Hackett — The Twist and Magnetization of a Steel Tube. 429 



used, giving a line for which $H j SI = O40. These values rule out definitely 

 the theory of Du Bois and Grotrian. The demagnetizing factor for a steel 

 tube is approximately the mean of the factors for a solid cylinder of equal 

 radius and a cylinder of equal sectional area. 



Conclusion. 



The results of the investigation on magnetization have shown that the 

 ideal conditions required for the experimental proof of the theoretical formula 

 for the twist cannot be completely satisfied for all values of the pitch-angle. 

 As we have seen, the introduction of the demagnetizing factor makes a closer 

 approach to these conditions ; but then the variations in susceptibility become 

 apparent. It is not therefore surprising that it was not found possible to 

 demonstrate the variation of the twist with sin 2a for a constant spiral 

 magnetic field for all angles and intensity more exactly thau is shown in 

 fig. 2. A limited success was, however, attained with field of low intensity, 

 S = 10, which is interesting, as it gave an independent determination of the 

 demagnetizing factor. It has been shown that a spiral field whose components 

 are S=(S+ s) sin «, and F = Scos a, should give a net resultant field S, if a 

 suitable value of s be selected. 



A value of s = 1'65 was taken for S = 10, giving H = (10 + 165) sin a and 

 F = 10 cos a. The twist was observed f©r various pitch-angles in the neigh- 

 bourhood of 45°, when both fields were applied simultaneously, as in the 

 measurement of the "initial" spiral magnetization. The maximum twist 

 occurred for a pitch-angle of 47"5°. It was inferred from this result that a 

 set of fields given by H = (10 + 2.5) sin a and F = 10 cos a would bring the 

 maximum to 45°. The observations obtained are given in Table II, and it 

 will be seen how closely they follow the sin2a-law in the neighbourhood of 

 the maximum. 



Table II. 



Twist of a Steel Tube for a Spiral Field S = 10. 



Circular and longitudinal fields applied simultaneously. 



Comparison with the spiral curve for magnetization for the same fields, given 

 in fig. 4, marked S = 10, shows that the twist is symmetrical, even over the 



