Progress of the Aurora Borealis. 307 
enlarge in width, by a change exactly the reverse of that by 
which it had become narrower in its progress towards that point. 
The light near the vertex of the arch again assumes the appear- 
ance of pencils of rays, parallel to the magnetic meridian ; and 
the rays near its eastern and western extremities no longer main- 
tain their parallelism with the line of the zone, but again form 
an angle with it, which gradually increases as the zone gets 
farther south ; for they still direct themselves towards the 
zenith, or a point within the limits of 10° to the southward of 
it. This enlargement and gradual change of appearance of the 
luminous space have always gone on continuously, so long as it 
has been visible in its progress towards the south. But in the 
observations which I have yet made, the meteor never reached 
above 25° or 30° to the southward of the zenith ; having become 
gradually indistinct; and having vanished entirely before pass- 
ing that limit. I have, indeed, seen the meteor near the south- 
ern horizon; but it is of extremely rare occurrence in that 
quarter ; and I cannot say whether, in any case of that descrip- 
tion, it had travelled from the north ; never having observed any 
of them sufficiently early, to be able to determine that point. I 
can only say that the bundles of rays are vertical, or nearly so, 
in the south, as they are in the north. 
Such is the order of appearances presented by the aurora 
borealis, when it has been observed under the most favourable 
circumstances. It is very seldom, however, that all the succes- 
sive phenomena now described have been observed continuously 
on the same evening ; but those observed at any one particular 
time, have always been entirely consistent with the above descrip- 
tion ; and I shall now enter a little into a detail of the varieties 
which present themselves. 
It very frequently happens that the twilight appearance on 
the northern horizon is all that is visible ; and the phenomenon 
begins and ends with that. In this case, the meteor is seldom 
of long continuance; but, during the time that it lasts, the lu- 
minous space gradually enlarges itself towards the south. It 
then gradually disappears, frequently, to be succeeded by an- 
other, appearing low on the horizon, to enlarge, and afterwards 
disappear, as its predecessor had done. 
It happens also very frequently that the meteor, even when 
u 2 
