158 Dr. E. H. Hall on the " Rotational Coefficient " 



several times greater, and that in platinum several times less, 

 than the effect in gold or silver. The fact of greatest interest, 

 however, was that, if we called the direction of the transverse 

 effect in iron -f , that in the diamagnetic metals, and in nickel 

 and platinum also, would be — *. 



In view of this remarkable disagreement in behaviour be- 

 tween the two strongly magnetic metals iron and nickel, it 

 seemed highly desirable to make a quantitative investigation 

 of the effect in nickel as soon as possible, and extend the ex- 

 amination to the other strongly magnetic metal cobalt. Most 

 of the experiments to be described in this article relate, there- 

 fore, to nickel and cobalt. The examination of the latter was 

 a hasty one, and may well be described first. 



No thin strips of the metal being at hand, a slice was sawn 

 from a small block of moderately pure cast cobalt and worked 

 into the form of a cross. To the extremity of each arm of this 

 cross was soldered a thin strip of copper 2 or 3 centim. long, 

 for the purpose of making the electrical connexions. The 

 cross of cobalt with the copper strips attached was now fast- 

 ened with hard cement to a strip of glass and worked down 

 with a file to sufficient thinness. Before placing the cross 

 upon the glass its thickness, and that of the glass also, was 

 measured by the calipers. After cementing the two together, 

 the total thickness was found, and, again, the thickness of the 

 whole after the cross had been filed down. The thickness of 

 the cross in its final condition was thus estimated at '45 mil- 

 lim., to which value an uncertainty of perhaps 10 or 15 per 

 cent, attaches. 



With this apparatus it was found that the direction of the 

 transverse effect in cobalt is +, i. e. the same as that in iron. 



Mx V 



As to the magnitude of the effect, — ^— was found to be 



44 x 10 10 , placing cobalt between silver and iron. The speci- 

 men of cobalt used, however, contained some nickel (how 

 much is not known accurately): and this doubtless counter- 

 acted in part the effect of the cobalt. It seems probable, 

 however, that, allowing for all errors, the transverse effect in 

 cobalt is less than that in iron, other things being equal. The 

 magnetic field used was about 9000 (cm.-gr.-sec), stronger 

 than has yet been used with iron. 



* These signs are given to avoid tedious repetitions. I have here called 

 the effect in iron -f- simply because its direction in this metal is that which 

 the conductor itself bearing the current would follow, if free to move 

 across the lines of magnetic force under the action of the ordinary " pon- 

 deromotive " force. No significance further than this is at present 

 attached to this choice of signs, 



