Stress and Strain on the Properties of Matter. 495 



The mean value of all the observations on the mean thermal 

 ■capacity of German-silver between 20° C. and 60° C. is 0-09424, but 

 this number requires to be slightly corrected if we are to assume 

 the thermal capacity of water at 0° C. to be unity. According to 

 Regnault,* the thermal capacity of water at f 0. == 1 + 0-00004* + 

 0-0000009^, and hence the thermal capacity at 20° 0. = 1-00116. 

 Again the apparent weight of water in air was taken instead of the 

 weight in vacuo. Now the correction of the apparent weight W of 

 a, mass of water when weighed with brass weights amounts to 

 W . 0012 (1 - = w • 0-00106. On the whole, therefore, we 

 must take 0*09424 x 1-00222, i.e., 0*09445, as the corrected thermal 

 capacity. A similar correction was applied in the case of all the 

 other observations. No correction was made in consideration of the 

 fact that the whole of the stem of the thermometer employed in the 

 calorimeter was not immersed in the source of heat, as according to 

 Kopp, if N be the number of degrees extant from the source of heat, 

 T the observed temperature, and t the atmospheric temperature, a the 

 correction to be applied, 



«=N(T-Q x -0001545.f 



Now if we assume the temperature of the air to be 20° C, and the 

 temperature of the water before and after the immersion of the 

 heated wire to be 18° C. and 22° C. respectively (this representing 

 rather an extreme case), the correction would only introduce an altera- 

 tion in the value of the thermal capacity of - 015 per cent. With 

 the thermometer used in the heated air-chamber a correction, not 

 however exceeding in any case 0*3 per cent., was, when necessary, 

 applied by the aid of Kopp's formula. 



In Table I are given the results obtained with the different metals, 

 and in Table II a comparison of these results with those obtained by 

 other experimenters. 



* " Mem. Acad. Sciences," xxi, p. 729. 



f Eecent investigations have shown that this formula can only be regarded as 

 roughly approximate, but the approximation is quite sufficient for the present 

 purpose. 



