HALL AND AYRES. — HEAT CONDUCTION IN IRON. 307 



A comparison of these results with those obtained by the end to end 

 method with short bars, cut thickness-wise from the shib, shows that the 

 two methods gave almost identical results. Of course, it is possible that 

 the bars used in the two methods differed considerably in thermo-electric 

 quality, and that some error in one or the other method compensated for 

 and obscured this difference of quality ; but it is much more reasonable 

 to conclude that the slab from which all of the bars were cut had practi- 

 cally the same thermo-electric quality crosswise as thickness-wise, and 

 that the accuracy of each method of testing this quality is affirmed by its 

 concordance in results with the other method. The results of both 

 methods were used for plotting a line from which values of the copper- 

 iron thermo e. m. f. could be derived for purposes of interpolation. This 

 line is a curve ascending with increase of temperature, and slightly con- 

 cave upward. The divergence of this line from true rectitude is prob- 

 ably not very significant. There is in the corresponding curve for the 

 thermo e. m. f. of the copper-German silver junctions, described in 

 the preceding pages, a divergence of about the same relative amount in 

 the same direction. It is possible that this peculiarity of both lines is 

 due to some idiosyncrasy of the thermometers used in the thermo-electric 

 tests. The same thermometers were used in all these tests ; and tliere- 

 fore, as the method of calculation of conductivity involved the ratio of 

 tlie e. m. f. of copper-iron and copper-German silver, no final error as 

 to conductivity results from any small imperfections of these thermom- 

 eters. E. H. II. 



APPENDIX II. 



Measurement of Electric Conductivity of the Cast Iron. 



One of the 2 cm. bars described in Appendix I. was used for this 

 determination. Four copper wires were attached to this bar by elec- 

 trolytic deposit of copper. Two of the wires were about 0.08 cm. in 

 diameter ; these were attached to the flat ends of the bar, and served to 

 carry in and out an electric current of about 0.25 ampere. The other 

 two wires were much finer, about 0.018 cm. in diameter; these were at- 

 tached at two points about 1.7 cm. apart, each being about 0.15 cm. from 

 one end of the bai", and were used for making connection with a poten- 

 tiometer. The bar was submerged in oil during the measurements. 

 The temperature of the oil was controlled by water flowing through 



