NO. II STRUCTURE OF THE ATOM PARSON 23 



We can now derive constitutions for the atoms of all the elements 

 in the manner shown in the accompanying table. Hydrogen, with 

 one magneton only, is followed by a gap ; then comes Helium with a 

 group of eight (represented by " y "), and the table goes on regularly 

 with Lithium (y + i), Beryllium (y + 2), Boron (y + 3), and so on. 

 While this works out very well in the short periods, it is evident that 

 for the long periods the plan must be modified ; for Manganese 

 (3y + 7) behaves very differently from Chlorine (2y + 7). Now a 

 comparison of Vanadium with Phosphorus, Chromium with Sulphur, 

 Manganese with Chlorine, and the Iron-Cobalt-Nickel trio with 

 Argon, shows that these metals of the long period have just the 

 properties that we should expect if there were no tendency in the 

 systems represented by 3y + 5, 3y + 6, 3y + 7 to form a fourth group 

 of eight, and 4y were really 3y + 8. To represent this, I have placed 

 a bar over the number referring to the valence magnetons, thus : 

 3y + 5> 3y + 6, 3y + 7, 3y + 8. This state of affairs, which accounts 

 very well for the differences between what are usually called sub- 

 groups A and B, is carried on, in a diminishing degree, through 



. . 37 + 9 3Y + IO 



Copper, Zinc, and Gallium, with the constitutions ^k i a 



(3y + H) (4y + i) 4y + 2 



>|^ > and the overdue group of eight is assumed not to be 



4y + 3 

 firmly established until Germanium (4y + 4) in group IV is reached. 

 The constitutions assigned to these elements will be discussed in §13 

 of this paper. 



I am unable to see any good reason for the non-formation of this 

 group of eight, or to suggest any simple additional assumption that 

 would secure it. It may be observed that each long period begins 

 with an odd number of groups of eight already within the atom, but 

 that is not likely to be of any particular significance. The non- 

 formation, in certain cases, of the group of eight must then be classed 

 as a subsidiary assumption (§§15, 16) ; but I have shown, in what 

 follows, how well in accordance with the most various facts are the 

 deductions that can be made from it. The tautomerism which has, as 

 one result, been ascribed to the atoms of Copper, Zinc, and Gallium 

 (and their analogues) seems to be a particularly fruitful conception 

 (see §§13-15)- 



