482 M. Hermann Herwig on the Independence of Temperature 



And yet, hitherto, in this direction very few statements have 

 been made, and, one may say, those scarcely in harmony with 

 one another. While some of the older observers hold that the 

 heat-conducting power is in general constant, Wiedemann and 

 Franz appear to admit for copper none, for Rose's metal a slight 



o 



variability of it with the temperature. Angstrom* gives for 

 copper a variability amounting to less than half, and for iron to 

 fully one half of the value of that which holds for electric con- 

 duction. Forbes f obtained for iron a still less variability than 



o 



Angstrom. 



In face of these various statements, M. Lorenz J has recently 

 asserted the independence of temperature of the conductivity for 

 heat of pure metals which remain homogeneous, and accounted 

 for the observed variabilities by the production of thermoelectric 

 currents in consequence of unequal heating of the metals. In 

 fact, iron (for which chiefly, according to the above account, a 

 variable power of conduction is found) is shown by the experi- 

 ments of Le Roux§ to be peculiarly favourable for such thermo- 

 electric actions. 



Accordingly it seems to me that it is still an open question 

 whether, and to what degree, the heat-conducting power of me- 

 tals depends on their temperature, and that, up to the present 

 time, two simple solutions of it have been proposed. These solu- 

 tions would be contained in Wiedemann and Franz's law, sup- 

 posed to be generally valid, that the heat-conducting power is 

 just as variable as the electric conduction-power, and, on the 

 other hand, in Lorenz's law, according to which the former 

 power is invariable for all temperatures. 



But if the position of the question be viewed thus, it is abso- 

 lutely necessary, for its decision, to submit to investigation 

 a pure metal which remains homogeneous. And of this kind, 

 perhaps, only mercury is known. Mercury can be obtained 

 purer than any other metal ; and no variations of internal struc- 

 ture can be admitted to take place in it in consequence of partial 

 heating. The latter is excluded by the experiments of Magnus ||, 

 according to which warm and cold mercury brought into contact 

 never give rise to electromotive action. Mercury would, accord- 

 ingly, be the most suitable metal for this investigation. On the 

 other hand, there are two circumstances opposed to the choice of 

 mercury — first, that the variability of its galvanic conductivity 

 is exceptionally less than that of the rest of the metals, and, 

 secondly, the fluidity of this metal. The former circumstance 



* Pogg. Ann. vol. cxviii. p. 429. t Edinb. Phil. Trans, vol. xxiv. 

 % Pogg. Ann. vol. cxlvii. p. 435 ; Phil. Mag. S. 4. vol. xlvi. p. fi5. 

 § Ann. de Chim. et de Phys. Ser. 4. vol. x. 

 j| Pogg. Ann. vol. lxxxiii. p. 495. 



