354 De la Rive on the Aurora Borealis. 
sort of effervescence, and a constant change of form, before it 
the explosion violent. Sometimes the columns of light proceed- 
ing from the luminous are are mixed with blackish or smoky 
columns; sometimes they rise simultaneously from different 
points of the horizon; or they may unite in a sea of flames of 
r 
a corona. Rarely the aurora continues till the corona is on 
clouds are grouped and arranged like the auroral columns; and 
in this case they appear to disturb the magnetic needle. After @ 
brilliant aurora, the trains of clouds in the morning have some- 
times been found to indicate the positions of as many luminous 
columns during the ni 
The absolute height of the aurora has been variously estima~’ 
ted. Fora long time it was supposed that it might be ascer- 
tained by the observations of distant observers on the corona: 
but it is now well known that the corona is only an effect of per 
spective, due to the apparent convergence of rays which are pa! 
allel to the dipping needle ; so that each sees his own corona, aS 
each his own rainbow. Moreover the aspect of the phenomenon 
depends on the position of the observer, The seat of the aurora 
is in the upper regions of the atmosphere; but sometimes it ap- 
pears to be produced within less elevated regions, where clouds 
are formed. Such observations as those of Capt. Franklin ap 
pear to establish the latter conclusion, who saw an aurora which 
lighted up the under surface of the clouds, whilst Mr. Kendall, 
two to three miles distant, saw no light whatever, although 
awake and constantly observing the sky. Captain Parry also as- 
