VELOCITY OF POLYMORPHIC CHANGES BETWEEN SOLIDS. 73 



Discussion. 



A simple interpretation of the results is complicated by several 

 obscurin.uj factors, which we will first discuss. There is, in the first 

 place, the efl'ect of the heat set free din-ing the transition. The effect 

 of this is that the temperature at which the transition is running is not 

 the observed temperature of the bath, but differs from it by an un- 

 known amount depending on the magnitude of the heat of transition, 

 the transition velocity, and the thermal conductivity of the surround- 

 ing envelope. If the transition is very rapid, the measured rate may 

 be entirely controlled by the rate of dissipation of the latent heat, the 

 two phases being always under equilibrium conditions at the mo- 

 mentary actual temperature of the interior. As heat is conducted 

 away from the interior, the pressure so changes as to assume the value 

 appropriate to the temperature of the interior, and it is this rate of 

 change of temperature which is really measured. This is preeminently 

 the case on a melting curve. Of course it is not possible to carry the 

 solid any distance at all into the domain of the liquid, and the apparent 

 rate of change of pressure is onl}' the rate at which pressure follows 

 the return of temperature along the equilibrium line. The reverse 

 displacement, that of the liquid into the domain of the solid, is possible 

 to carry out, because liquids may be subcooled. But even in the 

 reverse case, the latent heat of freezing is so large, and its rate of 

 dissipation is so slow compared with the rate of crystallization of the 

 liquid, that the apparent rate is governed by the thermal conductivity. 

 The result is that the apparent rate of melting of most solids is exactly 

 the same as the apparent rate of freezing. Figure 18 for mercury 

 shows an example of this. In some cases, however, the velocity of 

 crystallization is unusually slow, so that an appreciable fraction of the 

 latent heat is conducted away during the freezing, and it may be 

 possible to detect the difference of velocity between melting and 

 freezing. Figure 19 shows such results for benzophenone. If the . 

 conductivity were perfect, results of this character would be shown by 

 all liquids; the rate of melting would be infinite and the rate of freez- 

 ing would be a different characteristic rate for every substance. As a 

 consequence of this large disturbing factor, it has not been worth while 

 to try to collect any data for the rate of melting or freezing under 

 pressure. 



Slow thermal dissipation is evidently going to be a disturbing factor 

 also in the case of solid transitions, but to a much less degree than for 



