The Sizes and Masses of the Stars 85
Class G, the brighter component of Capella. This star is
70 times as bright as the Sun and has 4.6 times its mass.
Assuming that its surface brightness equals the Sun’s we
obtain a volume 600 as great, and a mean density 1/130
of the Sun’s, or between eight and nine times that of air
under standard conditions. For Arcturus, with a volume
14,000 times the Sun’s 10 times the Sun’s mass would be
a liberal estimate. The corresponding density is 4/5 that
of air. Finally for Antares, with its enormous bulk, even
if we assign the exaggerated and improbable mass of 100
times that of the Sun, the density comes out one-millionth
of the Sun’s, or 1/900 that of ordinary air. The actual
density is probably somewhat less, but more than 1/5
as great.
These values are so extraordinarily small that confirma-
tion is desirable—and the eclipsing variables provide it.
Of those of “later” spectral type, the majority are dense
stars, one or two nearly equaling the Sun, and one having
almost twice the Sun’s density. But in a few cases the
density is very much less. In one case it is about 1/100
that of the Sun, in two others 1/600 and 1/2000, and
finally in one remarkable star, W Crucis, there is clear
evidence of a density only 1/500,000 of the Sun’s, or
about 1/400 that of air. With this independent evidence,
the reality of these low densities is placed beyond doubt.
But when we realize that the mean density of Antares or
W Crucis is less than that of what was once considered a
fairly good vacuum, before air-pumps reached their present
degree of perfection, we can but ask, Can a body of such
low density possibly be opaque, and shine as if it were
solid? We must not forget, however, that even a per-
fectly pure gas inevitably scatters some light away from a
beam which passes through it, owing to the action of the