(77) 



b. }lALh-effect. 

 The observations 

 pensation i). 



were made by means of the method of com- 



c. Transverse thermomacjnetic difference of iwtential. 



For measuring' this phenomenon the upper and lower borders of 

 the plate must be connected through the galvanometer; as moreover 

 the current of heat cannot be reversed momentaneously, we are 

 obliged to observe the change in the deflection on the reversal of 

 the magnetic field. These reversals occurred every minute; the 

 readings were taken just before reversing. 



In this measurement the rotation of the isothermal lines causes a 

 disturbance. Though the rotation is not considerable, yet the distur- 

 bances are not to be neglected ; for on the leversal of the field the 

 sign of the difference of temperature between the borders is changed ; 

 moreover we have to do with a thermo-electric couple bismuth- 

 copper; in fact the disturbances assumed sometimes a value of 60 

 scale-divisions, which would cause an error in the difference of po- 

 tential of 1800 C. G. S. units. 



The current of heat was obtained by conducting steam of 100° 

 through one of the tubes, and water at the temperature of the room 

 through the other tube. The slope of temperature, which we must 

 know in order to calculate the coefficient of this effect, is not the 

 same in different sections, in consequence of the loss of the heat 

 absorbed by the surroundings. Therefore I have measured the tem- 

 perature in the three places and found for A ± 34°, B 50° and 

 C 28°. Thence the slope of temperature given below was deduced 

 by graphical interpolation. 



') See communication of May 30, 1896, p. 47. Oomm. W. 26, p. 1. 



