ASTRONOMY: H. SHAPLEY 
481 
gone, implies that the spectra are all Kkewise strictly comparable. 
We conclude, therefore, that in surface brightness and volume, and 
probably in mass, density, and other physical properties, these 110 
stars are almost identical. 
Whether the isolated variables of this so-called cluster type also are 
exactly ahke in luminosity, we have at present no means of knowing, 
other than analogy, because their distances are unknown and their 
apparent magnitudes differ greatly. To investigate further the pos- 
sible generality of a law of constant median magnitude, some special 
studies of the variables in other clusters have been made. The ex- 
tended discussion will appear in the Astrophysical Journal. The results 
are summarized in the following numbered paragraphs. 
1. In Messier 5 Bailey has recently determined the periods and light 
curves of about 70 variables.^ Excluding those that are nearer the 
center than V .2, the results for which must be uncertain because of 
the crowding of images, and omitting also three for which the periods 
exceed a day, the median magnitude of the remaining 61, referred to 
Mount Wilson standards, is 15.26 ±0.01, the average deviation for a 
single star being ±0.075. Considering only the 30 light curves selected 
by Bailey as well-determined, the mean median magnitude is 15.25, 
with an average deviation of ±0.08. 
2. In Messier 15 are 51 known variables,^ but the light curves have 
not yet been determined. On three plates the extreme magnitude 
range of the variables has been measured and found to be: 
Maximum 14.98 Extremes of range 1.22 
Minimum 16.20 Median magnitude 15.59 
Two peculiarly bright stars, suspected of variability, are excluded. 
The range of variation is in very good agreement with the mean ranges 
for Messier 3 and 5, and, although the results are not final, the median 
magnitude is probably correct within a tenth of a magnitude. 
3. In the southern cluster co Centauri three subclasses of cluster 
variables are recognized. Treating each separately we have, for the 
stars whose classification is certain, the results of table 2, which are 
TABLE 2 
SUB CLASS 
NUMBER 
VARIABLES 
MEAN PERIOD 
MAXIMUM 
MAGNITUDE 
RANGE OF 
VARIATION 
MEDIAN 
MAGNITUDE 
AVERAGE 
DEVIATION 
days 
33 
0.586 
12.99 
1.11 
13.55 
±0.09 
b 
15 
0.752 
13.11 
0.87 
13.55 
±0.10 
28 
0.395 
13.33 
0.56 
13.61 
±0.09 
All 
76 
13.57 
±0.10 
