VELOCITY OF POLYMORPHIC CHANGES BETWEEN SOLIDS. 87 



tions receive enough energy to surmount the barrier, the transition is 

 cataclysmic; l)Ut on the other hand, is very slow if only once in a while 

 a molecule rccei\es enough energy to slip over. A transition increas- 

 ing in acceleration with increasing temperature means that the ])order 

 of the velocity domain becomes sharper at the higher temperature, 

 and a retarded acceleration at higher pressures means a more nebulous 

 boundary. In general the results above do show a more nebulous 

 boimdary at higher pressures. One would (>xpect, ther<>fore, that the 

 velocity domain of the phase of smaller volume would have the more 

 nebulous boundary. This is precisely what the greater acceleration 

 with falling as compared with rising pressure means. During the 

 falling pressure transition, the low fJressure modification, with the 

 sharper velocity l)oundary is tumbling over the hill into the lower 

 minimum, and is running with the greater speed. 



It is suggested by the figures that in extreme cases, when the inter- 

 vening hill is low, and the domain of temperature distribution is wide, 

 that not only should there be no region of indifference, but that the 

 equilibrium should be of the nature of a liquid-vapor equilibrium, and 

 should involve eqiuility of streams of molecules in two directions. 

 Such transitions must be of great velocity, and would not lie expected 

 to be within the range of these observations. This may perhaps be 

 the nature of the equilibrium in the neighborhood of a triple point 

 between two solids and a liquid, on the upper end of the ice I-III curve, 

 for example. 



This analysis does not pretend to be an explanation; it makes no 

 attempt to explain why the various factors vary with pressure and 

 temperature in the way in which we have supposed they may. It is 

 merely an attempt to state the nature of the elements that may enter 

 the problem. 



Summary. 



The rate at which one polymorphic modification is transformed 

 into another may be measured by the time rate of change of pressure 

 at constant temperature during the transition. Data for the velocity 

 of a number of such transitions are given in this paper. It is probable 

 that all the measurements of this paper have to do essentially with the 

 rate of advance of a single surface separating the two phases. The 

 rate of advance increases rapidly as pressure is displaced from the 

 pressure of equilibrium between the phases. As a rough comparative 



