370 Intelligence and Miscellaneous Articles, 



p. 291) ; which is that electromotive forces are produced in a plate 

 of bismuth, antimony, cobalt, nickel, or iron, traversed by a current 

 of heat and brought into a powerful magnetic field in such a 

 manner that the lines of force cut the plate at right angles. We 

 called the phenomenon " thermomagnetic effect," and in bismuth we 

 observed both a transverse current, one, that is to say, at right angles 

 to the current of heat, and a longitudinal one, that is, acting in the 

 direction of the current ; the former, which is by far the stronger, 

 changes its direction when the direction of the current to which the 

 magnetic field is due is reversed. 



It was obvious to attempt a reversal of the phenomenon, that is, 

 to demonstrate an inequality of temperature in a plate traversed 

 by a voltaic current, which was produced by the action of magnetic 

 forces. I used a bismuth plate 3*1 centim. in length, 2-4 centim. 

 in breadth, and about 0*04 centim. in thickness ; stout copper 

 wires were soldered to the whole extent of the short sides of the 

 copper plate ; they served to transmit the current J through the 

 plate ; in the middle of one long side the junction of an argentan- 

 copper thermoelement was soldered to the plate, while the other 

 junction dipped in a vessel of water kept at the temperature of 

 the room. The thermoelement was connected with a reflecting 

 galvanometer. The current of two Bunsen's elements was passed 

 through the plate, the intensity of which could be varied by a rheo- 

 stat with a stout wire, and could be measured by a reflecting 

 tangent- galvanometer. Owing to heating of the plate, a deflection 

 of the needle is observed, which soon becomes stationary ; the loss 

 of heat by the plate is diminished by enclosing it in wadding. 



The plate was placed between the elongated pole-pieces of an 

 electromagnet. If the latter is excited, a change in the setting of 

 the needle is observed, and after some time (|-1 minute) it acquires 

 a pretty constant position ; on opening the magnetizing current, 

 the needle reverts to its former position. The difference of the 

 adjustments is obtained with sufficient accuracy to read off each 

 time half a minute after the opening or closing of the magnetizing 

 current. Experiment showed that the positions of the needle were 

 on opposite sides of the stationary positions if the direction of the 

 currents was alternated ; and further that the deflections changed 

 with the direction of the current J which traversed the plate. 



The direction of the deflections showed that the temperature of 

 the junction fastened to the plate was always raised, when we 

 reach the junction from the place where the current J enters the 

 plate by a motion in the same direction as the currents replacing 

 the magnetic field. If the phenomenon could be regarded as a 

 displacement of the stream-lines of the plate, by which the stream 

 was condensed in one part and dilated in another, the displacement 

 would take place in agreement with the ponderomotive action 

 according to Ampere's rule. The deflections of the needle on each 

 side are rather considerable (as much as 60 divisions), although 

 when the direction of the current J is unchanged, the absolute 



