cleus to form the coma, and then gradually repelled
from the sun to form the tail. If this process went on indefinitely
every comet would, in the course of ages, be entirely dissipated. This
result has actually happened in the case of some known comets, the best
established example of which is that of Biela, in which the process of
disintegration was clearly followed. As the amount of matter lost by a
comet at any one return cannot be estimated, and may be very small, it
is impossible to set any limit to the period during which its life may
continue. It is still an unsettled question whether, in every case, the
evaporation will ultimately cease, leaving a residuum as permanent as
any other mass of matter.

The next question in logical order is one of great difficulty. It is
whether the nucleus of a comet is an opaque solid body, a cluster of
such bodies, or a mass of particles of extreme tenuity. Some light is
thrown on this and other questions by the spectroscope. This instrument
shows in the spectrum of nearly every comet three bright bands,
recognized as those of hydrocarbons. The obvious conclusion is that the
light forming these bands is not reflected sunlight, but light radiated
by the gaseous hydrocarbons. Since a gas at so great a distance from the
sun cannot be heated to incandescence, the question arises how
incandescence is excited. The generalizations of recent years growing
out of the phenomena of radioactivity make it highly probable that the
source is to be found in some form of electrical excitation, produced by
electrons or other corpuscles thrown out by the sun. The resemblance of
the cometary spectrum to the spectrum of hydrocarbons in the Geissler
tube lends great plausibility to this view. It is remarkable that the
great comet of 1882 also showed the bright lines of sodium with such
intensity that they were observed in daylight by R. Copeland and W. O.
Lohse. In addition to these gaseous spectra, all but the fainter comets
show a continuous spectrum, crossed by the Fraunhofer lines, which is
doubtless due to reflected sunlight. It happens that, since the
spectroscope has been perfected, no comet of great brilliancy has been
favourably situated for observation. Until the opportunity is offered,
the conclusions to be derived from spectroscopic observation cannot be
further extended.

PLATE I.

[Illustration: FIG. 1.--COMET 1892, I. (SWIFT), 1892, APRIL 26.
By permission of Lick Observatory (E