428 Scientific Proceedings, Royal Dublin Society. 



magnetization were found to be appreciably higher than the corresponding 

 values for the left-hand series, showing that the susceptibility was greater 

 for these directions. One example may be given. For an applied spiral field 

 of 20 units, at a pitch-angle of sin a = 02, the longitudinal intensity of 

 magnetization was 138 for the left-hand spiral, and 230 for the right-hand 

 spiral. Similar observations have been made by Williams in his investiga- 

 tions on wires. A limit is, therefore, reached. It is not possible to place all 

 the observations in a theoretical framework, based on constant susceptibility 

 in all directions for a given field. The variations are not serious, being less 

 than 5 per cent, throughout most of the range examined, except in the 

 extreme case quoted above. If we wish to obtain exact concordance between 

 theory and experiment, we shall have to operate in a region where the 

 susceptibility is approximately constant, which seems to be the case from 

 45° to 90°, corresponding to a range of sin a to - 7 to 1/0, and sin 2a from 

 1-0 to 0-0. 



Determination of the demagnetizing factor. — It has already been pointed 

 out that there are no reliable experimental data for the demagnetizing factor 

 of steel tubes. Grotrian 1 arrived at the conclusion that it was the same as 

 that of a cylinder of equal diameter. Du Bois, 2 from observations on bundles 

 of wires, concluded that it was equal to the factor for a circular cylinder 

 of the same sectional area. The demagnetizing factor depends on the ratio of 

 the length /diameter, and from the values given by Schuddemagen, Grotrian's 

 theory gives iVxl0 3 = 2-2 - 5, while the theory of DuBois gives iVxl0 3 = 7-7 , 7. 

 It may be stated that neither theory is true, and that the true value of iVlies 

 nearer the mean of these values. The easiest method to demonstrate this is 

 to make use of the property of the anhysteretic curve of magnetization 

 obtained by superimposing a diminishing alternating field, as explained in the 

 earlier part of this paper. It was shown by Steinhaus and Gumlich that the 

 anhysteretic curve has an infinite slope at the origin. The curve obtained from 

 the steel tube has a finite slope, due to the demagnetizing factor. The curve 

 has to be sheared back by measuring the magnetic field from the line H= NI. 

 If the sheared curve is to have an infinite slope at the origin, the line 

 H = NI must be a tangent to the experimental curve at the origin. The 

 equation of the experimental curve for very small fields, therefore, gives the 

 value of N. It is found that the equation of the curve at the origin is 

 H = 0-033/ for magnetic fields < 0'2 c. g. s. units. As the deflections observed 

 were very small, larger values of H ranging from 0-2 to 0'6 c. g. s. units were 



1 Grotrian, Ann der Physik, vol. 52, 1892; vol. 54, 1894. 



2 Du Bois, Magnetic Circuit. 



