OP ARTS AND SCIENCES. 



267 



iron or of nickel appeared in the calculations, Naccari (Atti della 

 R. Accademia delle Scienze dl Torino, Vol. XXIII., 1887) was 

 followed as an authority. He gives these values : — 



As any curve drawn, even when it runs smoothly and accurately 

 through the points given by observation, may not represent the 

 exact condition of things at intermediate points, and as there 

 is liability to considerable error in determining the gradient of a 

 line the curvature of which is everywhere changing, the conduc- 

 tivity was calculated for two points upon each curve of steady 

 flow, both points lying between the two thermometers nearest the 

 heated end of the bar. The difference between the two values thus 

 found from any one curve is not to be explained by variation of 

 conductivity with temperature. It is due, in the main, to errors 

 of experiment or calculation. As there are two calculated values 

 of conductivity from each curve of steady flow, and two such curves 

 for each bar, there are four calculations of conductivity for each 

 bar, but no two of these four are entirely independent, for the two 

 sets of observations on the rate of cooling of each bar were com- 

 bined for use with each curve of steady flow in that bar. The 

 results are here presented : — 



Southern Cast-Iron. 



