HIGH-PRESSURE PHYSICS — BRIDGMAN 209 



to lower the melting point. This effect was found for water compar- 

 atively early ; the agreement between the experimentally determined 

 and the theoretically calculated lowering of the freezing point was in 

 fact one of the early triumphs of the then young science of thermo- 

 dynamics. Much later, I found that the melting point of bismuth is 

 similarly lowered by application of pressure, as would be expected. 

 What, it may be asked, can we expect to be the ultimate course of the 

 freezing curve of water as we raise pressure indefinitely? In the 

 case of water, it is found experimentally that, at first, increasing 

 pressure only accentuates the effect, for at higher pressures tlie ab- 

 normal increase of volume on freezing becomes larger, and therefore 

 the melting point is depressed at a continually accelerating rate. One 

 would expect that this could not continue indefinitely, and in fact it 

 does not. It was Tammann who first found how water extricates 

 itself from its dilemma. At —22° C. and 2,200 atmospheres ordinary 

 ice abruptly gives up the unequal struggle and collapses into a new 

 kind of ice, a very large decrease of volume occurring at the same time. 

 In fact, the decrease of volume is so large that the new solid has a 

 smaller volume than the liquid from which it freezes. Tliis means 

 that the melting point of the new form of ice increases with rising pres- 

 sure. Not only this, but, as the pressure continues to rise, the new 

 ice discovered by Tammann eventually becomes unstable in its turn 

 and is replaced by a succession of others, with still smaller volumes 

 and more rapidly rising melting points. In all, seven kinds of ice have 

 so far been discovered. The last of these may be heated without melt- 

 ing to the temperature of melting solder, provided a pressure of 45,000 

 atmospheres is applied. 



By analogy, it was expected that a new kind of bismuth would 

 appear at high pressures to replace the ordinary abnormal bismuth, 

 and that the melting point of the new bismuth would rise with pres- 

 sure. Search for this hypothetical bismuth was diligently made by 

 workers in the high-pressure field, but for a long time with no success. 

 Eventually the new modification was found, but at a pressure con- 

 siderably higher than had been anticipated, namely at 25,000 atmos- 

 pheres, more than 10 times the pressure which produces the analogous 

 transition in ice. Furthermore, there are still other transitions of 

 bismuth at even higher pressures, as there are for ice. Figure 4 

 shows transitions at 45,000, 65,000, and 90,000 atmospheres, in addi- 

 tion to the one already mentioned at 25,000 atmospheres. In reality, 

 the transition at 25,000 is double, there being two transitions so close 

 together — within a thousand atmospheres of each other — that it would 

 have been confusing to try to separate them in the diagram. Thus 

 altogether there are five known transitions of bismuth, or six poly- 

 morphic modifications of the solid. This is the same as the known 



