290 Mr. H. Tomlinson on the Effect of Magnetization 



gated with great completeness the influence of magnetization 

 on the therm oelectrical properties of steel of different tempers. 

 Finally, Ewing has entered very fully * into the changes 

 effected by longitudinal magnetization in iron when under 

 different amounts of longitudinal stress. Thomson has also 

 shownf that nickel is rendered by longitudinal magnetization 

 thermoelectrically positive to unmagnetized nickel ; whilst the 

 author has found J cobalt when under longitudinal magneti- 

 zation to be negative to the unmagnetized metal. 



The experiment now about to be described was made nearly 

 at the same period as the experiment on the effect of mag- 

 netization on the electrical resistance of bismuth and with the 

 same bar. This bar was 25 centim. long and 0*33 centim. in 

 diameter; it was placed in the axis of a magnetizing solenoid, 

 S, specially constructed to avoid imparting heat to the mag- 

 netized core§; a preliminary examination proved that there 

 was certainly no error arising from this cause. The ar- 

 rangements are sufficiently shown in fig. 1, where S is the 



Fig.]. 



.. s BATTERY \ j. 



GALVANOMETER 



solenoid and A B the bar. The bar was encircled by two 

 little air-chambers, C and D, through one of which steam was 

 passed, and through the other water at a temperature of about 

 16° C. The ends A and B were connected by copper wires 

 with a very sensitive Thomson's reflecting-galvanometer, and 

 were well buried in sawdust. 



Since a bar of bismuth can never be obtained in a perfectly 

 homogeneous condition throughout its whole length, there was 

 a considerable thermoelectrical current already in existence 



* Phil. Trans. Part ii. 1886. f Loe. cit. 



I Proc. Roy. Soc. No. 241 (1885). 



§ For a description of this solenoid see Phil. Trans. 1883, loc. cit. 



