270 Distribution of Magnetic Induction in a Long Iron Bar. 



value near the point of greatest demagnetizing force, i. e. about 

 13,000 ; but stress cannot be laid on this owing to the 

 reading-off from one curve to the other. It would seem, 

 however, that the assumption of a constant demagnetizing 

 factor is not quite satisfactory; consequently the magneto- 

 metric method with cylindrical rods, although extremely 

 useful for comparative work, must be used with much caution 

 in determinations of an absolute character. 



On looking at curves Q we note that the induction drops 

 most quickly somewhere in the neighbourhood of maximum 

 permeability, that is, the induction leaks out more when the 

 bar is the better conductor. This at first seems peculiar ; but 

 on referring to set T, one sees that in the cases where the 

 permeability is increasing as we go along the bar, the in- 

 duction keeps in the bar, and the quicker the permeability 

 diminishes the more the induction leaks out. In fact, the 

 leakage at any place depends on the average permeability of 

 the part nearer the end of the bar, and not on the value at 

 that particular place. 



An interesting relationship is brought out in Table II. It 

 will be noticed that the ratio of B OT and B^ ranges from about 

 2 per cent, less than unity to about 4 per cent. more. This 

 is not a great variation from unity considering the number of 

 operations to be gone through before arriving at the figures. 

 It would thus appear that the induction in the iron and that 

 in the air so arrange themselves as to give a mean induction 

 in the bar equal to that which the mean II would have pro- 

 duced had it been applied to the bar and no leakages taken 

 place. 



A very striking point is the great alteration in equivalent 

 length of the bar ; on the ellipsoidal assumption it is § the 

 length, and Kohlrausch's number for a cylinder is |k We 

 see that, far from this being a constant number, it varies from 

 *9 to *74 ; so that any determination depending on constancy 

 of this quantity will be somewhat vitiated. A direct expe- 

 riment was made to test the amount of displacement of the 

 poles as follows: — A piece of iron wire was inserted in the 

 magnetizing-solenoid used above, and a brass wire was attached 

 so that it could be slid up or down, always inside the solenoid. 

 It was first adjusted so as to be exactly opposite the needle of 

 a magnetometer. Various magnetizina-currents were then 

 sent round the solenoid, and the wire was moved along inside 

 until the maximum deflection was produced : the following 

 numbers were obtained ; — 



