Validity, and Residual Affinity. 307 



(2) The existence of the so-called molecular compounds is 

 due to the validity or varying valency of the non-metallic 

 radicals, certainly not, as a rule, to the very high validity of 

 the metallic atom. Thus the great difficulty experienced in 

 " washing " certain precipitated oxides free from potash is, I 

 take it, due to the validity of oxygen. And in spinelle, 

 A1 2 3 . MgO, the tetravalent oxygen makes it probably 



0=A1\ 



|| >0=0=Mg. 



0=A1/ 



(3) In particular, the numerous grades of bodies containing 

 H 2 is due to the tetravalency (or quadrivalidity) of oxygen 

 conditionally stimulated in some unknown way (say, to a 

 change of vibration-period or to a development of residual 

 charge) by chemical proximity to other non-metallic or 

 negative atom or atoms. 



(4) So-called catalytic change is often, if not always, due 

 to this development or existence of validity. This is seen in 

 the behaviour of hydrochloric acid as contrasted with liquid 

 hydrogen chloride. 



(5) So-called polymeric changes in many cases, e. g. 



CH 3 CH.O v 

 H 3 3 N 3 3 and para-aldehyd, | )OCH . OH 3 , 



CH 3 CH . 

 are similarly often due to the tetravalency of oxygen. These 

 seem to be perfectly analogous to the cases of allotropy. 



(6) In such cases we must be prepared to extend the idea 

 of nucleus hitherto confined to carbon (and extended on 

 unsatisfactory evidence to several of the metals*) to a chlorine 

 nucleus, an oxygen nucleus, &c. The formula 



oi==ci x 



A1-C1==C1-A1 

 N C1==C1 / 



is at least as probable as 



iVl iy Cl 3 

 A1 IV CL 



* Of the numerous cases usually instanced that of tin seems to me 



SnEt 3 



alone good. The evidence of Sn 2 Et 6 is strongly in favour of | ; 



SnEt 3 



whereas Sn 2 Cl 4 at low temperatures is ambiguous, and probably 



