Thermal Conductivity of Solid Insulators. 705' 



References. 



Bingham (1907). Anier. Chem. J. xxxviii. p. 91. 



Christoff (1906). Zeit. physik. Chem. lv. p. 622. 



Gilbault (1897). Zeit. physik. Chem. xxiv. p. 384. 



Handl (1872). Wien. Akad. Am. 1872, p. 125. 



Holmes and Sageman (1906, 1909, 1918). J. C. S. lxxxix. p. 1774 j. 

 xcv. p. 1919 ; cxiii. p. 2147. 



Nicol (1883, 1884, 1886). Phil. Mag. (5) xv. p. 91 ; (5) xvii. p. 537 ; 

 (5) xxi. p. 70. 



Thomson (1914). Phil. Mag. (6) xxvii. p. 757. 



Tjraube (1906). Ber. deut. chem. Ges. xlii. p. 86. 



LXXIII. The Thermal Conductivity of Solid Insulators. By 

 W. M. Thornton, D.Sc, JD.Eng., Professor of Electrical 

 Engineering in Armstrong College, JS T ewcastle-on-Tyne * . 



THE mode o£ conduction of heat in substances such as 

 quartz or paraffin-wax is shown, by their deviation 

 from the Wiedemann-Franz law, to be different from that 

 in metals. Free electrons are few, and heat energy consists 

 o£ elastic vibrations of atoms about fixed positions. From 

 the following experimental results it is probable that the 

 forces that control the elastic movement of atoms in solids 

 are those by which sound is transmitted. In the latter case 

 all movement in the wave-front is in the same phase; in the 

 transmission of heat there is complete confusion of motion, 

 and its rate of propagation is by comparison extremely slow. 

 Both movements are, however, quite definite in type, and 

 there should therefore be some relation between the con- 

 ductivity of heat and the velocity of sound in a solid r 

 insulating medium. From inspection of the recorded values 

 of these constants it was observed that while the square of 

 the velocity of sound V is equal to the ratio of the elasticity E 

 to the density p, the coefficient of conductivity of heat k is 

 equal numerically to their product, so that k = l$p = Y 2 p 2 . 

 This simple relation holds with surprising accuracy over a 

 wide range of solids, as shown by the following table. There 

 are few materials for which all the coefficients are known 

 and, with the exception of certain glasses, practically none 

 in which the measurements have been made on materials of 

 known composition or in the same state. The figures given 

 are the most probable values taken from the tables of Landolt 

 and Bornstein, Kaye and Laby, Everett, and Kempe ; and 

 for convenience reference is made to these rather than the 



* Communicated by the Author. 



