The Light of the Stars 51
other at regular intervals. In some such cases two stars,
separated by hardly more than their own diameter, have
spectra differing as widely as A and K. The stars of
such a pair have doubtless been formed out of the same
original mass, and it is quite incredible that practically
all the hydrogen has segregated in one component, and
practically all the metals in the other.
No: we must look to some difference in the physical
conditions of the stars’ atmospheres to explain the dif-
ferences in their spectra, and to differences in some one
important particular. There is no doubt at present that
this condition is the temperature of the atmosphere and
of the star’s surface below it. Practically sufficient evi-
dence of this is found in the colors of the stars. Those
which are “early” in the spectral sequence—of Classes O,
B and A—are conspicuously white or bluish; stars of
Class F are yellowish; and of Class G yellow, like the
Sun; those of Class K reddish, and of Class M red—as
are also those of Class R, while the N-stars are the red-
dest in the sky, and the only ones which are found be-
yond the limits of color mentioned earlier, which corre-
spond approximately to Classes B and M.
It has long been realized that the behavior of the line-
absorption in stellar spectra is in general agreement with
this hypothesis, on which the В-stars are the hottest, and
the М-stars the coldest of those in the main sequence.
But the full interpretation of the results has come only
very recently, since the processes by which spectra are
produced have begun to be understood. The great astro-
physical importance of these considerations was pointed
out, a year or two ago, by Dr. Saha, Professor at Cal-
cutta, and much of what follows is borrowed from his
discussions.