twilight returns peri- 

 odically in the south, marking his diurnal motion, and the direction 

 of his incident rays. When the direct lines of light are so far with- 

 drawn, as no longer to come within the angle of visible atmospheric 

 refraction, the rays are refracted from the receding twilight by the 

 vapor of the aurora in that quarter nearest the sun ; consequently, 

 near the pole the aurora is seen in the south, which is not opposite, 

 but in the direction of the illuminating source. 



4. It commonly appears as soon as it is dark, and disappears be- 

 fore midnight, which further indicates its dependence on solar light; 

 for owing to the greater refractive power, and the altitude of the va- 

 por, it reflects the light, as long as either the direct or refracted rays 

 reach an elevation that will fall upon it. It is then visible upon the 

 earth, but when the sun is in the nadir, the light is so withdrawn at 

 that point of time, as to be invisible between the zenith and horizon. 

 Within the temperate zone it has sometimes been seen at midnight 

 and in the zenith, both which anomalies must result from the light 

 reflected from one mass to another ; but these instances are of rare 

 occurrence. There is a region, however, a sort of neutral ground, 

 between the point where it ceases to appear in the north, and be- 

 gins to appear in the south, where it is seen in every quarter of 

 the heavens, probably comprehending a belt around the globe, be- 

 tween N. Lat. 65° and the Arctic circle. Over these latitudes the 

 sun's rays, from their increased obliquity, enlighten a tract of the 

 heavens, without being intercepted by the figure of the earth, and 

 at the same time illuminate the air at such an angle, that the refract- 

 ing medium renders the light indiscriminately visible throughout the 

 hemisphere.* 



* I think the intensity of the light probably owing to two causes; the quantity of 

 the vapor present, and the precise degree of the sun's obliquity. I cannot think that 

 the increased refraction is caused by the increase of atmospherical density, for the 

 barometrical pressure was not at its maximum when the aurora was most splendid 

 at Winter Island in lat. 66°. Nor does the difference of density between the trop- 

 ics and the high latitudes seem sufficient to account for the brilliancy of the arctic 



Vol. XIX.— No. 2. 32 



