8 Dr. J. G. Gray and Mr. A. D. Ross 



It seems necessary to add that this only holds for a series 

 of fields which are in ascending magnitude, and it is essential 

 that the specimen be initially neutral. For the method to 

 yield accurate results the hysteresis loop must be not only 

 closed but symmetrical. If, however, the neutral condition 

 of the specimen has been interfered with by magnetic or 

 thermal operations, the method is not applicable. This will 

 be apparent from fig. 3, which is intended to show the results 



Fiq-. 



obtained on applying the method to the examination of a 

 specimen which had been exposed to thermal treatment. On 

 establishing the field -fHi the induction became +Bj ; on 

 reversing the field the induction became -r-B 2 , and B 2 differs 

 widely from Bj. The change of induction is B 3 +B 2 , and 

 the kick of the galvanometer is proportional to this change, 

 and hence proportional neither to 2B } nor to 2B 2 . The 

 method of reversals is therefore not applicable unless the 

 specimen is in the strictly cyclic condition. On once more 

 reversing the field and increasing it to -f H 2 , the induction 

 followed the curve A'D'C. If the field +H 2 had been 

 applied in the first place, a widely different value of the 

 induction would have been obtained. TJius the magneti- 

 zation curve yielded by this method, when applied to a 

 specimen in the condition brought about by thermal treatment, 

 depends largely upon the procedure. 



Table II. shows the results obtained on starting and 



