But our interest in cometary orbits has been added to by the discovery
that, owing to the causes just cited, a comet, if it does not separate
into discrete parts like Biela's, must in time have its parts spread
out so as to cover a sensible part of the orbit, and that, when the
earth passes through such part of a comet's orbit, a meteor shower is
the result.

A magnificent meteor shower was seen in America on November 12th-13th,
1833, when the paths of the meteors all seemed to radiate from a point
in the constellation Leo. A similar display had been witnessed in
Mexico by Humboldt and Bonpland on November 12th, 1799. H. A. Newton
traced such records back to October 13th, A.D. 902. The orbital motion
of a cloud or stream of small particles was indicated. The period
favoured by H. A. Newton was 354-1/2 days; another suggestion was 375-1/2
days, and another 33-1/4 years. He noticed that the advance of the date
of the shower between 902 and 1833, at the rate of one day in seventy
years, meant a progression of the node of the orbit. Adams undertook
to calculate what the amount would be on all the five suppositions
that had been made about the period. After a laborious work, he found
that none gave one day in seventy years except the 33-1/4-year period,
which did so exactly. H. A. Newton predicted a return of the shower on
the night of November 13th-14th, 1866. He is now dead; but many of us
are alive to recall the wonder and enthusiasm with which we saw this
prediction being fulfilled by the grandest display of meteors ever
seen by anyone now alive.

The _progression_ of the nodes proved the path of the meteor
stream to be retrograde. The _radiant_ had almost the exact
longitude of the point towards which the earth was moving. This proved