Theory of the Chemical Elements. 877 



weight 3*25x48^7-41 or 3*8. This is consistent with the 

 theory that the polonium a rays are actually helium. 



This digression has been somewhat long, but it serves a 

 useful purpose. Nebulium and proton 1 uorine, even if not 

 primary substances from which others are built up in this 

 particular way, are not hypothetical, as a consideration of 

 spectra shows. 



We may now proceed further with the main argument. 

 Of the inert gases, we have already considered helium, 

 argon, and neon. 



Krypton and Xenon. — The inert gases may be divided 

 into the argon and neon groups, if the arrangement in 

 double columns of the Periodic Table be adopted, that is to 

 say, the arrangement which places potassium, rubidium, 

 and caesium in one subgroup, and sodium, copper, silver 

 in another of the same column. In the argon subgroup, 

 the sequence is argon, krypton, xenon, an unknown gas, 

 and niton. The only other member of the parallel group at 

 present known is neon, and it is followed by four vacant 

 spaces. Sir William Ramsay has suggested that the vacant 

 spaces should be occupied by emanations, possibly present 

 in the atmosphere, but decomposed during the process of 

 separation. We shall endeavour to fill ihese spaces, guided 

 in our selection of formulas by the groupings of primary atoms 

 found effective for other elements. 



AVe consider krypton and xenon in the first place. Argon 

 has already been classified as He 10 , and neon as (PfH) 6 . It 

 is preferable to write 



A = 5He 2 , Ne = 2(PfH) 3 , 



for the groups He 2 and (PfH) 3 appear to have some 

 significance. 



Consider now the groupings 



5{Nu 4 (P£H) 3 }, 5{He 4 (PfH) 3 } 



with an obvious relation to each other and to those of argon 

 and neon. If they represent elements, the calculated atomic 

 weights of these elements are respectively 83*08 and 130*28. 

 The present International atomic weights of krypton and 

 xenon are 82*9 and 130*2, which, allowing for errors of 

 experiment, especially likely in the case of krypton, are in 

 very close accord with these calculated numbers. Without 

 further discussion, we shall consider these formulae to repre- 

 sent the atoms of krypton and xenon. 



Passing now to the neon group, the first vacancy occurs 



