that, if there be no physical connection between the stars of a pair,
the largest is the nearest, and has the greatest parallax. So, by
noting the distance between the pair at different times of the year, a
delicate test of parallax is provided, unaffected by major
instrumental errors.

Herschel did, indeed, discover changes of distance, but not of the
character to indicate parallax. Following this by further observation,
he found that the motions were not uniform nor rectilinear, and by a
clear analysis of the movements he established the remarkable and
wholly unexpected fact that in all these cases the motion is due to a
revolution about their common centre of gravity.[11] He gave the
approximate period of revolution of some of these: Castor, 342 years;
delta Serpentis, 375 years; gamma Leonis, 1,200 years; epsilon Bootis,
1,681 years.

Twenty years later Sir John Herschel and Sir James South, after
re-examination of these stars, confirmed[12] and extended the results,
one pair of Coronae having in the interval completed more than a whole
revolution.

It is, then, to Sir William Herschel that we owe the extension of the
law of gravitation, beyond the limits of the solar system, to the
whole universe. His observations were confirmed by F.G.W. Struve (born
1793, died 1864), who carried on the work at Dorpat. But it was first
to Savary,[13] and later to Encke and Sir John Herschel, that we owe
the computation of the elliptic elements of these stars; also the
resulting identification of their law of force with Newton's force of
gravitation applied to the solar system, and the force that makes an
apple fall to the ground. As Grant well says in his _History_: