of the
earth, it was found that there were still small outstanding differences
which must be due to the motion of the star itself--its proper motion.
The quantity of this motion was not well known for any star when
HERSCHEL'S researches began. Before they were concluded, however,
MASKELYNE had deduced the proper motions of thirty-six stars--the
fundamental stars, so called--which included in their number _Sirius_,
_Procyon_, _Arcturus_, and generally the brightest stars.

It is _a priori_ evident that stars, in general, must have proper
motions, when once we admit the universality of gravitation. That any
fixed star should be entirely at rest would require that the attractions
on all sides of it should be exactly balanced. Any change in the
position of this star would break up this balance, and thus, in general,
it follows that stars must be in motion, since all of them cannot occupy
such a critical position as has to be assumed. If but one fixed star is
in motion, this affects all the rest, and we cannot doubt but that
every star, our sun included, is in motion by an amount which varies
from small to great. If the sun alone had a motion, and the other stars
were at rest, the consequence of this would be that all the fixed stars
would appear to be retreating _en masse_ from that point in the sky
towards which we were moving. Those nearest us would move more rapidly,
those more distant less so. And in the same way, the stars from which
the solar system was receding would seem to be approaching each other.
If the stars, instead of being quite at rest, as just supposed, had
motions proper to themselves, then we should have a double complexity.
They would still appear to an observer in the solar system to have
motions, and part of these motions would be truly proper to the stars,
and part would be due to the advance of the sun itself in space.

Observations can show us only the _resultant_ of these two motions.
It is for reasoning to separate this resultant into its two components.
At first the question is to determine whether the results of observation
indicate any solar motion at all. If there is none, the proper motions
of stars will be directed along all possible lines. If the sun does
truly move, then there will be a general agreement in the resultant
motions of the stars near the ends of the line along which it moves,
while those at the sides, so to speak, will show comparatively less
systematic effect. It is as if