Thermal Conductivities. 17 



On the determination of the Thermal Conductivities 

 of bad conductors. By Charles H. Lees, M.Sc, 

 Bishop Berkeley Fellow of Owens College. Com- 

 municated by R. F. Gwyther, M.A. 



, {Received December 2nd^ i8go.) 



The experiments of Senarmont and others* have shewn 

 that anisotropic bodies possess different thermal conductivi- 

 ties in different directions, and that in the majority of 

 transparent bodies the conductivity in any direction in- 

 creases as the index of refraction of light in that direction 

 diminishes. The methods by which these results have been 

 obtained, differ little from that of Senarmont, which 

 depended on the melting round a central heated point, of a 

 thin layer of wax previously spread over the surface of a 

 crystal. In general, the area melted is an ellipse, the ratio 

 of the axes of which is equal to the square root of the ratio 

 of the conductivities along them. Such methods, therefore, 

 give no information with respect to the absolute values of 

 the conductivities, and a glance at the different results 

 obtained in successive experiments on the same crystal 

 shews that even the comparative values cannot be trusted 

 except as rough approximations. 



The recent determination by Prof Kundt of the indices 

 of refractionf of a number of metals, and the fact that he 

 finds they stand in the same order with respect to 



* References may be found in any text book, e.g.y Wiillner, Lehrbuch der 

 Experimentalphysik iii. , p. 310. 



+ Strictly we can only speak of an index of refraction of a substance when 

 Snell's Law sin z7sinr= constant, is obeyed. Kundt therefore defines the 

 index for a metal as the limit when i—Q of sin z'/sinr. The recent experi- 

 ments of Du Bois and Rubens {Phil. Mag. [5] XXX. 365) have shown that for 

 iron, cobalt, and nickel, Snell's Law is approximately true, but that there is 

 considerable deviation from it in silver and gold. 



B 



