1884.J Stress and Strain on the Properties of Matter. Ill 



thermal capacity." The following table shows that the rate of increase 

 per unit of thermal expansibility is at C., and therefore at any 

 temperature, equal to the increase per unit of the " variable capa- 

 city " . 



As a consequence of the above-mentioned relations we can from a 

 knowledge of the rate of increase of the "variable thermal capacity " 

 determine the expansibility at any temperature between two limits, 

 provided we know the mean coefficient of expansion between those 

 limits. Again, should the rate of increase of torsionability prove for 

 all metals, as it is for iron and copper, equal to the rate of increase of 

 expansibility, we shall, by the method of torsional vibrations, be able 

 to determine with any degree of accuracy we please any one of the 

 three rates of increase. 



It is shown in the paper that the thermal capacity per unit mass is 

 nearly two and a half times the " real capacity," so that only two- 

 fifths of the whole thermal energy which we may impart to a mass of 

 metal goes towards raising the temperature, the remaining three- 

 fifths being expended in internal and external work. The external 

 work is practically insensible in ordinary cases. Of the internal 

 work, that expended against bulk-elasticity amounts in the limiting 

 cases from j-^^ to T^^O~O ^ the whole, and, though greater than the 

 external work, is almost insensible ; moreover, there seems to be no 

 relationship whatever between the whole thermal capacity per unit 

 volume and the work done against bulk-elasticity. 



Raoul Pictet has concluded* that the amplitude of the oscillation 

 of molecules around their positions of equilibrium may be taken as 

 corresponding to temperature, and in the case of several metals has 

 shown that 



T X /3 X =a constant, 

 * " Nature," 1879, p. 356. 



