84 BRIDGMAN. 



sides of the transition point, when there is no appreciable region of 

 indifference, one at least of the transition velocity curves must have 

 a bend at the equilibrium point, as shown in Figure 22. But the 

 curves of both Figures 21 and 22 involve non-analytical singularities 

 of a kind that we are not willing to admit in our physical phenomena, 

 at least not on this scale of magnitude. 



The fact that polymorphic ecjuilibrium is not a dynamic equili- 

 brium is significant w^th respect to the random distribution of velocity 

 of temperature agitation among the molecules. In a vapor there is 

 almost certainly, and in a liquid quite probably, a very close approach 

 to Maxwell's distribution of velocities. In such a distribution there 

 are always some molecules with a velocity so much above the average 

 that they can pierce the transition layer and enter the other phase. 



D. 



o. 



Fig. 21. Fig. 22. 



Figure 21. Hypothetical curves showing what must be the nature of the- 

 velocity between two polymorphic modifications in those cases where there is a 

 band of indifference. 



Figure 22. Hypothetical curves showing what must be the nature of the 

 transition velocity between two polymorphic modifications in order to explain 

 asymmetry of velocity in those cases where there is no band of indifference. 



The fact of the existence of the region of indifference shows that with 

 polymorphs there are no molecules at the equilibrium point with 

 velocities high enough above the average to pierce the transition layer 

 (or its equivalent) and enter the other phase. That is, in a crystal 

 the random temperature agitation is random only within a restricted 

 range ; the velocity never rises above a definite limit. This is certainly 

 consistent with the known definiteness of crystal structure as shown 

 by X-ray photographs. The region within which the velocities are 

 confined in all probability does not have a sharp boundary, but is 

 more or less nebulous at the edges. We recognize that the abruptness 

 of this boundary within which the velocities are confined may vary 

 greatly from substance to substance, and may also vary with the same 

 substance at different pressures and temperatures. 



We have ruled out the reciprocal passage of molecules between 



