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



15 



Fig. 3 continued. 



' The formula (a) is a type of mono-substituted benzene 

 derivative which gives mainly 0- and £-di-derivatives on 



further substitution, and Y may be di-valent O, £<?r-valent N, 

 halogen, etc. On the other hand (6) is the type which mainly 

 yields larger quantities of m-di-derivative, and here Y may 



be the di-valent O of >CO or >X0 2 , the /er-valent N of 



-CN, the halogen of quaternary ammonium salts, 



-N(CH 3 ) 3 .C1, or of -CHC1 2 or -CC1 3 . In the case of 

 toluene it is not improbable that the hydrogen of the methyl 

 group is the " directive " or " key " atom, so that in addition 

 to (a) and (b) it may be necessary to add a third type with 



- + 

 -Y-X, as side group with o- and ^-directive influence. 



Schemes very similar to (a) and (b) have been used pre- 

 viously by Fry, Vorlander and others, but with less stress on 

 the influence of the "key" atoms as the paramount one; 

 moreover, when Fry attaches alternating + and - symbols to 

 successive atoms he does so on an ad hoc basis and is at no pains 

 to show that this alternation is so general as the present writer 

 has attempted to demonstrate. The principle of alternating 

 latent polarities does not appear to follow from Thomson's 

 theory of a bond, and Fry's formulas are often wholly at 

 variance with the principle in question. 



The writer raises the case of raeta-substitution in some 

 .detail, as he drew attention to its connexion with the case of 

 a-substitution in the fatty series many years ago. Formula (c) 

 in Fig. 3 represents his conception of the reactive phase of the 

 benzaldehyde molecule preparatory to attack by a substituting 

 agent in the m-position ; comparison of this formula with (ii), 

 Fig. 2, will make the exact analogy clear. 



The corresponding phase of nitrobenzene prior to attack 



