Manchester Memoirs, Vol. Ixiv. (1920), No. 3 11 



In past years the writer endeavoured to develop schemes for 

 reactions of carbon compounds on the assumption that those 

 reactions were usually ionic in character. A mass of evidence 

 has accumulated during the intervening years which tends 

 to show that even those changes which in the main proceed 

 through the ions have a parallel non-ionic course also, and 

 that frequently the two proceed simultaneously and with the 

 same end results, polarity considerations evidently affecting 

 both identically. In seeking schemes to represent the course 

 of such changes the writer holds that it is most satisfactory 

 to seek first an explanation of the reaction in its ionic phase 

 and subsequently to deduce the scheme for the reaction between 

 the non-ionised components, the advantage of this course 

 being that in nearly all instances only one of the radicles (or 

 ions) of the electrolyte is chemically active and the ionic 

 reaction must involve this one only at the critical stage — for 

 example, OH r in alkaline saponification and CN ; in cyano- 

 hydrin formation. These ions are weak in comparison with, 

 sav, N(V ; they have therefore less tendency to a stable 

 diffuse distribution of their valencies, and tend to lose 

 their ionic state by concentrating their diffuse valency on a 

 single atom of another substance, especially if that atom 

 itself has a decided polar character, inherent or induced. So, 

 for instance, HCN, HS0 3 H, etc., do not unite with an 

 ethvlenic bond, unless this is polarised as in a/3-unsatu rated 

 ketones, and are thus distinguished from HO'Cl, ON'N0 2 

 and Br 2 , for example, which frequently attach themselves to 

 non-polarised ethvlenic bonds. 



In order to indicate, necessarily very crudely owing to 

 considerations of space, how the writer conceives these various 

 aspects to contribute to the mechanism of chemical changes, 

 he has selected three cases to which he attaches special interest. 

 In the diagrams used he has followed Robinson in dividing 

 a distributed valencv into fractions, usually three, which total 

 to the exact value of the single undistributed valency. A 

 single valency is as usual indicated by an uninterrupted line 

 and the fractions by dotted lines. The "latent polarities >r 

 of the atoms are indicated by the signs + or -, so that a 

 molecule is electrically neutral when an equal number of free 

 valencies proceed from positively and negatively polar atoms, 

 while an ion has an excess of free valencies (which total to a 

 whole number) proceeding from atoms of the one kind. 



The first case is that of the formation of acetonecyano- 

 hydrin which must be conceived as due either to the formation 

 of a complex ion, from cyanion and acetone, or by direct addition 

 of a metallic cyanide to acetone (compare Trans. Chem. Soc, 



