I.T.I'II.UTIVK IX DICKS OF THE ELEMENTS. 



347 



hypothesis of simple relationships running through ever}- family of element*, it must 

 lie admitted that the figures are not encouraging. 



The most hopeful series is that of the alkali elements, in which we find that the 

 constants for K, Rb and Cs are 7 '85, 12'1 and 19'2, numbers which stand very nearly 

 in the relation of 2, 3 and 5 ; but lithium and sodium do not show any tendency to 

 fall into their places, and none of the other series of elements promise any better 

 success. It must, however, again be repeated that the refraction equivalents are 

 often misleading, so that an unpromising appearance ought not to daunt us. 



The third point which is remarkable in the relative refractivities of the elements is 

 their connection with the valency. In the periodic arrangement of the elements, 

 N, O, and Ne are followed, respectively, by P, S, and A, and the refractivity of each 

 oi the latter set is four times that ot each of the former. These, therefore, are, 

 probably, in some sense homologous elements, and we may consider together the 

 series Ne, O, N, and A, Cl, S and P. In the latter set we find that A, with the 

 highest atomic weight, but zero valency, retards light least, and that, as we pass 

 to Cl, S, and P, the refractivity increases with the valency, in spite of the decrease in 

 atomic weight. The same thing occurs in the series Ne, O, N ; and here it is 

 remarkable that the refractivity of carbon is. in all probability, even greater than 

 that of nitrogen, l>eing at least 320 for C a , and probably as high as 3f>0. 



So far as can )>e judged from the scanty knowledge we possess, the retarding 

 jx)wer diminishes from tin's point as we move across the table from right to left, that 

 of Li being less than that of carbon. But the case of boron presents extremely 

 interesting features. 



GHIRA* has measured the indices of BC1., and BBr^ in the liquid state, and the 

 following table shows the values of the refractivity of boron which can be inferred 

 from his results, on the assumption that the additive law for compounds holds exactly. 



TABLE XII. 



* 'R. Accad. Lincei Rendiconti,' 1893, vol. 2, iem. 1, p. 312. 



2 Y 2 



