Hackett — The Twist and Magnetisation of a Steel Tube. 427 



which, on this theory, should give a resultant field 8 = 10. It is seen that the 

 linear relation with sin a is closely followed for the interval - 7> sin a <l - 0. 

 The deviations increase as the pitch-angle diminishes. This is evidently not 

 due to the demagnetizing effect. Experiment shows that the longitudinal 

 magnetization remains too small, even if the correction term is greatly 

 increased. We have reached the limit of the correction for the demagne- 

 tizing factor, and we shall have to explain this result by a failure in the other 



-1000 



800 



600 



400 





200 



A 6 



Initial spiral magnetization. 

 Fig. 4. 



assumptions. Either (1) the susceptibility is not the same in all directions, 

 or (2) the condition of simultaneity is not satisfied, and the circular com- 

 ponent grows up appreciably more rapidly than the longitudinal component 

 of the field. 



Experiments show that the first alternative gives a sufficient explanation. 

 All the above observations were on left-hand spiral magnetization. On 

 reversing the circular field the values of the twist for right-hand spiral 



