276 
ASTRONOMY: H. SHAPLEY 
STUDIES OF MAGNITUDES IN STAR CLUSTERS. VI. THE RELA- 
TION OF BLUE STARS AND VARIABLES TO GALACTIC PLANES 
By Harlow Shapley 
MOUNT WILSON SOLAR OBSERVATORY, CARNEGIE INSTITUTION OF WASHINGTON 
Communicated by G. E. Hale, February 19. 1917 
The possibility of the invariable presence of axes of symmetry in 
globular clusters is attested by the results given in a recent communi- 
cation by Mr. Pease and the present writer.' The observed elliptical 
distribution of stars may denote a prolate spheroidal form; or, as ap- 
pears more probable, it may signify the projection of a discoidal figure, 
in which case the axis of symmetry represents the lines of ntersection 
with the celestial sphere of the central plane of a system analogous in 
form to our own Galaxy. In any case the phenomenon of symmetrical 
elongation may be highly significant in problems of stellar dynamics, 
and as such is deserving of further study. 
One method of investigation is the study in globular clusters of the 
distribution of those objects which show a peculiar arrangement in our 
galactic system with respect to the plane of the Milky Way. It does 
not follow, of course, that they should also show condensation along 
a symmetrical plane in all large stellar systems, but it will be of interest 
and of much importance to know whether or not they do. 
Stars of spectral class B and the longer-period Cepheid variables 
are two types of objects which in the local system show a strong gal- 
actic preference. Data bearing on both of these are available for 
Messier 13 — the only globular cluster for which a detailed study of 
magnitudes and colors has been made.^ Earlier discussions of the dis- 
tribution with respect to the center of the brightest thousand stars 
of Messier 13 failed to show a definite tendency toward symmetrical 
elongation,^ but, from an investigation of some 30,000 faint stars in 
this system, we now know with considerable accuracy the position of 
its axis of symmetry. With the orientation of the density ellipse as 
a starting point, the bright stars have been rediscussed from the 
standpoint of color index. The method of treatment is sufficiently indi- 
cated in the earlier communication relative to axes of symmetry. 
The smooth dotted curve in figure 1 is based upon the count of 
10,000 stars between the seventeenth and nineteenth magnitudes, and is 
inserted merely to show the true position of the axis, and, by means 
the amplitude of the curve, to indicate the degree of the ellipticity. 
The ordinates are the percentage deviations of the number of stars in 
each 30° sector from the mean number for all sectors ; the abscissae are 
