148 ANNUAL REPORT SMITHSONIAN INSTITUTION, 19 3 5 



the stars in our immediate neighborhood, but for more distant stars 

 it is superior since its accuracy is independent of the distance. Thus 

 the two methods supplement each other in a most valuable way, and 

 between the two we have acquired a knowledge of the luminosities 

 and distances of about 6,000 individual stars. 



When we begin to consider these results, we are led to some very 

 interesting conclusions. In the first place, we find that the apparent 

 brightness of a star as we see it may bear very little relationship to 

 its real brightness or luminosity. Sirius, apparently the brightest 

 star in the sky, is comparatively near us and gives out about 25 times 

 as much light as our sun; Canopus, the second brightest star, is 

 very far away and is almost certainly 10,000 times as luminous as 

 the sun. Similarly Procyon, one of the brightest of the stars in the 

 winter constellations, has 5 times the luminosity of our sun, while 

 Rigel, the brightest star in Orion and of nearly the same apparent 

 brightness as Procyon, gives out from 10,000 to 15,000 times as much 

 light as the sun. The color of Rigel is bluish white, and its tempera- 

 ture is very high, so that its surface brightness is great. On the 

 other hand, Betelgeuse, the other chief star in Orion, is red and has 

 a low temperature and surface brightness; its diameter, however, is 

 so enormous — over 200,000,000 miles — that its luminosity is nearly 

 1,500 times that of the sun. 



The contrast between the luminosities of these giant stars and the 

 faint dwarf stars is very great. The faintest star intrinsically of 

 which we have any knowledge is the small companion of the nearest 

 star in tlie sky, a Centauri. This star has a distance of 4.3 light- 

 years, and its luminosity is 0.00006 of that of the sun. About a 

 dozen stars are known the luminosity of which is less than 0.0001 

 part of that of the sun. So we find that among the stars already 

 studied the luminosity or candlepower varies through a range of 

 at least 200,000,000. This factor would be multiplied at least a 

 thousandfold if we were to include the brightest of the new or 

 temporary stars which suddenly blaze out and die away within a 

 few days or weeks. The luminosities of some of these stars must 

 be at least a million times that of our sun. So we find that the Bib- 

 lical statement that "one star differs from another star in glory" 

 is even more true of the stars as they really are than as they 

 appeared to the eyes of the shepherds of Palestine. 



A very remarkable result is found when the stars are grouped 

 according to spectral type, or surface temperature, and their true 

 luminosities. (Fig. 5.) The resulting diagram resembles a reversed 

 7, with the faint low-temperature stars lying along the stem of the 

 figure and the bright low-temperature stars along the upper horizon- 

 tal bar. Between the two there is a wide gap in which few or no 



