Variation of Thermal Conductivity with Temperature. 511 



300° hotter than the other. But no calculation of the varia- 

 tion coefficient of conductivity is likely to be possible by 



dk 

 means of equations from which the term -^ had been omitted. 



I am, Gentlemen, 



Your obedient Servant, 



Oliver J. Lodge. 



30. With the exception of the correction now indicated in 

 equations (1) and (3), the first fifteen sections of the paper 

 are quite unaffected by the slip, and may remain as they stand, 

 except that I have now a little more to say on the subject of 

 §§ 5-9. 



Professor Tait has been kind enough to send me a copy of 

 his researches on " Thermal and Electric Conductivity,"read 

 before the Royal Society of Edinburgh in March and June 

 1878 ; and I find that he has given up his enticing speculation 

 as to the inverse variation of thermometric conductivity with 

 absolute temperature — and in fact that he believes iron to be 

 possibly exceptional in the inverse connexion of conductivity 

 and temperature, all other metals which he has subjected to 

 experimental observation showing a slight increase of conduc- 

 tivity as the temperature rises. Prof. Tait's results are thus 

 in opposition to the results of Prof. Angstrom for copper ; 

 but since Prof. Angstrom, in the interpretation of his very 

 ingenious method of experiment, used the ordinary Fourier 

 equations, formed on the supposition that k is constant, and 

 that rate of cooling is proportional to excess of temperature, 

 Prof. Tait does not consider his observations competent to 

 decide a point as to the variability of k. Without venturing 

 an opinion of my own on the subject, it is evident that this 

 opposition is an additional reason for attacking the important 

 question of the law of the variation of thermal conductivity 

 with temperature. 



Prof. Tait finds that a linear function of the temperature, 

 k = a-\rbt, will express the value of the thermal conductivity 

 according to his experimental results, at least in their present 

 preliminary stage ; and we saw in § 8 that Prof. Forbes's 

 results for iron could be expressed nearly as 



h= -207(1 --00144:0; 



hence instead of the law of variation of thermometric conduc- 

 tivity, 



k _ A 

 cp~ b + t 



