^. 



POLYMORPHISM AT HIGH PRESSURES. 185 



them out of these nicely fitted positions into other positions further 

 apart, of less mean internal pressure, and greater freedom for tempera- 

 ture agitation. Figures 3G and 87 show a possible scheme for the 

 low and high temperature modifications. The effect of increased 

 pressure is ol)\iously to compact the phase of smaller volume, so that 

 increased tt'iiiperature is neetled to pull the molecules apart. As a 

 general rule, we would expect the phase of smaller volume to be more 

 incompressible, but if the molecule is unequally compressible in 

 different directions, as it nmst be, it is evident that under the proper 

 conditions the phase of larger volume may be more incompressible. 



The models of Figures 3G and 37 are suggestive in several other 

 particulars. It may Ix' mentioned in the first place that these units 

 may be built u}) into several other structures, two at least of which 

 have a still smaller volume than that of Figure 3(i, thus showing the 

 possibility of a number of polymorphic forms. The arrangement of 

 Figure 36 is evidently one which will show considerable persistence; 

 it will stand a good deal of superheating and the transition velocity 

 will be low, whereas that of Figure 37 will stand relatively slight super- 

 heating (or subcooling) and the transition velocity will be high. 

 Figure 36 also illustrates a point made in a previous paper,*^ namely 

 that if an atom passes from one modification to another it must rise 

 from its position of ecjuilibrium and pass through an intermediate 

 position of greater potential energy. It is evitlent that some initial 

 work will be required to pull an atom from the position of Figure 36, 

 even if this work is more than returned when it settles down into the 

 final position of equilibrium. In this way the band of indifference may 

 be accounted for. It is also evident that if the two phases of Figm-es 

 36 and 37 are in contact there will be a field of force over the surface 

 of each phase in which the atoms will tend to orient themselves in the 

 appropriate positions. The tendency to pass from one phase to an- 

 other is not an affair of absolute instability of one phase, but is a 

 relative instability shown only in the presence of the other phase; 

 this point was also mentioned in the paper just cited. 



One implication of the view that regards crystals as built up from 

 blocks of definite shape is to be especially insisted on. Only in excep- 

 tional cases will the edifice constructed from the blocks be such that 

 there are no unfilled crevices around the corners, and in no case where 

 there are two possible structures of different volumes will such empty 

 spaces be absent in at least one of the structures. These empty 



43 I. p. 86. 



