204 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1950 



high aifinities for oxygen (i. e., sodium, potassium, strontium, barium) 

 exist almost entirely in the silicate phase. A survey of the earth's 

 crust demonstrates that elements of low oxygen affinity, such as the 

 platinum metals, exist in exceedingly low concentrations when com- 

 pared with neighboring elements of high oxygen affinity. If we 

 assume that a metallic phase exists within the earth, the low abun- 

 dances of these elements in the crust can be explained on the basis that 

 they exist in considerably higher abundance in the deep-seated regions 

 of the earth in association with metallic iron. 



With these general ideas in mind Goldschmidt, together with several 

 chemists who had become interested in the problem, determined the 

 concentrations of many elements in iron and stony meteorites. Gold- 

 schmidt then utilized the data in 1937 to compile the first fairly 

 complete table of relative abundances. 



One difficulty associated with the compilation of such an abundance 

 table was that of coupling meteoritic abundances to solar abundances. 

 Goldschmidt utilized the then existent solar data and adjusted meteor- 

 itic silicon so that it would be equal to solar silicon. A second difficulty 

 was that of combining iron meteorites with stone meteorites in proper 

 proportions. Unfortunately iron and stone meteorites fall through 

 the atmosphere in different ways, with the result that there is a higher 

 probabilit}'^ of observing a stone fall than an iron fall. Stony mete- 

 orites tend to break into fragments while passing through the atmos- 

 phere, thus producing more spectacular displays than do iron mete- 

 orites. On the other hand, many meteorites reach the ground without 

 actually having been seen to fall. As stony meteorites appear to the 

 untrained eye to be rocks, many of them are never collected. Iron 

 meteorites, being more unusual, are picked up more frequently. 



Goldschmidt, in the absence of adequate information, chose a ratio 

 of metal to silicate of 1 : 5 ; but observations of both the eartli and the 

 sun lead us to believe that perhaps it should be closer to 3 : 5. 



An approximate figure for the ratio of metal to silicate can be ob- 

 tained by calculating the weight of the earth's core relative to the 

 earth as a whole. The core of the earth is, of course, compressed owing 

 to the tremendous pressures in the interior ; at the center, the pressure 

 is approximately 3 million atmospheres. We do not know experimen- 

 tally the compressibility of iron at such high pressures, but in recent 

 years a number of theoretical studies have been made. Utilizing the 

 Fermi-Thomas statistical atomic model, calculations of potential fields 

 and charge densities in metals as a function of lattice spacing can be 

 made. Such calculations make possible the determination of pressure- 

 volume relationships at extremely high pressures. It has been found 

 that the results of such a calculation made on iron are compatible 

 with estimates of the densities within the earth's core derived from 



