72 Recent Advances in Stellar Astronomy
affected by the presence of heavy dark bands in the green,
yellow and red. In this way it may be calculated that
the surface temperature of the one N-star which was
observed by Coblentz is about 2500°, and that of Alpha
Herculis but a little more than 20000.
With our estimates of temperature thus confirmed, we
may proceed to calculate the relative surface brightness
of the stars. Taking the Sun as typical of Class G, we
find that a star of the same size, but of Class B0, should
be perhaps thirty times as bright, one of Class AO ten
or twelve times as bright, one of Class FO about three
times. The surface brightness of a KO star should be
about 1/7 of the Sun’s and that of an М-star about 1/50.
All these values, of course, are averages, and individual
stars may sometimes run somewhat above or below them,
especially at the extremes of the spectral scale. As an
extreme example (to judge by its luminous efficiency),
Alpha Herculis may have a surface brightness as low as
1/500 that of the Sun.
We are now in a position to form an idea of the real
sizes of the stars. Take, for example, Sirius, which is
25 times as bright as the Sun, and of Class A. Taking
its surface brightness as ten times the Sun’s, its super-
ficial area comes out 2½ times the Sun’s and its diameter
60 per cent, greater than that of the Sun, or 1,200,000
miles. Procyon, of Class F5, six times as bright as the
Sun, and probably twice as bright per unit of surface,
must be nearly equal to Sirius in diameter. Alpha Cen-
tauri, which is almost exactly similar to the Sun in
brightness as well as in spectrum, must be also very nearly
of the Sun’s diameter, 866,000 miles. Passing to redder
stars, we may take 61 Cygni. The components of this
well-known double star, one of our closest neighbors in