( 73 ) 



rliffereiice of temperature (i. e. the difference of temperature per linear 

 cM. for a slope of potential 1 and a magnetic field 1) 



'' (1) 



and for the rotation of the equipotential lines occurring in the 

 IlALL-effect 



u^9n-v' _g^^^ ^^^ 



li 9n +V ffp 



In a similar manner he deduces : 



Coefficient of the thermomagnetic difference of potential 



9p9n (u + v) _ 



"S'n -\- vgp 



(3) 



Coefficient of the thermomagnetic difference of temperature (i. e. 

 the rotation of the isothermal lines) 



U V (go — (Jn) 



^^ ^^z=d (4) 



«</« + " ffp 



When these four phenomena have been observed, u, v, g,, and .</„ 

 may be calculated. Proceeding to this calculation, Riecke says : „Eiue 

 thermomagnetische Teinperaturdifferenz scheint hier (bei Wismuth) 

 nicht vorhandeu zu sein." Therefore he takes f/^. = ê^h, which consi- 

 derably simplifies the formulae. 



This remark however is not justified. As early as 1887 Leduc') 

 observed the rotation of the isothermal lines in bismuth ; this obser- 

 vation was corroborated by the experiments of VOK Ettingshausen"). 

 The rotation iiowever is not very large and hence one might think, 

 that Riecke's assumption does not remain far behind the truth. 

 In order to decide this question I have repeated the calculation with 

 the data used by RiECKE, augmented by von Ettingshausen's 

 result for the rotation of the isothermal lines. The latter found 

 with a plate of bismuth, 2,2 cM. broad and probably ^) 4,8 cM. 



') C. K. lO-J., p. 178't, 1887. 



2) Wied. Ann. 33, p. 135, 1S88. 



^) See VON Ettingshausen nnd Nebnst, VVied. Ann. 33., p. 477, 18SS, 



