Mechanism of Molecular Action, 803 



This result requires that for molecules of Q l5 Q 2 , &c. to 

 react in accordance with the back action they must collide 

 with a relative velocity lying between v' and (t/ + &.''). 



Each of these postulated mechanisms of forward action gives 

 a corresponding mechanism of back action that is consistent 

 and readily intelligible. The other assumptions as to forward 

 action which have been tried give corresponding mechanisms 

 of back action which are either inconsistent or fantastic ; 

 for example, the assumption that for molecules of P l5 P 2 , &c. 

 to react in accordance with the forward action they must 

 collide with a relative velocity of approach less than i\ 

 requires that for molecules of Q 1? Q 2 , &c. to react in ac- 

 cordance with the back action, they must collide with a 

 relative velocity of approach less than yit and greater than ///, 

 where /ju—fi f = v t a mechanism it is very difficult to imagine. 



It should be noticed that in both mechanism (1) and 

 mechanism (2) the index of the exponential is of the form 



positive quantity 



for both the forward and the back action. 



It would seem, then, that all chemical actions are accele- 

 rated by rise of temperature ; that is to say, in reversible 

 actions both the forward and the back actions are accelerated. 

 Cases where an action seems at first sight to be retarded are 

 due, therefore, to a back action which increases more rapidly 

 than the forward action ; there is, in fact, a brisker inter- 

 change of particles between the two states, though whether 

 an action seems to be retarded or seems to be accelerated, 

 depends upon the relative rate of increase of its constituent 

 forward and back actions. 



As experimental evidence oE these results there can be 

 used 



(1) Measurements of the velocity of an action at various 



temperatures. 



(2) Measurements of the equilibrium constant of an action 



at various temperatures. 

 Unfortunately, the forms involving 



e ' ■ \/0 and e ' • ^ 



are influenced so much more readily by the exponential than 

 by the surd factor, that it is not possible to determine from 



