M. G. AViedemanu on ihe Transmission of Heat in Metals, 397 



According to this, a difference of temperature and a resistance 

 of passage arc only observed when heat passes from a better to 

 a worse conducting metal. No such difference M'as observed in 

 the passage of heat from a worse to a better conductor, nor in 

 the passage through the place of contact of two metal rods of 

 the same material. 



These contradictions suggested some deficiency in the method 

 of observation. 



If the thermo-electric element be applied to equally warm 

 points of heated bars of different conductibilities, the heat of 

 the point touched communicates itself to the element. Without 

 regarding the difference which results from the difference in the 

 specific heats of both bars, more heat will pass to the part cooled 

 by contact with the thermo-electric element in the case of the 

 better than of the worse conducting bar. The first, therefore, 

 will appear comparatively warmer. This error may cause the 

 differences at the places of contact noted in the above Table, 

 which are the more prominent the greater the difference between 

 the conductibilities of the two metals. The warmer the place of 

 junction is, the more these differences are seen, on which account 

 they scarcely appear in the passage of heat from a worse to a 

 better conducting metal, where the latter ought, properly, to 

 appear warmer than the former. 



This error can, however, exercise no disturbing influence on 

 the relative conductibilities of the bars themselves. For on the 

 supposition that the conductibilities of the bars do not vary at 

 different temperatures, the changes of temperature at the several 

 points of the same bar, caused by the application of the thermo- 

 electric element, must be proportional to the temperatures 

 themselves. 



Hence the relative proportion of the temperatures observed at 

 different parts of the bar remains unchanged, and this is alone 

 decisive in calculating the relative conductibilities. 



It was necessary, in obtaining reliable results as to the passage 

 of heat, to make the loss of heat of the bars on the apjilication 

 of the thermo-electric element as small as possible. For this 

 purpose the following arrangement of the experiments was made. 



llound bars of different metals, of 13-2 millims. diameter and 

 157 millims. long, were made, and accurately ground plane at 

 one end. A bar of iron and a bar of copper of the same dia- 

 meter, but 666 miUims. long, and a bar of bismuth equally thick 



