HIGH-PRESSURE RESEARCH BRIDGMAN 163 



peratiire, transitions may take place with no volume change, and 

 others with no latent heat. As pressure rises, the slope of the transi- 

 tion line may either rise or fall, while the slope of the melting curve 

 always falls. There is a surprisingly large number of cases analo- 

 gous to that of water and ice, in which the modification stable at the 

 higher temperature has the smaller volume. We might be inclined 

 to think of this as an abnormal state of affairs, and might expect 

 that high pressure would tend to wipe it out, but such is not the 

 case, and these ice-type transitions persist at least to pressures 

 beyond our present reach. Another surprising fact is that in the 

 majority of cases the modification with the smaller volume is the 

 more compressible. This is the same behavior as that shown by 

 water and ice, and can not help being of significance for atomic 

 structure. It means that the atom has outlying parts which have 

 centers of force of strength sufficient, in conjunction with the out- 

 lying parts of other atoms, to build up an open-w^ork structure, 

 and that these outlying parts have more than usual stiffness, so 

 that the resultant open-work structure has a lower compressibility 

 than the more dense structure which the atoms are forced to assume 

 by a pressure sufficiently high to push the outlying centers of 

 force out of register. From some points of vieAv, the thought of 

 atoms as having ■•' knobs " is certainly suggestive, although so crude 

 a picture is repulsive to some persons. 



One feature of the behavior of polymorphic forms, which is 

 of great significance, may be studied to advantage wnth the high- 

 pressure tool. This is the speed with which one modification trans- 

 forms itself to another. A particular advantage of this tool is 

 the nicety and rapidity of control which it offers. Changing the 

 pressure on a two-phase sj^stem is equivalent to changing the tem- 

 perature, but, whereas the actual temperature can be changed 

 only slowly and nonuniformly throughout the interior of the 

 apparatus, and considerable time is required for attainment of 

 equilibrium, the effective temperature may be changed instantly 

 and uniformly by any desired amount by a slroke of the pressure 

 pump. By changing the conditions in this way in the immediate 

 neighborhood of a transition point, I have studied a considerable 

 number of solid transitions under pressure. There is an essential 

 difference betw^een a change from one solid to another and that 

 from a solid or a vapor to a liquid. If a vapor and liquid or 

 solid and liquid coexist, then at any definite temperature a single 

 definite pressure of equilibrium will be automatically established. 

 The ordinary kinetic mechanism makes this understandable. 



Equilibrium between vapor and liquid, for example, is main- 

 tained by the action of two streams of matter, one continually con- 

 densing on the liquid from the vapor, the other continually evapo- 



