PARAMAGNETIC DIAMAGiNETIC SUBSTANCES 293 



two sides and its lower edge was suspended in the position of the 

 plate, and a known current C was sent through the tinfoil strip. 

 If the length of the lower edge was /, the vertical force was CZH, 

 since the action on the sides was negligible, being practically 

 horizontal. Hence H was determinate by the balance. 



The value of K for bismuth was found to be the same within 



(//) Elevation 



the limits of error with fields varying from 1600 to 10,000. The 

 following results were obtained : 



10/c 



Bismuth . . . 12'25 to 12-55 



Antimony . . . 0*714 



Sulphur . . . -0-765 



Marble . . -0'7 



Aluminium . . . +1*88 



Townsend's experiment.* The method was devised to 

 measure K for liquids, and was in principle like the first method of 

 v. Ettiiigsiwuseii. A tall jar, which could contain the liquid to 

 be experimented on, was placed within a solenoid which served as 

 a primary coil. Round the middle of this was a secondary coil. 

 There were a compensating primary and a secondary respectively in 

 series with the coils round the jar, and a galvanometer with 

 commutator to rectify the current was in the secondary circuit. 

 The two sets of coils were so arranged and so connected up that 

 when the jar was empty an alternating current of frequency 16 in 

 the primaries produced no current in the secondaries. If there are 



* Phil. Tnin*.. A. (1896), p. 533. 



