'80 BRIDGMAN. 



and therefore the observed transition velocity is very possibly greater 

 than if the reaction had been allowed to run its natural course. In 

 the following, this small residual effect will be entirely disregarded, if 

 indeed it exists in most cases, as being an extraneous effect. 



The general behavior of the acceleration curves next calls for com- 

 ment. One feature common to nearly all the curves is that the acceler- 

 ation with falling pressure is greater than with rising. This is what 

 one would expect; that reaction runs more rapidly which is pushed 

 by the external pressure. There may be a few exceptions, however, 

 as in the case of NH4NO3. 



With regard to the variation of the acceleration along the transition 

 curves we may have either increasing or decreasing acceleration at the 

 higher temperatures. It is to be recognized that on a normal transi- 

 tion curve there are two opposing tendencies; increase of pressure 

 would be expected to decrease the transition acceleration, because of 

 increase of viscosity, while increasing temperature would increase it. 

 Phosphorus is a well marked example of decreasing acceleration, and 

 camphor of increasing. The curves for AgXOa are remarkable; the 

 acceleration passes through a maximum, much more pronounced for 

 falling than for rising pressure, in the region of the rapid change of 

 •direction of the transition curve. It has already been stated that this 

 is a region of anomalous behavior of other physical properties also. 

 The accelerations with rising and falling pressure are usually affected 

 in the same way as we move along a transition line, but the magni- 

 tude of the effects may be different, and in exceptional cases possibly 

 the signs may be unlike. 



The behavior of the region of indifference next concerns us. It is 

 first to be noted that the apparent equilibrium within the region of 

 indifference is somewhat different in its nature from the so-called 

 ■"false equilibria" with which we are familiar in such cases as diamond 

 and graphite or hardened steel. The usual explanation of the failure 

 to react in such cases is that one of the phases has been cooled so far 

 below the temperature of transition that the velocity of transition has 

 become inappreciable because of the enormously increased viscosity. 

 A similar example is afforded by those liquids which may be subcooled 

 so far as to become glassy. In the cases described in this paper, 

 however, the reaction does not run when two phases are in contact 

 at only a slight distance from the equilibrium temperature. Such 

 immobility cannot be a viscosity effect, because at lower temperatures, 

 where the viscosity is greater, the reaction may run with high velocity. 



Phenomena consistent with these have been observed before. In 



