166 ANNUAL BEPORT SMITHSONIAN INSTITUTION, 1925 



more suggestive in formulating a theory of resistance. Some im- 

 portant relations are brought out by the pressure tool that might 

 not otherwise be known. Thus, the fact that pressure increases the 

 resistance of solid bismuth, which is abnormal, but decreases that of 

 the liquid, which is normal, strongly suggests that the abnormal 

 behavior of solid bisnuith is connected in some way with a pecu- 

 liaritj' of the crystal structure. But, on the other hand, the fact 

 that pressure increases the resistance of both solid and liquid lithium 

 shows that the abnoi-mality is in some way more deep-seated for 

 this metal and is doubtless connected with some characteristic of 

 the'atom as a whole. The pressure effects bring out the fact that the 

 conduction mechanism is not as simple as we might have supposed 

 from the fact that the temperature coefficient of all metals is nearly 

 the same. 



There is, of course, no adequate theory of metallic conduction; 

 the classical free electron theory has had to be given up, and there 

 is no satisfactory substitute. I have myself made an attempt at 

 finding the significance of the various pressure effects, and have 

 been led by them to a very general conception of conduction which 

 I believe must be incorporated into the finally accepted theory. 

 This conception is that the atoms play an essentially positive part 

 in conduction and, in some way which we do not at present under- 

 stand, make it possible for electrons to pass from one part of a 

 metal to another. According to the classical conception, the role of 

 the atoms was entirely secondary anci negative; the atoms were 

 the source of the free electrons, but, having once provided the elec- 

 trons, their role was merely that of trouble makers, getting in the 

 way of the electrons and preventing them from moving about as 

 they wished. But it now api)ears that the electrons can not drift 

 about without the intervention of the atoms. It is as if the atoms 

 hand on the electrons from one to another when the}' are lined 

 up in certain wa^^s, or as if tliere were tracks between the intricate 

 maze of quantum orbits within the atom, along which the conduc- 

 tion electrons may travel, so that there is an opportunity for long 

 flights b}' the electrons when the tracks are properly aligned. All 

 of the normal and abnormal effects of pressure are understandable 

 in terms of a picture like this if we only suppose that the atom 

 itself, when brought into close proximity to its fellows, behaves 

 in the unsimple way that the facts of compressibility and poly- 

 morphic transition compel us to suppose. It is natural to suppose 

 that these definite tracks are in some Avay connected Avith quantum 

 conditions, and perhaps with high quantum numbers, but the de- 

 tails are too complicated for our working out at present. 



The effects of pressure on thermal e. m. f. I have not attempted 

 to incorporate into any definite theory, but they do suggest one 



