THE AURORA BOREALI3. 



one and sometimes both of its extremities would desert the horizon, and then 

 its folds would become more numerous and marked, the bow would change its 

 character, and assume the form of a long sheet of rays returning into itself, 

 and consisting of several parts forming graceful curves, as represented in fig. 5. 



Fig. 5. 



The brightness of the rays would vary suddenly, sometimes surpassing in 

 splendor stars of the first magnitude ; these rays would rapidly dart out, and 

 curves would be formed and developed like the folds of a serpent ; then the 

 rays would effect various colors, the base would be red, the middle green, and 

 the remainder would preserve its clear yellow hue. Such was the arrange- 

 ment which the colors always preserved; they were of admirable transparency, 

 the base exhibiting blood-red, and the green of the middle being that of the 

 pale emerald ; the brightness would diminish, the colors disappear, and all be 

 extinguished, sometimes suddenly, and sometimes by slow degrees. After 

 this disappearance, fragments of the bow would be reproduced, would continue 

 their upward movement, and approach the zenith ; the rays, by the effect of 

 perspective, would be gradually shortened ; the thickness of the arc, which 

 presented then the appearance of a large zone of parallel rays (fig. 6), would 

 be estimated ; then the vertex of the bow would reach the magnetic zenith, or 

 the point to which the south pole of the dipping needle is directed. At that 

 moment the rays would be seen in the direction of their feet. If they were 

 colored, they would appear as a large red band, through which the green tints 

 of their superior parts could be distinguished ; and if the wave of light above 

 mentioned passed along them, their feet would form a long sinuous undulating 

 zone, while, throughout all these changes, the rays would never suffer any os- 

 cillation in the direction of their axis, and would constantly preserve their 

 mutual parallelisms. 



While these appearances are manifested, new bows are formed, either com- 

 mencing in the same diffuse manner, or with vivid and ready-formed rays : 

 they succeed each other, passing through nearly the same phases, and arrange 

 themselves at certain distances from each other. As many as nine have been 



