MOUNT WILSON OBSERVATORY. 199 



intricacies of the stellar universe. Before him all attempts to make order out 

 of seeming chaos had been in vain. Double and multiple star-systems, 

 globular clusters, and irregular star-groups moving together in space were 

 known. But the vast mass of stars had yielded no sign of larger relation- 

 ship and the constitution of the Galaxy was a sealed mystery. Kapteyn's great 

 discovery of the two star-streams, which comprise between them a large 

 proportion of all stars whose motions are known, pointed the way that many 

 astronomers have since pursued with success. His carefully devised plan for 

 the intensive study of the stars in selected areas of the sky, toward the realiza- 

 tion of which observatories in all parts of the world have contributed, will be 

 continued, it may be hoped, by his friends and collaborators. We at Mount 

 Wilson, who have profited greatly by Kapteyn's wide vision and wise counsel, 

 and have enjoyed the advantage of his personal friendship, shall be glad to do 

 our full share toward its completion. 



A different mode of approach to the problem of the structure of the universe, 

 pursued at Mount Wilson with marked success, is that of Dr. Shapley, who 

 has made use for this purpose of his photometric studies of the stars in globular 

 clusters. His conclusion that the galactic system is vastly larger than was 

 formerly supposed, though attacked in some quarters, has received substantial 

 support from several recent studies. Shapley's appointment as director of the 

 Harvard Observatory deprives us of another able investigator, but we shall 

 hope to continue to cooperate with him in the study of stellar problems. 



Turning now to the results of the year, we may first refer to the extensive 

 investigation by Mr. Scares on the masses and densities of the stars. Our 

 knowledge of absolute magnitudes has been greatly extended by the spectro- 

 scopic results of Adams and his associates. With these and other data, 

 especially those for visual binaries, a progressive change of absolute magnitude 

 with spectral type is found, ranging from —1.6 for BO through A, F, G, 

 and K dwarfs to -f 9.8 for Ma. For the same types the geometric mean 

 masses for single stars, derived from the hypothetical absolute magnitudes of 

 binaries, vary progressively from 10 to 0.6, respectively, in terms of the sun's 

 mass. The small dispersion in mass among visual binaries of known parallax 

 is noteworthy, and, as Russell has pointed out, may be a consequence of the 

 use of the spectroscopic method in which line-intensity perhaps depends wholly 

 upon the temperature and density in a star's atmosphere (p. 231). By an 

 indirect method, however, Russell has confirmed the conclusion that the dis- 

 persion in mass is small. 



The densities, effective temperatures, surface brightness, and volumes of 

 both giant and dwarf stars have been derived by Scares from a discussion of 

 this material. The later-type dwarfs are found to be considerably hotter than 

 giants of the same type and to be about 500,000 times as dense. The surface 

 brightness, checked by Pease's measurements of stellar diameters for a few 

 giants, ranges from -f 4.8 for stars at the lowest temperature to —2.7 magni- 

 tudes at the highest (sun = 0). These results, with some exceptions, support 

 Eddington's theory of the radiative equihbrium of giant stars and the pulsa- 

 tion theory of Cepheids and tend to confirm the empirical proportionality of 

 mass to radius. Changes of temperature and pressure at various points within 

 a star are shown by theoretical computations to be such as to maintain an ap- 

 proximately constant degree of ionization throughout, implying a nearly con- 



