198 ANNUAL REPORT SMITHSONIAN INSTITUTION, 195 



whose mass numbers are integer multiples of four are more abundant 

 than other species. It appeared further that the abundances of ele- 

 ments fall off rapidly with increasing atomic number. 



The "odd-even" and the "integer multiple of 4" abundance regulari- 

 ties are in themselves sufficient to permit an important conclusion. 

 Wliatever the process might have been that led to the formation of 

 elements, it seems clear that the elements were formed in relative 

 amounts which depended, at least in part, upon the nuclear properties 

 of their constituent isotopes. But if one is to theorize further on 

 the question of the origin of the elements and the relationships between 

 abundance and nuclear composition, it is important that we obtain 

 more quantitative data than is obtainable through a study of the 

 earth's crust alone, which could not be expected to be a representative 

 sample of matter in our cosmos. 



THE COMPOSITION OF STARS 



There are numerous dark lines or "absorption lines" in the sun's 

 spectrum which are found to be located at the same characteristic 

 frequencies which are observed in the emission spectra of elements 

 studied in the laboratories of our earth. In this manner many ele- 

 ments which exist in the earth's crust have been identified in the sun's 

 atmosphere. No elements have been found in the sun's atmosphere 

 which do not exist on earth, though helium was discovered in the sun 

 before it was isolated and identified terrestrially. The spectra of 

 other stars are similar in nature to the sun's. Characteristic dark 

 absorption lines are observed, leading to the conclusion that the stars 

 are similar to our sun in general structure and composition. 



The fact that absorption lines are observed demonstrates that the 

 continuous radiation emitted from a star's surface passes through a 

 layer of relatively cool gas surrounding the star, the reversing layer, 

 and so the elements in stellar atmospheres can be positively identified. 

 But the task of converting the intensities of the lines observed in 

 stellar spectra into relative numbers of atoms of the various species 

 which exist in stellar atmospheres is most difficult. 



Thanks to the herculean efforts of early workers in the field, such as 

 Henry Norris Russell, C. H. Payne, and C. E. Moore, and recent 

 developments by astrophysicists such as A. Unsold, B. Stromgren, 

 D. Menzel, L. Aller, and J. Greenstein, quantitative conversion of 

 spectral intensities into relative numbers of atoms is now possible. 

 The theory which permits the conversion of spectral line intensities 

 into relative numbers of atoms is straightforward, but exceedingly 

 complex, and need not be discussed here in any detail. It involves 

 a detailed knowledge of the quantum-mechanical behavior of atomic 

 species as functions of temperature and density, and in the presence 



