252 
RELATIONS BETWEEN THE SPECTRA AND 
OTHER CHARACTERISTICS OF THE 
STARS.* 7 
Ue 
Brightness and Spectral Class. 
Eyes thus made a rapid survey of the general 
field, I shall now ask your attention in greater 
detail to certain relations which have been the more 
special objects of my study. 
Let us begin with the rela- 
tions between the spectra and 
the real brightness of the 
stars. These have been dis- 
cussed by many investigators 
—notably by Kapteyn and 
Hertzsprung—and many of 
the facts which will be 
brought before you are not 
new; but the observational 
material here presented is, | 
believe, much more extensive 
than has hitherto been 
assembled. We _ can_ only 
determine the real brightness 
of 2 star when we know its 
distance; but the recent 
accumulation of direct 
measures of parallax, and the 
discovery of several moving 
clusters of stars the distances 
of which can be determined, 
put at our disposal far more 
extensive data than were 
available a few years ago. 
Fig. 1 shows graphically 
the results derived from all 
the direct measures of 
parallax available in the 
spring of 1913 (when the 
diagram was constructed). 
The spectral class appears as 
the horizontal coordinate, 
while the vertical one is the 
absolute magnitude, accord- 
ing to Kapteyn’s definition— 
that is, the visual magnitude 
which each star would appear 
to have if it should be 
brought up to a standard dis- 
tance, corresponding to a 
NATURE 
[May 7, 1914 
dots, representing the results derived from the poor 
parallaxes, should scarcely be used as a basis for any 
argument. The solid black dots represent stars the 
parallaxes of which depend on the mean of two or 
more determinations; the open circles, those observed 
but once. In the latter case, only the results of those 
observers whose work appears to be nearly free from 
systematic error have been included, and in all cases 
the observed parallaxes have been corrected for the 
probable mean parallax of the comparison stars to 
parallax of o-1” (no account 
being taken of any possible 
absorption of light in space). 
The absolute magnitude, —5, 
at the top of the diagram, 
corresponds to a_ luminosity 
7500. times that of the sun, 
the absolute magnitude of 
which is 4-7. The absolute 
magnitude 14, at the bottom, 
corresponds to 1/5000 of the 
sun’s luminosity. The larger 
dots denote the stars for 
which the computed probable error of the 
parallax is less than 42 per cent. of the parallax itself, 
so that the probable error of the resulting absolute 
magnitude is less than +1-om, This is a fairly 
tolerant criterion for a ‘‘ good parallax,’’ and the small 
* An address delivered before a joint meeting of the Astronomical and 
Astrophysical Society of America and Section A of the American Association 
for the Advancement of Science, at Atlanta, Georgia, December 30, 1913, 
with a few additions, by Prof. H. N. Rus-ell. Continued from p. 230. 
NO.) 22235 ViOL.193)| 
teste 
which they were referred. The large open circles in 
the upper part of the diagram represent mean results 
| for numerous bright stars of small proper-motion 
(about 120 altogether) the observed parallaxes of which 
scarcely exceed their probable errors. In this casé the 
best thing to do is to take means of the observed 
parallaxes and magnitudes for suitable groups of stars, 
| and then calculate the absolute magnitudes of the 
typical stars thus defined. These will not exactly 
