284 SCIENCE PROGRESS 



helium and neon the reverse is the case, and the atomic weight 

 increases more than twice as fast as the refractivity. But the 

 most extraordinary fact is the great jump from the row which 

 contains neon to that which contains argon, when the re- 

 fractivity increases fourfold ; coupled with the fact that this 

 relation is exactly repeated in the case of nitrogen and phos- 

 phorus, oxygen and sulphur, fluorine and chlorine. These facts 

 point to the disconcerting conclusion that no formula can exist 

 which will connect refractivities directly with atomic weights, 

 for while the refractivities of adjacent numbers of two neigh- 

 bouring groups are in the same ratio, their atomic weights 

 are not. 



The next point, which is of great importance, is that 

 refractivity is evidently connected closely with valency. An 

 examination of the horizontal row containing nitrogen, oxygen, 

 fluorine, and neon shows that the last named, with the highest 

 atomic weight, 20, has much the lowest refractivity. Next 

 comes the monovalent fluorine, then the divalent oxygen, and 

 finally trivalent nitrogen with the lowest atomic weight and 

 the highest refractivity. But this is not all. If we could 

 measure the refractivity of gaseous carbon, it would, on the 

 same scale, lie somewhere between 330 and 360. It thus ap- 

 pears that increase of valency is accompanied by increase of 

 refractivity. In what proportion the increase occurs it is not 

 possible to say at present, for the atomic weight is not constant, 

 and the relation between refractivity and atomic weight is not 

 known ; in other words, we have no data for calculating what 

 the refractivities of N, O, and F would be if they all had an 

 atomic weight of 20. It is clear, however, that a large portion 

 of the refractivity is independent of the valency ; otherwise the 

 non-valent elements would not retard light at all. 



Gladstone thought that the product of specific refractivity 

 by the square root of the chemical equivalent was constant, 

 but the figures do not support this view. 



Perhaps some light may be shed on the problem by a study 

 of the table of the elements given above. 



The elements are there shown in the order of their atomic 

 weights, as in Mendeleef s table ; but the distribution of the 

 rows is so arranged that the elements whose refractivities are 

 in the ratio of 2, 3, 5 in neighbouring groups are adjacent. The 

 result is to exhibit the table in the shape of a pyramid rather 



