82 Recent Advances in Stellar Astronomy
parallax, and the diameters in seconds of arc of all the
stars in the sky—except perhaps for a few abnormal
objects.
We can be sure then, that, while the dwarf stars are
all much alike in size, the giants are huge affairs—so huge,
in some cases, that their size would be almost incredible
if the evidence were not conclusive. The bulk of these
bodies is amazing. Arcturus—not a large star as giants
go—must have about fourteen thousand times the cubic
content of the Sun, while Antares exceeds the Sun in
volume a hundred million fold ! One naturally asks, Is the
quantity of matter in such a star—its mass—proportion-
ally as great as its bulk? To answer this question in the
case of any star, we must have recourse to gravitational
theory. We can measure the mass of an astronomical body
only by the effects of its attraction upon the motion of
some neighboring body—and this limits our knowledge of
stellar masses to double stars. Fortunately these are so
numerous and so well distributed among the various
spectral classes, both giant and dwarf, that we can be
reasonably sure that they present a fair sample of the
general run. When the stars of a binary pair have com-
pleted a large part of a revolution (or more) while under
observation, their orbit can be computed, and then, if the
parallax is known, the mass of the pair is easily found.
Even if the motion is so slow that only a few degrees of
the orbit have been described in a century, it is still possible
to get valuable information by a statistical process, which
gives values which may be considerably too large or too
small for any individual star, but leads to an average
which will be nearly correct for groups of twenty stars
or more.
Combining these methods, several hundred systems be-