RECENT ADVANCES IN SCIENCE 441 



The well-known fact that good conductors of electricity are 

 good conductors of heat also receives an explanation ; for just 

 as gas molecules tend to diffuse from a region where their 

 temperature, i.e. their kinetic energy, is high to a region where 

 it is low and thus transmit heat, so the free electrons at the 

 heated end of a metal rod tend to diffuse to the cool end. This 

 does not imply a motion of electrons as a whole ; equal numbers 

 of electrons diffuse in either direction, but those coming from 

 the warm end have more energy than those from the cool end, 

 and gradually impart some of their increment of energy to the 

 molecules at the cool end, so that we have transport of heat 

 (without transport of electricity) in one direction. Metals being 

 rich in free electrons ought therefore to be good conductors 

 of heat. Further, the well-known Weidemann-Franz law con- 

 cerning the ratio of thermal to electric conductivity can be 

 deduced. This ratio should be a constant for all metals and 

 proportional to the absolute temperature. Now there is in 

 fact for many metals a close approximation to such constancy 

 and dependence on the temperature on the part of the ratio 

 referred to. 



This hypothesis, however, plausible as it seems, is not free 

 from serious difficulties ; one of the most important concerns 

 the relation between the electric conductivity and the tempera- 

 ture. In pure metals the conductivity varies almost inversely 

 as the absolute temperature at ordinary temperatures ; in alloys 

 there seems to be no simple relation ; yet the above theory 

 leads to an inverse proportion between the conductivity and 

 the square root of the temperature, unless further postulates 

 are introduced concerning change of molecular structure with 

 temperature. There are other objections which cannot be 

 dealt with here. Suffice it to say that of late years the theory 

 has received some very adverse criticism, and as far back as 

 1900 Thomson advanced an alternative hypothesis which can 

 be found developed in his Corpuscular Theory of Matter. In- 

 deed, indications of the hypothesis may be found in his earlier 

 work, Applications of Dynamics to Physics and Chemistry . 

 However, this hypothesis is not as widely known as the " free 

 electron " theory, but recent work, it would appear, is likely to 

 bring the former into considerable prominence. 



Kammerlingh Onnes has recently made some remarkable 

 investigations on the resistance of metals at the temperature 



29 



