Circular Magnetization on Magnetized Iron Wire. 429 



fields the permeability is very small. Therefore the circular 

 field would have at first only a slight effect in diminishing the 

 longitudinal intensity. On the other hand, in lower fields 

 the permeability is many times greater, and hence the effect 

 of the circular field is much more marked, until the direction 

 of the resultant field swings around nearer to the direction of 

 the circular field, when the rate of change in the longitudinal 

 intensity becomes very slow. 



Temperature- Curves, fig. 2. — The curious wave which occurs 

 at the upper end of each high-temperature curve {in vacuo) 

 may be partly due to the sudden change in the temperature- 

 coefficient of iron at high temperatures *, and partly also, in this 

 particular case, to the effect on the vacuum of gas given off 

 from the wire. The pump may not have been able to exhaust 

 at a sufficiently high rate. 



Hopkinson (Phil. Trans, vol. clxxx.) investigated the resist- 

 ance-temperature curves of soft iron and steel at high tempe- 

 ratures up to 900° C. The temperature was inferred from 

 the resistance of a copper wire enclosed with his specimens, 

 apparently on the assumption of a constant temperature- 

 coefficient for copper. He found a sudden drop in the tem- 

 perature-coefficient for soft iron and steel between 800° and 

 900° C. beyond the critical point. It seems desirable that 

 this should be tested up to higher temperatures by comparison 

 with a platinum pyrometer. 



Method of Distinguishing the Effects of Temperature and 

 of Circular Magnetization. 



By treating the ordinates of the curves in fig. 1 as one 

 component of the resultant intensity the temperature-variation 

 of the magnetization of iron at high fields can be worked out 

 to a fairly accurate result. 



The first step in the reduction was to obtain a family of 

 curves {a) of average resultant I and H at different tem- 

 peratures. These were compared with a similar set (b) from 

 which temperature-effect had been eliminated. Then by 

 treating the drop between corresponding curves of the first 

 and second set — at the same resultant fields — as due to tem- 

 perature, the temperature-effect on the resultant intensity 

 was obtained. 



The average circular field in the wire was taken equal to 

 two thirds of the field at the periphery. This was compounded 

 with the longitudinal field to give the average resultant field 

 due to the two magnetizing forces. The longitudinal field 

 previous to compounding was corrected for the effect of the 



* A rapid increase of a similar character was observed by Callendar to 

 occur just below the critical point. 



