Intelligence and Miscellaneous Articles. 371 



value is not equally great for the two directions of the current 

 which excited the electromagnet. 



It will be seen from this statement that the heating or cooling 

 does not occur in such a way that we can regard the experiment as 

 a reversal of the transverse thermomagnetic phenomenon; from 

 aualogy with the Peltier effect the change of temperature must 

 be exactly opposite that which observation gives. The thermo- 

 magnetic action of the bismuth plate was seen to be very powerful, 

 and was quite normal ; it manifested itself in such a way, that from 

 the place where the thermal current entered the plate to the place 

 where the thermomagnetic current enters, the motion is opposite 

 to the direction of the held-magnets. 



In order to keep the results free from objection I placed the two 

 solderings of the thermoelement at two opposite places in the 

 middle of the long sides of the bismuth plate, but both plates 

 were carefully insulated from the plate by interposing thin sheets 

 of mica. 



After closing the current J only a slight alteration of the position 

 of the needle could be noticed, as both solderings had almost the 

 same temperature. The excitation of the magnetic field, produced 

 the action in the way given above, and with such regularity that 

 measurements could be made. In the homogeneous field, M = 6400 

 (C. Gr. S.) ; the deflections of the galvanometer needle i, correspond- 

 ing to the differences of temperature of the solderings, where the 

 strength of the current in the plate was I = 2-49 amperes^ was 



< A > i= ^} mean 9-1. < B > *-««- } mean _ I0 - 2 . 



(A) and (B) mean the two directions of the current in the plate ; 

 in (A) J flows from left to right through it. The two numbers 

 under each other are the deflections of the needle in parts of the 

 scale with opposite directions of the current which excites the 

 electromagnet, measured each time from the position of rest of the 

 needle when the magnetic field is not excited. In the first case 

 (the upper numbers), for an observer looking at the plate, the field 

 currents are in the direction of the hands of a watch (south pole 

 behiud the plate) ; the signs + signify that the deflections of the 

 galvanometer needle are towards the great or small numbers of the 

 scale. A deflection towards the large numbers represeuts an 

 increase of temperature of the soldering of the thermoelement at 

 the upper edge of the plate over the temperature of the other. If 

 no current flowed through the bottom bismuth plate (J=0), then, 

 w T hen the direction of the field current was reversed, there was no 

 action on the thermoelement. Tor J = 5*72 the mean of the 

 deflections at A was 17*0, at B 19-3 divisions, hence the difference 

 of temperature at the top and bottom edges (for the same M) is 

 proportional to the intensity, J, of the current in the plate. 



The observations were still more successful with a double plate of 



