312 On Valency, Validity, and Residual Affinity. 



elements such as oxygen." On the other hand, the view that 

 in the so-called molecular formula KC1 the elements are only 

 " apparently monovalent " leads the Professor to favour the 

 view of metallic linkings, each metallic atom being associated 

 with strictly monovalent chlorine atoms. Thus formulas such as 



Cl 2 =Pb— Pb=Cl 2 CI— Fe=Fe— CI 



II and II 



Cl 2 =Pb— Pb=Cl 2 CI CI 



are quoted as probable or possible, pending vapour-density 

 determinations. In face of the scanty evidence for di-metallic 

 linkings in most cases, and in view of the much stronger 

 evidence in favour of trivalent chlorine, I hold at present that 



the formula Fe\™_ Qj/Fe is preferable. Moreover the K 3 C1 3 

 formula seems less probable after the recent vapour-density 

 determination favouring the gas molecule KI. The Ag — Ag 

 linking is at least premature. 



(2) Professor Meyer so thoroughly argues for trivalent 

 aluminium except in A1 2 C1 6 , that only the acceptance of this 



/C1=C1 X 

 exception to prove the rule as Al^— C1=C1-^A1 is required. 



\C1=C1/ 

 It is only in the case of hydrocarbon compounds that the 

 strict monovalence of chlorine can be said to be demonstrated. 



(3) Professor Meyer thinks that Ba &c. are divalent in only 

 their more common and most stable compounds, but that " the 

 peroxides indicate a higher valency," presumably Ba IV 2 . He 

 quotes Geuther as viewing Be, Mg, Ca, Sr, Ba as having 

 valencies 2, 4, 6, 8, and 10. If once the peculiarities of the 

 non-metals in these molecular compounds are recognized, 

 these extraordinary valence powers will disappear. It gives 

 no explanation of the " molecular compounds" which even the 

 "uncommon" compounds with higher valencies themselves form. 



(4) The atoms Ru and Os are viewed as " undoubtedly 

 octovalent " in their volatile oxides M V1II 4 , and cobalt pro^ 

 bably in the purpureo-chloride (NH 3 ) 5 (JoCl 3 , as if from 

 H 5 Co ym Cl 3 . The valence numbers are u probably from 1 

 to 8." 



(5) In As 4 6 Professor Meyer does not link arsenic with 

 arsenic atoms, but suggests, with probability, a symmetrical 

 arrangement : — 



As— 0— As 

 /\ 

 O. 



A 



As— 0— As 



