288 ANNUAL, REPORT SMITHSONIAN INSTITUTION, 1920. 



bromine, and iodine). It is also seen in group 8, in the case of iron, 

 nickel, and cobalt, according to their atomic weights, though the 

 atomic numbers of nickel and cobalt are reversed in order. As Clarke 

 says : " We are dealing with an evident tendency of which the mean- 

 ing is yet to be discovered. That the abundance and associations of 

 the elements are connected with their position in the periodic system 

 seems, however, to be clear. The coincidences are many, the exceptions 

 are comparatively few." 



The relation of the abundance of the elements to the periodic law 

 has also been discussed recently by Harkins, 17 who holds that the 

 abundance of the elements is " related to the atomic number and not to 

 the periodic system," that the abundant elements are those of low 

 atomic weight with an atomic number less than 29, and that the ele- 

 ments with even-numbered atomic numbers surpass in abundance the 

 odd-numbered. It would take us too far from our proper subject to 

 discuss this very interesting topic here, but we may examine briefly a 

 hitherto unrecognized phase of the relation of the occurrence of the 

 elements to their position in the periodic table, shown b}^ the study of 

 minerals and of igneous rocks, and taking into consideration the 

 chemical relations of the various elements. 



THE PETROGENIC AND METALLOGENIC ELEMENTS. 



In Table III is presented the periodic classification of the elements, 

 as usually given. 18 The atomic weights are stated in round numbers, 

 most of the elements of the " rare earths " are omitted, as their rela- 

 tive positions are still in dispute, as are also the radio-active elements, 

 except radium. 



Here, as has been commonly recognized, the most abundant ele- 

 ments in the earth's crust, being of generally low atomic weight (or 

 atomic number), occupy the upper part of the scheme, forming the 

 series 1 to 4 of groups 1 to 8. These, with some others in series 6 

 and 8 to be mentioned presently, may be called the " rock elements," 

 as they are the essential elements, in greater or less amount, of the 

 igneous rocks of the earth's crust, of which they constitute at least 

 99.9 per cent by weight. 



In the lower part of the scheme are elements of higher atomic 

 weight which, with others, in series 5 and 7, to be mentioned later, 

 are but seldom, if ever, found in determinable quantities in igneous 

 rocks, but which occur chiefly as ores, or as native metals. These 

 may therefore be called the " ore elements." There would seem to 



17 Harkins, W. D., Jour. Amer. Chem. Soc, xxxix, p. 856, 1917. 



18 This table is based on that given by Clarke, TJ. S. Geol. Survey, Bull. 695, p. 37. 



