444 



SCIENCE 



[N. S. Vol. XL VI. No. 1193 



consideration of the radioactive elements 

 indicates that those which have odd atomic 

 numbers have either shorter periods, or else 

 are at present unknown. Now unfortu- 

 nately there is no known method of test- 

 ing the stability of the elements of low 

 atomic number, but it mig'ht seem, at first 

 thought, that the more stable atoms should 

 be the more abundantly formed, and to a 

 certain extent this is undoubtedly true. If 

 then, at the stage of evolution represented 

 by the solar system, or by the earth, it is 

 f oimd that the even-numbered elements are 

 more abundant than the odd, as seems to 

 be the case, then it might be assumed that 

 the even-numbered elements are on the 

 whole the more stable. However, there is 

 at least one other factor than stability 

 ■which must be considered in this eonnec- 

 ■tion. The formula of the even-numbered 

 (elements has been shown to be nlle'. Now, 

 since that for the odd-numbered elements 

 is nHe' + E^' it is evident that if the sup- 

 ply of the H^' needed by the elements was 

 relatively small at the time of their forma- 

 tion, not so much material would go into 

 this system, and this would be true whether 

 -the E^' represents three atoms of hydrogen 

 or one atom of some other element.^ 



In studying the relative abundance of 

 the elements the ideal method would be to 

 sample one or more solar systems at the 



3 "With regard to the latter alternative, it is at 

 least remarkable that the H3 occurs 11 times in the 

 system for the first 27 elements, while Hj and H 

 each occur only once, and it may also be men- 

 tioned that Fabry and Buisson have by interference 

 methods determined the atomic weight of nebulium 

 to be 2.7, and this they think indicates that its real 

 atomic weight is 3. Also, Campbell has found that 

 in the nebula N. 6. C. Index 418, situated in the 

 southern part of the constellation of Orion, the 

 nebulium spectrum is found farther from the in- 

 terior than that of helium, while the hydrogen spec- 

 trum extends out to a much greater distance still. 

 This, he thinks, indicates that the atomic weight of 

 nebulium lies between the values for hydrogen (1) 

 and helium (4). 



desired stage of evolution, and to make a 

 quantitative analysis for all of the 92 ele- 

 ments of the ordinary system. Since this 

 is evidently impossible, even in the case of 

 the earth, it might be considered that suffi- 

 ciently good data could be obtained from 

 the earth's crust, or the lithosphere. How- 

 ever, the part of the crust to which we 

 have access is relatively so thin, and has 



Atomic Number 

 Fig. 2. The Periodic Variation in the Abun- 

 dance of the Elements as the result of Atomic Evo- 

 lution. The data are given for 125 stone meteo- 

 rites, but the relations are true for meteorites in 

 general. Note that ten elements of even atomic 

 number make up 97.59 per cent, of the meteorites, 

 and seven odd-numbered elements, 2.41 per cent., 

 or 100 per cent, in all. Elements of atomic num- 

 ber gi-eater than 29 are present only in traces. 



been subjected to such far-reaching mag- 

 matic differentiation, and to such extensive 

 solubility effects, that it seems improbable 

 that the surface of the earth at all truly 

 represents its composition as a whole. 

 The meteorites, on the other hand, show 

 much less evidence of differentiative ef- 

 fects, and undoubtedly represent more 

 truly the average composition of our 

 planetary system. At least it might seem 

 proper to assume that the meteorites would 



