210 AISTNTTAL REPORT SMITHSONIAN INSTITUTION, 1951 



number of stable modifications of ice, so thafe the two substances 

 would seem to be quite closely parallel. 



There are two other fairly well-known elements which, like bismuth, 

 are abnormal in that they expand on freezing. These are antimony 

 and gallium. The same question arises with respect to them: will 

 they also have other modifications under pressure, and will the melt- 

 ing-point curve of the new modification rise instead of fall? Of 

 these two substances antimony is in many respects much like bismuth : 

 chemically it is closely related, and it crystallizes in the same system. 

 Figure 4 shows that it, too, has a transition under pressure but at a 

 pressure more than three times that of bismuth — 85,000 atmospheres 

 instead of 25,000. It may well be, therefore, that bismuth and anti- 

 mony are analogous in their polymorphic behavior. Proof of this 

 must, however, w^ait until higher pressures can be commanded in the 

 laboratory for it would appear that the pressure scale of the phe- 

 nomena in antimony is more than three times as great as the scale for 

 bismuth. Also the melting point of antimony is much higher than 

 that of bismuth, so that it has not been possible to find experimentally 

 whether the melting point of the new antimony is raised by pressure. 



The other abnormal element, gallium, proves to fall into line as 

 well, and here the pressure scale is fortunately smaller. A new gal- 

 lium appears at a pressure of about 12,000 atmospheres, and the 

 melting point rises with pressure. Gallium and water are parallel in 

 that gallium has a high-pressure modification which is totally unstable 

 thermodynamically with respect to the other modifications ; similarly 

 there is a totally unstable form of ice that appears at temperatures 

 below 0° C. and at pressures above 4,000 atmospheres. 



Considering all three substances — water, bismuth, and gallium — 

 it would thus appear that pressure ultimately wipes out the ordinary 

 abnormal forms and they become normal, at least to the extent that 

 their melting point rises with increasing pressure. 



Figure 4 shows another example of polymorphism induced by pres- 

 sure ; barium possesses two transitions and three modifications. Poly- 

 morphism is indeed a phenomenon commonly encountered in high- 

 pressure research. Among the several hundred substances that I have 

 examined, about one-third show polymorphic transitions. The phase 

 diagrams of most of these substances have been examined, and it is 

 possible to show how the pressure at which a transition occurs changes 

 with temperature. Study of these phase diagrams reveals that many 

 of the modifications found under pressure are quite new, in that they 

 do not occur at atmospheric pressure at any temperature. Further- 

 more, there seems to be no tendency for the phenomenon to exhaust 

 itself as the pressure range is increased; on a statistical basis, the 

 probability that pressure will produce a new modification in an un- 



