VARIABLE STARS
181
/
maximum velocity of approach. This fact rules out any interpretation
of the variation as an occultation effect.
The absolute magnitude is a definite function of the period. This was
first shown by Miss Leavitt from a study of the variables in the Lesser
Magellanic Cloud. A full confirmation was obtained by H. Shapley from
the variables in globular clusters. In the same cluster the absolute mag
nitude differs from the apparent magnitude by a constant (depending on
the unknown distance of the cluster) so that the period-luminosity re
lation is given directly without the intervention of parallax. It is found
that the period determines the absolute magnitude to within a probable
error of ± 0 m -25. Having thus found the period-magnitude curve applicable
to all Cepheids except for the unknown constant, we proceed to anchor
the curve by combining our knowledge of the mean luminosity of the
nearer Cepheids (derived from their parallactic and cross motions) into
a single mean determination of the constant.
There is a progression of spectral type in the direction from A towards
M as the period (and luminosity) increases.
The Cepheids are more luminous than ordinary giant stars of the same
spectral class, although some giant stars of high luminosity, called pseudo-
Cepheids, are found which seem to resemble them very closely without
showing any light-variation. Cepheids and pseudo-Cepheids are sometimes
described as “super-giants.”
124. Tables 24 and 25 contain results for those Cepheids which have
been sufficiently investigated. The observational data have been taken
from a compilation by Margarete Güssow*.
In Table 24, column 3 gives for most stars the range of spectral type
since the type changes during the light-period. In column 4 an effective
temperature is assigned to correspond to the median spectral type. The
basis adopted in this assignment is 4900° for type G 0 with an increase of
log 10 T e by 0-0140 for each tenth of a type between M and A, so that
A 0 = 9300°, M 0 = 2600°. The temperatures are taken rather low, partly
because these stars are super-giants of low density, and partly because the
types here used (mostly due to Shapleyf) are systematically ^ or T % of a
type bluer than those assigned by Adams and Joy. The absolute visual
magnitude in column 6 is derived from the period by Shapley’s period-
luminosity curvej. Differentially these magnitudes should be correct to
within 0 m -25, but the zero point of the period-luminosity curve is not so
well determined, and there may be a constant correction applicable to the
whole series. I suspect also that there may be a progressive error (originat-
* “Kritische Zusammenstellung sämtlicher Beobachtungsergebnisse der Verän
derlichen vom S Cephei-Typus und Kritik der Eddingtonschen Pulsationstheorie”
(lithographed, Berlin, 1924).
f Astrophys. Journ. 44 , p. 274. t Ibid. 48 , pp. 114, 282.
