ground, and is transformed into an aurora borealis. The cirro-cumulus
and the hazes become luminous when they are traversed by sufficiently
energetic discharges of electricity, and when the light of day is no
longer present to overcome their more feeble light. Dr. Usher describes
an aurora borealis seen in the open day, at noon, May 24, 1778.

MM. Cornulier and Verdier are convinced, after carefully studying the
subject, that there are almost always aurorae boreales in the high
polar latitudes, and that their brilliancy alone is variable. This
conviction is in accordance with the very careful observations which
have now been made for four years in the northern hemisphere. It
appears, as the result of these, that the aurora borealis is visible
almost every clear night, but it does not show itself at all the
stations at the same time. From October to March there is scarcely a
night in which it may not be seen; but it is in February that it is
most brilliant. In 1850 it was observed two hundred and sixty-one
nights, and during 1851 two hundred and seven. The proportion of nights
in which the aurora is seen is much greater the nearer we are to the
magnetic pole.

De la Rive, from whose admirable treatise upon Electricity we have
borrowed our general views, and whose theory we have attempted to
illustrate in this paper, concludes that the aurora borealis is a
phenomenon which has its seat in the atmosphere, and consists in the
production of a luminous ring of greater or less diameter, having for
its centre the magnetic pole. Experiment shows, as we have seen, that,
on bringing about in rarefied air the reunion of the two electricities,
near the pole of a powerful artificial magnet, a small luminous ring is
produced, similar to that which constitutes the aurora borealis, and
animated by a similar movement of rotation. The aurora borealis would