c reports of early chroniclers, we meet the "double
comet" discovered by Liais at Olinda (Brazil), February 27, 1860, of
which the division appeared recent, and about to be carried
farther.[1241] But a division once established, separation must
continually progress. The periodic times of the fragments will never be
identical; one must drop a little behind the other at each revolution,
until at length they come to travel in remote parts of nearly the same
orbit. Thus the comet predicted by Klinkerfues and discovered by Pogson
had already lagged to the extent of twelve weeks, and we shall meet
instances farther on where the retardation is counted, not by weeks, but
by years. Here original identity emerges only from calculation and
comparison of orbits.

Comets, then, die, as Kepler wrote long ago, _sicut bombyces filo
fundendo_. This certainty, anticipated by Kirkwood in 1861, we have at
least acquired from the discovery of their generative connection with
meteors. Nay, their actual materials become, in smaller or larger
proportions, incorporated with our globe. It is not, indeed, universally
admitted that the ponderous masses of which, according to Daubree's
estimate,[1242] at least 600 fall annually from space upon the earth,
ever formed part of the bodies known to us as comets. Some follow
Tschermak in attributing to aerolites a totally different origin from
that of periodical shooting-stars. That no clear line of demarcation can
be drawn is no valid reason for asserting that no real distinction
exists; and it is certainly remarkable that a meteoric fusillade may be
kept up for hours without a single solid projectile reaching its
destination. It would seem as if the celestial army had been supplied
with blank cartridges. Yet, since a few detonating meteors have been
found to proceed from ascertained radiants of shooting-stars, it is
difficult to suppose that any generic difference separates them.

Their assimilation is further urged--though not with any demonstrative
force--by two instances, the only two on record, of the tangible descent
of an aerolite during the progress of a star-shower. On April 4, 1095,
the Saxon Chronicle informs us that stars fell "so thickly that no man
could count them," and adds that one of them having struck the ground in
France, a bystander "cast water upon it, which was raised in steam with
a great noise of boiling."[1243] And again, on November 27, 1885, while
the skirts of the Andromede