160 ANNUAL EEPORT SMITHSONIAN INSTITUTION, 1925 



individual shows small and perfectly characteristic divergences from 

 the mean, which do not become conspicuous until the liighest pres- 

 sures are reached. These suggest differences between the different 

 kinds of atoms or molecules, roughly corresponding to differences 

 of shape. If an atom has a '' knob " on it (local inequality in its 

 force field), it wnll interfere with its neighbors and produce per- 

 ceptible differences of behavior when the atoms are pushed close 

 together by high pressure. On the other hand, it is interesting 

 that water, which under ordinary conditions is a highly abnormal 

 liquid, under high pressures loses its abnormalities and returns to 

 normality. 



The kinetic theory of gases, extended to include the behavior of 

 liquids up to a few hundred atmospheres around the critical point, 

 has engaged almost exclusive attention up to the present time. There 

 is, so far as I know, no theory of liquids that, with any approach 

 to success, attempts to picture how the molecules in a liquid behave 

 when pushed into such close contact that there is considerable mutual 

 interference. 



The changes of volinne of the solid elements (metals mostly) are 

 in general much less than those of liquids, but they are nevertheless 

 significantly large, as in most cases it is possible to reduce the volume 

 by pressure to considerably less than it would be if the metal were 

 deprived of all heat motion b}' being cooled to absolute zero at 

 atmospheric pressure. 



Some of the metals are highly compressible, and coesium, in partic- 

 ular, loses more in volume at liigh pressures than ether. There is a 

 characteristic difference between liquids and solids in that solids do 

 not lose their compressibility nearly as rapidly at high pressures. 

 We would expect a difference of this sort; in solids, the atoms must 

 retain their regular crystalline arrangement at high pressures, so 

 that the free spaces around the edges and corners of the molecules 

 are never occupied and hence play a relatively small part in the phe- 

 nomena of compression. In a number of solids it seems that the per- 

 sistence of compressibility at high pressure can be explained only by 

 a compressibility of the atoms themselves. There is here an import- 

 ant problem for quantum theor}^ : To formulate the quantum con- 

 ditions when the electronic orbits interfere with each other as much 

 as they must in a strongly compressed solid. 



There is no adequate theory to explain the properties of metals. 

 In fact, the general nature of their structure is less well known than 

 is that of the structure of certain salts, compounds of two elements. 

 The data on compressibility now give us the means of finding the 

 first three derivatives of the forces between the atoms of a metal. 



