312 Mr. R. H. M. Bosanquet on 



caused so much friction that the cradle would rest in any 

 position within two or three degrees. Now there is a cen- 

 tral cup or thimble with a central wire, and a circular channel 

 round it, of small diameter, in which an excentric wire dips. 

 The friction is not objectionable. 



The course of experiment is this: — The bifilar is set at 

 right angles to the magnetic meridian. A current is sent 

 through, and the deflection measured. The current is mea- 

 sured by a new standard galvanometer, just constructed. It 

 has ten turns, of which any number can be used, parallel or 

 in series. 



The current is reversed, and the mean of the deflections, 



corrected for torsion, taken. The moment is then calculated 



from the constants of the bifilar. The magnetic induction 23 is 



then calculated, employing the observed distance between the 



poles. The magnetomotive force, (47rCn), is calculated. 



mi .• . magnetomotive force mi , .... 



lne resistance x is — -. — ; — s — -. . lne permeability 



magnetic induction ^ 



of the metal for the given value of 33 is then taken from Row- 

 land's table for soft iron (previous paper, p. 209). The 

 length of bar divided by this permeability is the resistance of 

 the metal, and is subtracted from x. 



The remainder is the resistance of the air-circuit for the 

 shape in question. Assuming that this resistance can be ex- 

 pressed in the form 2aR, we divide this remainder by R, and 

 obtain 2a, the air-resistance for unit radius. 



2«R / , where R ; is the radius of the steel, gives us the air- 

 resistance for the shape of the steel. This process is carried 

 out both for the whole bar and the separate pieces. These 

 air-resistances are then introduced into an equation which 

 states that the total resistances, joined up and separate, must 

 be inversely as the observed moments. Hence the resistance 

 of hard steel is found to be less than that of air in the propor- 

 tion of about 1 : 31. 



The resistance was then determined directly, by sending a 

 current through a coil wound about the steel itself. As large 

 a current as was safe was used. It increased the permanent 

 magnetism a little at first, but afterwards there was no further 

 permanent change. The resistance thus determined by tem- 

 porary change of moment was to that of air nearly as 1 : 32. 

 A former determination was rejected on account of a disturb- 

 ance in the permanent magnetism. There is no further room 

 for doubt that the behaviour of permanent magnetism in hard 

 steel is completely explained by the assumption of magneto- 

 motive force and resistance in the steel ; and if the resistance 

 be determined by division of the bar, the same numerical 

 result is obtained as by direct determination. 



