202 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1950 



fair accuracy the ratios of several elements in the interstellar gas. 

 He finds the atomic ratio of hydrogen to sodium to be 5 to 25 X 10®. 

 This value is considerably higher than the corresponding value for 

 the ratio of the abundances of these elements believed to exist in the 

 sun (about 0.7 X 10*') and may indicate that the interstellar gas 

 is deficient in higher atomic weight elements relative to stars. It is 

 noteworthy, however, that Stromgren's value for the titanium-sodium 

 ratio is about 3 X 10"^, which, within experimental error, agrees with 

 the best value for the abundance ratio of the same elements in the 

 solar system (6 X 10"-). 



In general, it appears probable that, with the exceptions of hydro- 

 gen, helium, and lighter elements whose abundances are shifted in 

 stars owing to thermonuclear reactions, the abundances of the ele- 

 ments in interstellar material lie very close to their abundances 

 in stars. 



THE EARTH AND METEORITES 



It has been mentioned that the study of stellar spectra gives rise 

 to abundance values which are in the very best cases precise only to 

 within a factor of 2. In addition, only a few elements can be 

 determined in stars with anything approaching this degree of ac- 

 curacy. If we are to extend our knowledge concerning abundances 

 to a wider range of elements, and with a greater degree of precision, 

 we must examine condensed material within our solar system : planets 

 and meteorites. 



Realizing that the crust of the earth constitutes a poor specimen of 

 gross material within our solar system, the late V. M. Goldschmidt, 

 who perhaps more than any one man can be considered to be the father 

 of modern geochemistry, studied the composition of meteorites. In 

 doing so, he followed the general concept which had originated many 

 decades previously : the average composition of these bodies which fall 

 to the earth from space is probably equivalent to the composition of the 

 earth as a whole. In view of the fact that meteorites, as distinct from 

 stellar spectra, can be analyzed quite precisely, it is important to 

 investigate the validity of Goldschmidt's hypothesis. 



What are the chemical relationships between meteorites and the 

 earth ? How is the earth related chemically to the sun and planets of 

 the solar system ? If we can ascertain these relationships, we will then 

 be in a position to utilize meteorites in an evaluation of elemental 

 abundances. 



A century ago, the scientist Boisse first suggested the possibility 

 assumed by Goldschmidt. Since that time considerable effort has been 

 expended by astronomers, geologists, geophysicists, and geochemists in 

 attempts to develop or to disprove Boisse's speculation. On the whole, 



