Law of Atomic Weights. 415 



15 and sesqui-radius 1. Between these lies the vacant 

 sesqui-radius 16. 



"Note. — By a sesqui-radius is meant a radius along with 

 the inner part of the opposite radius. 



" 8. It is proved in the memoir that the unoccupied sesqui- 

 radius is not arbitrarily introduced into the diagram, but has 

 a real existence in nature. 



" 9. The natural chemical groupings of the elements come 

 out with conspicuous distinctness, e.g.,~F, CI, Br, I, on sesqui- 

 radius 15. 



" 10. Shaded prominences are introduced to point out the 

 elements of greatest atomic volume in the solid state, and 

 shaded sectors to indicate those of least atomic volume. It 

 is easily seen that the diagram can be made to convey other 

 useful information in a similarly convenient form. 



" 11. If, as seems probable, the Logarithmic Law is the 

 law of nature, there appear to be three elements lighter than 

 hydrogen, which may be called infra-fluorine, infra-oxygen, 

 and infra-nitrogen. Again, whether the real law be the 

 logarithmic or the elliptic law, there are six missing elements 

 between hvdro£en and lithium. Infra-nitrogen has a con- 



JO o 



jectured atomic weight of about one- sixty-fourth the atomic 

 weight of hydrogen. This estimate is arrived at by com- 

 puting the deviation by the law referred to above in section 5, 

 and applying it to the logarithmic curve. 



"12. It will be found convenient to number the radii from 

 1 to 16, and to letter the coils from a to g, using for coil a 

 — the innermost coil — only three-quarters of a complete coil. 

 By this arrangement the numbering of the radii will be that 

 which is most convenient to chemists. 



" Thus thallium, being on radius 11 and coil /, may be 

 briefly designated 11/j or/11. Uranium is g 6. Infra-nitro- 

 gen, infra-oxygen, and infra-fluorine, are a 5, a 6, and a 7. 



" 13. The best determinations of atomic weights that could 

 be procured were used in the investigation, but those entered 

 on the diagram are only the rough values in common use. 



" August 1888.'' " G. JOHXSTOXE STONEY." 



[Note added August 1902. — No. 11 of the above ' Obser- 

 vations ' contains conjectures depending on extrapolation, 

 which I am unwilling to put forward again without a caution 

 that they are now more doubtful than they seemed to be 

 fourteen years ago. They depend on the probability, what- 

 ever it is, that the logarithmic curve is to be employed for 

 elements with less atomic weight than lithium, rather than 

 the elliptic curve, or any other, if other there be, which can. 

 like these, thread it< wav amid all observed atomic weights 



