Self -Demagnetizing Factor of Bar Magnets. 737 



from the magnetic moment, and are presumably mean values 

 throughout the length of the bar, whilst our values o£ J are 

 the values deduced from the action of an exploring coil 

 wound round the equator of each bar, and presumably measure 

 the maximum value of c/. As the self- demagnetizing action 

 of a bar depends on neither the mean value, nor the maximum 

 value of e/, as we have seen, but on a mean that is impossible 

 to calculate unless the actual surface distribution of the 

 magnetism is known, it appeared to us preferable to take 

 the value of c/ that can be ascertained with precision at the 

 place where the self-demagnetizing force has its minimum, 

 namely the centre of the bar. 



One point of criticism on Biborg Mann's results may be 

 permitted us. To give us confidence in our results, we have 

 throughout used substantial bars of 1*128 cm. in diameter, 

 and have raised the lengths. Biboro- Mann used a single ■'■ 



© © © 



cylinder of iron 11*850 cm. in length, originally of a diameter 

 1*526 cm., therefore of a dimension-ratio of 7*76. This he 

 turned down successively to smaller and smaller diameters 

 until he reached a diameter of 0*237 cm., giving a dimension- 

 ratio of 50. How he contrived to turn so thin a wire is 

 remarkable. It would have a sectional area of only 0*0561 

 sq. cm. Further, while his cylinder was 11*850 cm. in length, 

 his magnetizing coil was only 30 centimetres long and 4 cm. 

 in diameter. The ends of his rod were therefore at points 

 only 2j diameters distant from open ends of the coil, where 

 therefore the value of the field would differ by some 

 2iper cent, from the value of the uniform field at the middle 

 of the coil. 



ii 

 Part III. — Experimental. On the Values of the Self- 

 Demagnetizing Factor for Bar-magnets of rectan- 

 gular CROSS-SECTIONS OF VARIOUS PROPORTIONS. 



We are not aware that any previous investigator has 

 determined the self-demagnetizing factor for square bars or 

 flat bars of rectangular section such as are often used in 

 magnetic work. 



A priori we should expect the self-demagnetizing factors 

 to be less than for bars of equal section of circular form and 

 equal length ; since the greater perimeter of the rectangular 

 forms is magnetically equivalent to giving to the end parts 

 a polar expansion, reducing the reluctance of the air-paths 

 of the external magnetic flux, and so bettering the magnetic 

 circuit. And such has proved to be the case. 



The experiments were made in exactly the same manner 

 Phil. Mag. S. 6. Vol. 17. No. 101. May 1909. 3 D 



