200 



PROCEEDINGS OP THE AMERICAN ACADEMY 



us to determine any difFerences in molecular heating, — if the heat 

 developed by magnetizing and demagnetizing i.s due to molecular heat- 

 ing. If it is due entirely to induction currents in the metals, the 

 slight changes in electrical resistance produced by small quantities of 

 sulphur, of phosphorus, and of carbon would be inappreciable in the 

 masses of iron which we used, and we should expect to obtain under 

 the same conditions the same rise in temperature for the different 

 specimens of steel. Our previous work* on cobalt and nickel must 

 therefore have been affected by some error. 



We next determined to ascertain if the heating was confined to 

 the surface of the metallic cores. Theory indicates this to be the 

 case, whether we adopt the hypothesis that the heat is due to magnet- 

 ization and demagnetization, or the hypothesis that it is produced by 

 induction currents. We could not find, however, any experiments 

 upon this point. The bars were prepared as follows. Each one was 

 bored one half its length. At the outer end of the hole a shoidder 

 was turned in order that a short piece of glass tubing could be ce- 

 mented in. One thermometer was j^laced in the mercury surrounding 

 the bar of iron, and another was hung in the hole in the centre of the 

 bar, the hole being also filled with mercury. It was difficult to dis- 

 tinguish between the conduction of heat and the evolution of heat. 

 The rise of temperature indicated by the inner thermometer, however, 

 was probably entirely due to conduction of heat, as can be seen by 

 comparing the amounts of mercury surrounding the two thermometers. 

 The following table exhibits the results obtained. In the third 

 column an arbitrary designation is given to the electro-magnetic coils 

 which were used in the experiments. 



* Proceedings of the American Academy, 1878-79, p. 114. 



