292 Intelligence and Miscellaneous Articles. 



electromagnet is established, a permanent galvanic current is then 

 observed in a galvanometer connected with the electrodes of the 

 plate, which lie nearly on an isothermal line. The direction of this 

 current changes with the direction of magnetization, and with the 

 direction of the current of heat in the plate ; if the bismuth is 

 heated on both sides, the action of the magnet disappears. 



The electromotive force which produces the current is propor- 

 tional to the strength of the magnetic field, to the distance of the 

 electrodes, probably also to the fall of heat along the plate ; it 

 seems independent of the thickness of the plate. 



It was obvious to suppose that the cause of the electromotive force 

 might be thermoelectrical : the temperature of the two electrodes 

 (the copper wires soldered to the bismuth plates) might have been 

 altered under the influence of magnetism. Direct experiments 

 with thermoelements, which were placed carefully insulated be- 

 tween two plates traversed by a current of heat, showed no change 

 of temperature in consequence of magnetic action ; nor was there 

 any when, instead of electrodes, thermoelements (argentan-copper) 

 were soldered to the plates ; the electromotive force was also seen 

 to be independent of the nature of the electrodes. Hence there is 

 no deflection of the heat current in the bismuth plate in conse- 

 quence of magnetic forces. 



If the electrodes lie in the direction of the current of heat, if 

 they are therefore anisothermal, and if the thermoelectric current 

 which takes place between them even without a magnetic field is 

 compensated, then, when the field is produced, an electromotive 

 force is produced in one or the other direction, but mostly varying 

 in strength. 



In eight bismuth plates from different sources, the direction of 

 the " transverse " currents, that is the " therm omagnetic currents " 

 at right angles to the current of heat has been ascertained to be the 

 same ; the current flowed in such a direction through the plate 

 that from the starting point of the current in the plate, to the 

 starting point of the current produced, we get a motion opposed 

 to the direction of the current producing the field. Only in 

 one plate, in preparing which the metal was rapidly cooled, was 

 there seen a different behaviour ; after melting and slow cooling 

 this bismuth also came within the above rule. 



As respects magnitude, we observed that using a magnetic field 

 of the absolute strength 5000 (C.G-.S.) in an almost rectangular 

 plate 5 centim. in the side, and 1*9 millim in thickness, which on 

 one side was heated by means of a hot copper strip, while the other 

 side was cooled by ice- water, an electromotive force of -^j of a volt 

 was observed. 



The direction of the " longitudinal " electromotive force, which, 

 as we observed, did not change with the field, was such that the 

 galvanic current flowed in the plate from the heated to the colder 

 electrode; yet here also individual differences seem to prevail. 

 The longitudinal effect was feebler with the magnetic forces applied 



