Mat 29j 1914] 



SCIENCE 



793 



data. It is clear, however, that the elements 

 whose lines are faint in the sun are, in gen- 

 eral, present in but very small proportions in 

 the earth's crust. 



It is very remarkable that the correspon- 

 dence of the two lists is so close, in view of 

 the radical differences in the methods of in- 

 vestigation, and the great differences in the 

 relative intensities of the lines in stellar 

 spectra of different t3T)es. Even in the bright- 

 line spectrum of the solar atmosphere, the 

 similarity is by no means as pronounced. 



Out of the first sixteen elements in either 

 list, only one, barium, has an atomic weight 

 exceeding 100, and but one other, strontium, 

 one greater than. 60. The significance of this 

 fact has frequently been discussed by geolo- 

 gists or by astronomers.' In both cases it has 

 been suggested that the heavier elements lie 

 for the most part deep within the body, and 

 out of reach; but Clarke gives good reasons 

 for believing that, even in the earth's interior, 

 the lighter elements are more abundant than 

 the heavier. This suggests that the faintness 

 or absence of the lines of the heavier metals 

 in the solar spectrum may be due largely to 

 the small proportions in which they occur, and 

 some confirmation of this is found in the fact 

 that, of the elements of atomic weight greater 

 than 180, only lead, which is the most abun- 

 dant in the earth's crust, appears at all in the 

 sun. But the rarity of these elements can 

 not be the whole explanation of their absence 

 from the solar spectrum, for although no lines 

 of Os, Ir or Pt occur in it, the stronger lines 

 of the equally rare elements Eu, Eh and Pd 

 (whose atomic weights are about half as great) 

 appear distinctly, though faintly. 



The element which is most disproportion- 

 ately conspicuous in the sun, in comparison 

 with its terrestrial abundance, is cobalt. 

 Nickel too is relatively high on the solar list. 

 This may be partly explained by the great 

 number of lines in the spectra of these ele- 

 ments, which gives them undue weight in a 

 spectroscopic count. It is also worthy of 

 notice that, if 25 per cent, of meteoric iron 



5 Compare Clarke, op. cit., p. 33, and Abbot, 

 "The Sun," pp. 92-94, 253-254. 



were added to a sample of the earth's crust, 

 and the composition of the resulting mixture 

 considered, iron would occupy the first place 

 among the metallic elements, nickel the eighth, 

 and cobalt the eleventh, and the discordance 

 with the solar list would disappear. Of the 

 elements abundant in the earth, and relatively 

 less conspicuous in the sun, silicon apparently 

 approaches the typical non-metallic elements 

 in its behavior, while aluminium has only 

 four lines in the observable region, and is thus 

 handicapped by the spectroscopic method of 

 detection. 



The principal differences in the order of the 

 metallic elements in the two lists are there- 

 fore easily explicable, with one conspicuous 

 exception. Potassium, which is one of the 

 principal constituents of the earth's crust, 

 and is fairly abundant in meteorites, shows as 

 the merest trace, if at all, in the solar spec- 

 trum, although many strong lines, both of the 

 principal and the subordinate series, lie in the 

 observed region. Their absence, in spite of 

 the presence of conspicuous lines of elements 

 which are far less abundant in terrestrial 

 materials, is remarkable, and would seem to 

 demand some special explanation. It is of 

 interest in this connection that potassium, 

 alone among the more common elements, is 

 slightly radio-active. If this indicates that 

 its atoms are relatively unstable, they might 

 break down under solar conditions; but this 

 is a highly speculative consideration. The 

 lines of the more strongly radio-active ele- 

 ments do not appear at all in the solar spec- 

 trum; but this may be accounted for by their 

 extreme rarity (on earth, at least) and their 

 high atomic weights. It should however be 

 mentioned that lithium, which is next to 

 potassium in abundance among the alkali 

 metals, and occurs in sensible proportions in 

 the earth's crust, but, so far as is known, is 

 not radio-active, is also practically absent 

 from the solar spectrum — though Adams' 

 points out that a very faint line, greatly 

 strengthened in sun-spots, at wave-length 

 6708.08 may represent the strongest line in the 

 lithium spectrum. 



6 Astrophysical Journal, Vol. 30, p. 92, 1909. 



