B— CHEMISTRY. 41 



physicist, there is no doubt that this mechanism enables its inventors to 

 correlate a large number of important generalisations, so that some real 

 truth must underlie it, although we may at present be in some doubt as 

 to what that truth exactly is. But I think these chemists have tended 

 to rely too much on supposed modifications of the linkages within the 

 molecule, and have not sufficiently considered the possibility of the forma- 

 tion of co-ordination compounds with the reagents employed, such as 

 those which I have suggested above. The effect of one atom in a molecule 

 in hastening the replacement of another may not be due merely to a 

 weakening of the attachment of the latter, but may be caused by the 

 formation of a co-ordinate link through the former, or this may promote 

 co-ordination through some other atom in the molecule. We know now 

 that even in purely ' organic ' compounds — quite apart from those organo- 

 metallic compounds which the old-fashioned organic chemist regarded 

 with so much distaste — co-ordination is of frequent occurrence. In the 

 particular form of the production of chelate rings, that is, in the form of 

 co-ordination between two atoms of the same molecule, it has been shown 

 to occur in (3-diketones and (3-ketoesters, in many ortho-substituted 

 phenols, and in a-keto-oximes, and to be responsible for much of the 

 chemical as well as the physical peculiarities of these substances ; and in 

 the more general form of association or ' molecular compound ' formation 

 its occurrence is widespread. Formerly the production of such compounds 

 was ascribed to some inferior and rather contemptible form of valency, 

 possibly to a force acting not between atoms at all, but between whole 

 molecules, and so the influence of their formation on what were regarded 

 as the reactions of genuine valencies was naturally taken little into account. 

 But we now realise that they owe their existence to the production of 

 co-ordinate links, and that the co-ordinate link is in essence the same as 

 a normal covalency. The co-ordinated hydrogen, for example, as in 



H— 0— H^O< , 

 \H 



is attached to each of the two oxygen atoms by means of two shared 

 electrons. The link on one side is just as genuine as that on the other, 

 although, owing to the difference in the states of the two oxygen atoms, 

 one of them may separate more easily. It therefore seems probable that 

 the formation of such a link may often be a preliminary stage to the 

 complete transference of the hydrogen from one point of attachment to 

 another, and that the possibility of its formation may be a necessary 

 condition of reaction. We have further to recognise another way in which 

 reaction may be promoted by co-ordination, which is illustrated by the 

 example I gave of the hydrolysis of an acid chloride. The formation of a 

 co-ordination compound between two molecules may bring two atoms 

 into proximity with one another, and so favour their reaction. In 

 developing this possibility we have to consider the stereochemical relations. 

 The study of chelate rings has shown us what forms of ring are most 

 stable ; owing to the weakness of the co-ordinate link which they contain 

 such rings afford, as I have already pointed out, a more delicate test of 

 strain than the ordinary rings of organic chemistry. Thus we find that 

 a chelate ring of six atoms, including double links, is formed with peculiar 



