288 Mr. H. Tomlinson on the Effect of Magnetization 



Provided that the susceptibility of bismuth be the same for 

 very high magnetizing forces as for low ones, it would follow 

 that if a bar of pure bismuth could be magnetized to unit 

 intensity* its resistance would be nearly trebled, whilst the 

 increase of resistance per unit would be at least 300,000 times 

 as great as the corresponding change of resistance in iron. 

 Considerations such as the above render it difficult to believe 

 that what is observed in bismuth in such experiments as these 

 is a real change in the specific resistance of the metal ; and 

 even with iron, nickel, and cobalt there seems to be evidence that 

 the whole of the observed change produced by magnetization 

 is not produced by mere rotation of the molecules. In the 

 author's experiments on iron and nickel he found that the 

 increase of resistance could be very closely represented by the 

 formula 



r J 



Ar . . 



where — denotes the increase of resistance per unit, and M/ 



and M,- are the magnetizing force and the magnetic intensity 

 respectively. From this formula it follows that, even if the 

 magnetizing force were so high that the ratio of increase of 

 intensity to increase of force was extremely small, the resist- 

 ance would nevertheless go on increasing very perceptibly 

 indeed with the force. The values of the coefficient h were 

 not very different in the two metals nickel and iron ; but the 

 coefficient a in nickel was about five times as great as in iron, 

 and was nearly double the coefficient h. It would be a matter 

 of considerable interest to ascertain whether it would really 

 happen that, when very great magnetizing forces were em- 

 ployed, the resistance of nickel would go on increasing very 

 perceptibly with the force t. 



If, however, on the one hand there is a difficulty in con- 

 ceiving how magnetization can so largely affect the true 

 specific resistance of bismuth, there is also an equal difficulty 

 in accounting otherwise for what is observed. Hall's pheno- 

 menon cannot certainly be credited in any of the experiments 

 which have yet been made with anything but a small fraction 

 of the whole observed effect. It is true that bismuth has been 

 found by Kighi and others to have a very large rotational 

 coefficient as compared with iron, nickel, or cobalt ; but 

 Ettingshausen and Nernst have shown % that whilst, with 



* This would require a magnetizing force of 71.000 C.G.S. units. 

 t The magnitudes of the forces used by the author never exceeded 

 200 C.G.S. units. X Loc. tit. 



