( 40i; ) 



f'oliuwiiig' till! method, iloricribed iu the "Vcrshig' der YcrguJ." ü[ April 

 21^t 1897, p. 493 i). 



As resistance- electrodes we used the same therino-elements which 

 were used in the measurements of the transverse phenomena, the 

 copper wires of which were again connected with one of the coils 

 of a dirtereutial-galvanometer. Fig. 1 shows the principle of this 

 method of observation. 



i^':= Element. 

 P=: Plate of 



bismuth. 

 A' = Rheotan- 

 wires. 

 Th := Thermo-elc- 

 ments. 

 IFj and Wr. = Resistance- 

 boxes. 



rig. 1. 



C\ = 0,258 



C„ = l.OOU 



Maa-uetic iield. 



Ferceutage increase 

 Observed ! Calculated 



Difference. 



46aü 



GlOO 

 9200 



9.7 

 14,5 



9,8 

 14,5 

 25,1 



+ 0,1 

 0.Ü 

 0,0 



From these observations we could also deduce an approximate 

 value for the specific resistance of this bismuth at about 20°, for 

 which 182,000 C.G.S was found. 



We ueed not be astonished at this number being much larger than 

 that of Fleming and Dewak, 116,000, which we used in the cal- 

 culations of the previous communication. For, our method, although 



') Colli m. Pliys. Liib; Leiden, N' 

 doctorate. 



37, p. 5. Compare also my thesis tor the 



