96 
ASTRONOMY: PEASE AND SHAPLEY 
AXES OF SYMMETRY IN GLOBULAR CLUSTERS 
By Francis G. Pease and Harlow Shapley 
MOUNT WILSON SOLAR OBSERVATORY, CARNEGIE INSTITUTION OF WASHINGTON 
Communicated by G. E. Hale, January 10, 1917 
The distribution of stars in globular clusters has been extensively 
studied in the past, chiefly from the standpoint of the groupings and 
superficial arrangements of the brightest stars or on the basis of the 
variation of density with distance from the center. ^ The distribution 
with respect to direction as well as distance has been considered briefly 
for ten clusters by Bailey, who found a number of asymmetrical irreg- 
ularities among the brighter stars.^ But deductions relative to the basic 
structure of globular clusters cannot be founded safely on the most 
luminous objects alone, for investigations of the colors in these systems 
show that the brighter stars are peculiar and perhaps not at all repre- 
sentative of the enormous number of faint stars which constitute the 
greater part of the mass. 
Several points wherein globular clusters resemble our galactic system 
of stars have been noted in recent papers, and it was suggested that 
planes of symmetry may characterize the distribution of stars in clusters, 
as does the Galaxy that of the stars in general.^ Although the thous- 
and brightest stars in Messier 13 gave no definite evidence of an ellip- 
tical distribution, a study with respect to direction from the center of 
the star-counts published for the bright southern cluster co Centauri* 
did verify the photographic appearance of an elongated form. In this 
cluster 6400 stars were counted, covering a considerable interval of 
apparent (and absolute) magnitude, while for only two of the ten clus- 
ters discussed by Bailey did the number appreciably exceed one thousand. 
The photographs of clusters with the 60-inch reflector at Mount Wil- 
son afford material particularly suited to the study of this problem. 
Plates have been obtained for nearly twenty different systems, and for 
some clusters exposures varying in length from one minute to several 
hours are available. A few plates record stars fainter than the twentieth 
magnitude. The method and progress of the counting have previously 
been reported.^ Briefly, the study of star density on each plate is 
based on counts made with the aid of a superposed reseau of small 
squares 31 ''.4 on a side. The details of the observations and counts 
will appear eventually in the Contributions of the Mount Wilson Solar 
Observatory. The present paper indicates one of the most important 
results, considered somewhat in detail for the well-known Hercules 
