ASTRONOMY: H. SHAPLEY 277 
angles of direction from the center (expressed as position angles), the 
data for opposite sectors being combined into means. Compared with 
this dotted curve it is apparent that the thousand brightest stars, rep- 
resented by the full line, show Httle trace of the elliptical distribution. 
Of this group of bright stars, 130 have negative color indices. They 
correspond, at least approximately, to the B-class stars in our galactic 
system, which show marked galactic concentration. Their distribution 
is plotted in figure 1 as a broken line. Though the accidental varia- 
tions are influential, because of the relatively small numbers concerned, 
the elongation is definitely shown and the degree of concentration ap- 
pears to be about double that for the faint stars of all color classes. 
0° 30° 60° 90° 120° 150° 0° 30° 60° 90° 
+ 40% 
+ 20 
0 
-20 
-40 
Full line, brightest thousand stars; broken line with crosses, blue stars; dotted line, all 
stars between magnitudes 17 and 19. Ordinates, percentage deviation from the mean; 
abscissae, angle of direction from the center. 
There are seven variable stars in the Hercules cluster. Four of them, 
including the two that are known certainly to be Cepheid variables, 
are in the sectors that contain the axis of symmetry, and one is in an 
adjoining sector. As some allowance for the inclination of the supposed 
plane to the line of sight must be made, this proportion is as much as 
could be expected, even if all these are Cepheids and in their distribu- 
tion a strict comparabiHty to the condition in our system is required. 
The elliptical form of cu Centauri, the largest and brightest globular 
cluster in the sky, is immediately evident upon photographs that show 
several thousands of its stars. The apparent elongation was mentioned 
by Bailey in discussing the variable stars ,^ but was not considered in 
his discussion of the distribution.^ That the star counts, as published 
by Bailey, verify the appearance of the photographs and show a dis- 
tinctly elliptical symmetry, has been noted in earlier papers.^ 
The data upon which this conclusion was based are now given in the 
table, the thousand stars nearer the center than 3' being excluded be- 
cause of the difficulty and consequent uncertainty in counting and ar- 
ranging the data in sectors. The successive lines of the table show that 
