Voiv. 6, 1920 
ASTRONOMY: H. SHAPLEY 
297 
It should be especially noted that only six stars (four of those above and 
two in table I) are brighter than — 4.0, and the brightest recorded star is 
fainter than — 5.5. We are able, moreover, to go much farther with this 
inference relative to an upper limit for stellar luminosity. The investiga- 
tions carried on at Mount Wilson during the last six years have yielded 
provisional but sufficiently reliable absolute magnitudes in more than 
fifty clusters ; they permit the statement that, of more than a million stars, 
less than a tenth of 1% are brighter than —4, photovisually. Since^the 
brightest giants are red, the corresponding upper limit for absolute photo- 
FIG. 4 - 
Luminosity curves for different intervals of color index (Table II). Ordinates are 
numbers of stars; abscissae are absolute photovisual magnitudes. 
graphic magnitude may be placed at —3. There appears to be scarcely 
a single star brighter than photovisual magnitude —6. 
The existence of a maximum luminosity, indicated by these results, is 
probably closely related to the upper limits of stellar mass suggested by 
Kddington's researches on giant stars ; it is highly significant in considera- 
tions of external galactic systems. 
It is also of interest that of the 1152 stars only three may be considered 
slightly abnormal in color, in that they fall outside the limits of table I. 
This speaks for the uniformity of stars throughout the universe, and 
