502 Mr. J. R. Clarke on the Thermal 



limited range) the specific heat at constant pressure, C . If 

 we suppose the liquid to be initially at temperature 0° and 

 at a pressure V t equal to the saturation pressure for 

 temperature t, the increase o£ entropy while the liquid is 

 heated to its saturation temperature under that pressure is 



I p . Using this expression we may apply known values 



of C to calculate the increase of entropy in the process 

 of heating the liquid. This integral gives the quantity 

 <j> w — [<f> ] where [<£ to ]o is the entropy of the liquid at tem- 

 perature 0° and pressure P^. For the purpose indicated 

 in the paper, namely, that of obtaining a " first approxima- 

 tion to the entropy-temperature diagram," the small quantity 

 \_(f> ] was ignored. Each constant-pressure line in the 

 diao-ram, along which the heating of the liquid was supposed 

 to take place, was treated as practically coinciding with the 

 liquid boundary curve : it actually lies slightly to the left 

 of that curve until the saturation temperature is reached. 



Yours faithfully, 



J. A. Ewing. 



P.S. There is apparently a clerical error in the second 

 expression for K w as printed on p. 212. It should read 



K ^C, + T(|)J.. 



LVIII. On the Thermal Conductivity of some Solid 

 Insulators. By J. "R. Clarke, M.Se. (Sheffield) *. 



IN a paper in the Philosophical Magazine for December 

 1919, Professor W. M. Thornton remarks that from an 

 inspection of the recorded values of the physical constants of 

 some heat insulators, it was observed that the thermal con- 

 ductivity of a substance is equal to the product of the 

 coefficient of elasticity and the density of the substance. 

 The agreement between this product and the thermal con- 

 ductivity was so pronounced, except in the case of some 

 complex organic bodies, that it invited further investigation. 

 Among the substances which showed this agreement were 

 various kinds of glass. The data used, however, were 

 obtained from tables of physical constants, and it is not 

 certain, nor even probable, that the sample of " flint glass," 



* Communicated by the Author. 



