74 Recent Advances in Stellar Astronomy
ures of parallax, appears to be 900 times as bright as
the Sun. With a surface brightness 1/50 that of the
latter, the estimated diameter reaches the enormous value
of 210 million miles. Antares, of similar spectral type,
and 3000 times brighter than the Sun, by the same token
comes out 390 million miles in diameter. These last
values are, however, exceptional. We know of no other
stars which bid fair to rival them, except perhaps Alpha
Herculis, which according to Adams, is 250 times as
bright as the Sun. With our previous estimate of sur-
face brightness, its diameter would be 300 million miles
—but this particular estimate is exceedingly uncertain.
These conclusions are so remarkable that they require
to be tested in every available way. Until within a year
or two, there was only one direct way of finding the dia-
meter of a star, and this could be used in only a few in-
stances. There are a good many stars which show regu-
lar, periodic, variations in their light, which give every
evidence of arising from eclipse—the star being really
double, with components so close together that they hide
one another, partially or completely, during every revolu-
tion. If we have good measures of the brightness of such
a star throughout its changes, we can find the relative
diameters and brightness of the two stars—can indeed
make a complete map or model of the system, except that
we do not know the actual dimensions, that is, on what
scale our model is drawn. But when, in addition, the star
has been observed with the spectroscope, and both com-
ponents are bright enough to be recorded, we may find the
velocity of their orbital motion, and so put a scale of miles
upon our model, and obtain the actual diameters of the
stars. About a dozen such systems have so far been in-
vestigated. All these stars but one are bigger than the