164 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1925 



rating from the li<iui(l into the vapor. When the velocity of the 

 two processes is tlie same we have equilibrium, and if equilibrium is 

 disturbed the two processes restore it. But with two solids the 

 behavior may be (althou«;h not necessarily) quite different, in that 

 two solids ma}^ coexist indefinitely in contact at a definite temperature 

 anywhere within a range of pressure. This is different from the 

 familiar persistence of a modification in a region of thermodynamic 

 instability which involves viscosity. In such a system it is usually 

 supposed that equilibrium w'ill be automaticall}' produced if the tem- 

 periiture is raised to the point of thermodj^namic equilibrium and 

 the two phases are in contact. The mechanism by which a definite 

 equilibrium is automatically set up does not exist in the solid. In 

 the solid it is probably near the truth to imagine the atoms as tied 

 to certain mean positions of equilibrium and as vibrating about these 

 positions, but never departing from them by more than a definite 

 amount. In the liquid or vapor, on the other hand, we ha\e some- 

 thing like Maxwell's distribution of velocities, and it is possible to 

 find a few atoms with any velocity that we please, no matter how 

 high above the mean. Suppose noAV two modifications of the solid 

 to be in contact. If an atom is going to change from one modifi- 

 cation to the other it will have to free itself from its first position 

 before it can settle down into the second. If the act of freeing itself 

 from the first position invohes too much of a departure from the 

 mean, the atom will not be able to make the change, even if the new 

 position, when once attained, is a position of greater stability. But 

 by a sufficiently large change of external conditions the atom may 

 be helped to free itself from the first position, and so attain the sec- 

 ond. I have studied in some detail the way in which the width of 

 the " domain of indifference " varies with pressure and temperature, 

 and also the variation of the velocity of transition outside the do- 

 main. Both show large and very different sorts of variation for 

 different substances. 



At present we have absolutely no theory of polymorphism; we 

 are acquiring by X-ray analysis a descriptive knowledge of the dif- 

 ferent atomic arrangements of polymorphs in some cases, but we 

 have no explanation to oll'er of why, under some conditions the atoms 

 build themselves into one kind of structure and sometimes into an- 

 other. For instance, who has any adequate explanation of why 

 diamonds are so rare and graphite so common? There is here a 

 curious example of the indirect progress that phj'sics sometimes 

 makes. It would seem to be a much simpler problem to explain 

 the ways in which an atom may combine with others of its own kind 

 than wnth those of a different kind, whereas in fact we have at least 

 the beginnings of an understanding of a great man}' chemical com- 

 pounds, but can not claim anything of the kind for polymorphs. I 

 can not help feeling that an understanding of the great multiplicity 



