302 Mr. J. F. Heyes on Valency, 



be convenient in electrolysis investigations, for instance, 

 when there is no ambiguity, to nse the phrase " residual 

 affinity/'' and the appropriateness of the analogy to " residual 

 charge " of a leyden jar is obvious. But it seems hard to 

 place so many of the compounds mentioned in my former 

 paper, and even PC1 5 or N(C 2 H 5 ) 4 OH, outside the pale of 

 honest " atomic " compounds — i. e. chemical molecules — and 

 alongside the innumerable hydrates which molecular, or pro- 

 phosphorus not yet known in the case of nitrogen. The alleged cases of 

 ChJ} S + C 2H 3 I and of gf^J S+CH 3 I 



(C 2 EL 



I CH 



giving distinct compounds (p. 27) of S -l qjt 3 is also interesting from 



(I 



the possibility of iodine being trivalent or at least tervalid. Then there 

 would be 



g^S^— C 2 H 5 and H 5 C 2 \ S==ICH3 



Sulphur is at least tetravalent in whatever sense iodine is trivalent. But, 



fOgH 5 



1 pti 



even assuming S IV < qtt 3 and the " equal value " of the " affinities " of 



the sulphur atom, we might expect a case of physical isomerism in 



CH 3v /CH 3 CH 3 . /C 2 H 5 



>S( and )S( 



r x c 2 h 5 r x ch 3 



where the methyl radicals hold ortho and meta " positions," as it were, 

 even on the maypole hypothesis. Since writing I find striking statements 

 by Professor Tilden on the trivalency of iodine written in 1876. On 

 page 181 of his ' Introduction to Chemical Philosophy ' (first edition) he 



even gave a closed-ring formula for KHgI 3 as K — I' | /Hg, pointing out 



that although " many chemists regard such compounds as formed by the 

 union of entire molecules of the constituent salts combined together by 

 some sort of adhesion differing from ordinary chemical affinity," yet ft if 

 they do they might naturally be expected to exhibit something of the 

 appearance and external characters of their proximate constituents." 

 He also considered the idea of " two extra units of combining power " 

 for oxygen "not wholly preposterous," and sketched crystallized 

 ZnS0 4 OH 2 . 60H 2 with " four bond " oxygen atoms throughout. This 

 was, however, with a view to diminish the number of " molecular combi- 

 nations," which in a sense it is now proposed to increase. There is much 

 logic and tempting simplicity in the view that an atom is always w-valent, 

 and that conflicting puzzles are "molecular," not "atomic," but this 

 logical view seems intimately connected with the idea that the atom of 

 N, for example, was born a manufactured article with tliree hooks, 

 or bonds, 



