The Sizes and Masses of the Stars 69
atmosphere. The deeper layers are undoubtedly hotter
than those nearer the surface, and it is therefore not sur-
prising that at the centre of the Sun’s disk, where we look
straight down into its atmosphere, and probably see
deeper into it, the light is brighter and bluer, correspond-
ing to a temperature about 400o higher than calculated
for the integrated light of the whole Sun.
When we attempt to estimate the temperatures of the
stars, our first reliance must be upon the method depend-
ing upon the color of their light—for, without a know-
ledge of their diameters, we cannot directly determine
the surface brightness, nor the emission of energy per
square centimeter. Extensive studies of the distribution
of energy in stellar spectra—which is a more precise
specification of the color of their light—have been made
by Wilsing and Scheiner at Potsdam, and by other ob-
servers. Their results are in good general agreement,
and indicate that the surface temperature of the M-stars
is about 3000° or a little higher; of the К-stars 4000°,
of the G’s 5400°—agreeing substantially with the color-
temperature of the Sun—; for Class F about 7000°, and
Class A, 11,000°. The typical stars of Class B must be
exceedingly hot, but it is very difficult to determine their
temperatures, as a small error in the observed colors makes
in this case a large one in the temperature. From all the
evidence, it may be estimated that in some cases their
temperature reaches 20,000°, or perhaps more. The ob-
served color indices of the stars agree excellently with
those which would be exhibited by perfect radiators at
these temperatures—which is not surprising, because they
depend on observation, in another manner, of the very
characteristics by means of which the temperatures were
calculated.