of Refraction of the Metals. 17 



crystallized. The thin films of metal constituting the bismuth 

 prisms, however, showed under the microscope no trace of 

 crystalline structure ; therefore the conductivity of the thin 

 uncrystallized film of bismuth used might very well be a quite 

 different one, and better than that of a crystallized rod. The 

 ratio of the conductivities for heat and electricity has also been 

 found to be very different by different observers. 



We may therefore, on the basis of the above numbers, come 

 to the conclusion that the electric conductivity in metals is 

 really proportional to the velocity of propagation of light- 

 waves of long period. In order to test this supposition, it 

 would be necessary to determine the index of refraction of 

 very long waves and the specific conductivity for the same 

 piece of metal, or at least for metal electrolytically deposited 

 in the same way. Whether such determinations could be 

 carried out at all accurately must remain for future discussion. 



If we conclude that the above proportionality has been 

 proved, at least approximately, by our experiments, there must 

 exist a very intimate connexion between the velocity of light 

 and the conductivity for heat in the metals ; for however 

 great the differences between individual observations may be, 

 it at least follows, from a comparison of the whole of the de- 

 terminations at our disposal, that the conducting powers for 

 electricity and heat in the metals are proportional to one 

 another. 



There thus exists at least a near proportionality between 

 the velocity of light, electric conductivity, and coefficient of 

 conduction of heat in the metals. This remarkable relation- 

 ship indicates a connexion between the motion of light in the 

 metals, the motion of electricity in the galvanic current, and 

 of heat in a current of heat. 



I have endeavoured to find an explanation of this by the 

 assumption that the conduction of heat in a metal depends 

 essentially upon radiation from one layer to the neighbouring 

 one, in which the radiation takes place with the velocity of 

 light in the metal in question ; and that, on the other hand, 

 what we call electricity moves in a metallic conductor tra- 

 versed by a galvanic current with the velocity which light 

 has in the metal. How far this view is tenable must be de- 

 cided by more extended experiments than I have as yet been 

 able to carry out. 



In conclusion I would further remark that, if the approxi- 

 mate proportionality shown in our experiments between 

 velocity of light and galvanic conductivity really exists, 

 certain consequences would follow which are capable of being 

 tested by experiment. I will only here mention two of them. 



Phil. Mag. S. 5. Vol. 26. No. 158. July 1888. C 



