eee 
Electrie Origin of Rigidity and Consequences. 423 
others in which the parts w? and W?6?/T? separately predo- 
minate. Thus when a metal melts, w? and the rigidity x 
vanish, because the electric axis in each molecule changes 
direction so quickly that the mean electric moment vanishes. 
For actions more rapid, small portions of a melted metal 
might exhibit high rigidity, all depending on Maxwell’s time 
of relaxation. In the next section the electrokinetic ener gy 
W*@/T° will play an important part in accounting for the 
mechanism of electric and thermal conduction in metals. 
4. The Electric Gyrostatic Property in Molecules and its Part 
in Metallic Conduction. Theory of Electric and Thermal 
Conductivity in Metals. 
We owe to Riecke (Wied. Ann. Ixvi. 1898) and to Drude 
(zbed. lxx. 1960) theories of electric and thermal conduction 
in metals wherein the positive and negative electrons act like 
the molecules of a perfect gas. Their theories are practically 
an application of the kinetic theory of gases to free electrons. 
These theories lead {0 some valuable results in the physics of 
electrons, as, for instance, when Drude finds that at a given 
temperature ‘the kinetic energy of a free electron is identical 
with that of a molecule of a perfect gas at the same tempe- 
rature. J. J. Thomson has sketched a similar theory of 
metallic conduction (‘ Nature,’ May 1900; Congres Inter- 
national de Physique, Paris, 1900, vol. iii.). These theories 
give the Wiedemann-Franz law of the approximate propor- 
tionality at ordinary temperatures between the electric and 
thermal conductivities of metals and the variation of their 
ratios inversely as the absolute temperature. In other words, 
the kinetic theory of electrons as gas makes the ratio of 
electric and thermal resistance at any temperature the same 
for all metals and directly proportional to the absolute tem- 
perature. But while the theory fares well with the ratio of 
the two conductivities, it makes no headway with an account 
of either conductivity taken separately. The reason is that 
the unfree electron pairs in the metallic atoms play a pro- 
roinent part in the mechanism of conduction, which will now 
be investigated. 
We have been led to assume in the last section that the 
pair of electrons in an atom of metal possesses electrokinetic 
energy by virtue of the rotation of the electrons round an 
electric axis; it must therefore have gyrostatic properties. 
Kach atom is equipped with an electrostatic doublet and an 
electrical gyrostat, both having the same axis. In a metallic 
conductor all the gyrostats vibrate in a field of electric force 
