required for their
separation.[36] The number of these objects was gradually increased by
fresh discoveries, until in 1781 (the same year in which Herschel
discovered Uranus) a list containing eighty double stars was published
by the astronomer Bode. These interesting objects claimed the attention
of Herschel during his memorable researches. The list of known doubles
rapidly swelled. Herschel's discoveries are to be enumerated by
hundreds, while he also commenced systematic measurements of the
distance by which the stars were separated, and the direction in which
the line joining them pointed. It was these measurements which
ultimately led to one of the most important and instructive of all
Herschel's discoveries. When, in the course of years, his observations
were repeated, Herschel found that in some cases the relative position
of the stars had changed. He was thus led to the discovery that in many
of the double stars the components are so related that they revolve
around each other. Mark the importance of this result. We must remember
that the stars are suns, comparable, it may be, with our sun in
magnitude; so that here we have the astonishing spectacle of pairs of
suns in mutual revolution. There is nothing very surprising in the fact
that movements should be observed, for in all probability every body in
the universe is in motion. It is the particular character of the
movement which is specially interesting and instructive.

It had been imagined that the proximity of the two stars forming a
double must be only accidental. It was thought that amid the vast host
of stars in the heavens it not unfrequently happened that one star was
so nearly behind another (as seen from the earth) that when the two were
viewed in the telescope they produced the effect of a double star. No
doubt many of the so-called double stars are produced in this way.
Herschel's discovery shows that this explanation will not always answer,
but that in many cases we really have two stars close together, and in
motion round their common centre of gravity.

When the measurements of the distances and the positions of double stars
had been accumulated during many years, they were taken over by the
mathematicians to be treated by their methods. There is one peculiarity
about double star observations: they have not--they cannot have--the
accuracy which the computer of an orbit demands. If the distance between
the pair of stars forming a binary be four