46 SECTIONAL ADDRESSES 



From this point of view the necessity for the change is independent of 

 the particular mechanism by which it may be effected, but different 

 mechanisms will bring it about with different velocities. 



There would seem an evident possibility, in these ionic interchanges, 

 of the hydroxyl or other anion (or the ammonia) interacting at one of the 

 other three faces of the tetrahedron, in which case no inversion of con- 

 figuration would ensue. The di-pole moment, which must be associated 

 with the link between the carbon atom and the radical which is going to 

 be eliminated as an anion, will, however, have a directive effect on the 

 attacking anion (or the pofer ammonia molecule), and this will cause the 

 probability of approach to be greatest on the opposite face of the 

 tetrahedron, especially when the other substituents are relatively in- 

 different. Elimination of the ionisable group will also be facilitated 

 most when the movement of the nucleus is directly away from it. Thus 

 it would appear that the mechanism effecting the change with the greatest 

 velocity would be that which produces an inversion of configuration. 



In these ionic interchanges the reaction-mechanism is sufficiently clear 

 to enable us to discuss their stereochemical aspect. Reactions such as 

 those of phosphorus pentachloride or thionyl chloride on hydroxylic 

 substances, or even of moist silver oxide on halogen compounds, are 

 evidently more complex, and their mechanism seems at present too 

 imperfectly known to justify speculation on the configurational changes 

 by which they may or may not be accompanied. 



A phenomenon which was generally regarded, though not with very 

 good reason, as closely related to the Walden inversion was that of trans- 

 addition at a double bond. The two phenomena had, however, this in 

 common, that both showed the inadequacy of the old conception of the 

 solid carbon atom with fixed valency poles. The two solid tetrahedra 

 attached edge to edge and opening on one angle to allow of the attachment 

 of the two addenda gave too crude a picture. 



The ethylenic link is now regarded rather as a unit, formed from two 

 pairs of electrons of opposite spins, and possessing torsional rigidity. 

 When combination with, for example, a chlorine molecule occurs, two 

 electrons of opposite spins are taken to share in binding the two chlorine 

 atoms, leaving two, also of opposite spins, to form the single link by which 

 the two carbon atoms remain united. 



Experiment shows that this process can take place so as to lead either 

 to cis- or to irfi«5-addition, and, since these may occur concurrently, we 

 may conclude that, so far as the readjustment of the link is concerned, 

 there is no great difference in the facility with which the two types of 

 change take place. 



The result appears to be determined partly by the energy relations 

 between the possible products, and, since chlorine shows a greater 

 tendency to CM-addition than bromine, it is probable that simple mechanical 

 factors, such as the distance between the centres of the two atoms to be 

 added, have an important effect. It is easy to see that an increase in 

 this distance might favour traiis-zddition. 



Closely related to the phenomenon of /?'«K5-addition is that of ^rani- 

 elimination. It has long been recognised that in ethylene derivatives the 



