100 



C. jBarus — Apparent Hysteresis. 



merits with a thinner wire of iron, this was kept submerged in 

 the tube of flowing water specified. The residual diminutions 

 of rigidity given by the line h in figure 1 may, therefore, be 

 of thermal origin. 



4. The decrement of rigidity due to the effect of circular 

 magnetization on a submerged iron wire (L = 41, D = *024, 

 C = 3) is so nearly like the above results for nickel, that a few 

 sample curves will suffice to introduce the subject. The 

 diameter being relatively small larger twists are admissible as 

 lying well within the elastic limits. 



In figure 3 the twists are alternately +180° and —180°, 

 though the observations begin and end with zero. The arrows 

 are a sufficient guide as to the sequence of results. The lines 



'- 0° +180' — -m° — HW^ 



I I I I 



Mian. 



a and h show the decrement of rigidity due to the initial and 

 subsequent circular magnetizations for each twist. In figure 4 

 three twists are successively applied. Lines corresponding to 

 the initial and final magnetizations are omitted. An inspec- 

 tion of all the figures shows that the displacements of the zero 

 due to twisting between two angles a and —a, is less than the 

 aggregated displacements due to twisting from a to 0° and 0° 

 to —a, under otherwise identical conditions. We have here 

 also the key to certain results in a previous paper in which 

 increasing positive twists return with larger positive displace- 

 ments and increasing (absolutely) negative twists return with 

 increasing negative displacements. 



