Numerics of the Elements. 153 



Lanthanum. — Oleve's result (confirming Brauner's) becomes 

 La = 137*58, if we take the calculated numerics = 15*94, 

 S = 31*82. The third member of Group X. has the numeric 

 137*64. 



Cerium. — Brauner has most exhaustively confirmed Robin- 

 son's long series of results. We may therefore take mean 

 experimental Ce as finally settled at 139*89. Hence Ce is 

 the sixth member of Group X., its calculated numeric being 

 139*82. 



Didymium. — According to Cleve, didymium free from 

 samarium has the numeric 141*70 (recalculated as for La). 

 It is accordingly the ninth member of Group X. ; for which 

 y = 141*61. 



[The above three numerics constitute an exact triad as 

 regards their order in this group, the values of x being 3, 6, 9 

 respectively.] 



Samarium. — Cleve's new result is (recalculated as previ- 

 ously) Bm = 149*36. This agrees exactly with #=49 in 

 Group X. 



Erbium. — The balance of evidence is, on the whole, favour- 

 able to Er = 165, nearly. This is the value of y when #=0 

 in Group XII. 



Ytterbium. — The most trustworthy results, on recalculation, 

 make Yb = 172*22. The tenth member of Group XII. has 

 the numeric 172*13. 



Tantalum. — Marignac^s best number is undoubtedly the 

 one which depends on the ratio 2K 2 S0 4 : Ta 2 5 , because the 

 method of calculation necessarily diminishes the effect of 

 experimental errors. Putting K = 38*92, and otherwise re- 

 calculating as before, Ta = 181*84. The agreement with y 2 (in 

 Group XIII.) =181*82 is still closer than indicated in Part I. 



Thorinum. — Nilsson's experiments (1883) are perhaps more 

 valuable than those of his predecessors. His mean numeric 

 (also recalculated) is Th = 230*47. The seventh member of 

 Group XVI. has the numeric 230*45. 



Uranium. — In Part I., U = 239*70, on the basis of Peligot's 

 results. If the calculated values of C and O be taken, U = 

 239*64, corresponding to the fifty-ninth numeric (239*67) 

 in Group XVI. A glance at the Table (infra) shows that 

 position to be quite anomalous, and that the numeric has 

 been much over-estimated. 



Scandium, Yttrium, and Platinum have been examined 

 afresh. The numerics assigned to them in Part I. are con- 

 firmed. 



[Note. — I have observed that the earlier values of numerics 

 are apt to be too high.] 



