106 C. Bar us — Apparent Hysteresis. 



the others equidistant from it. These results are shown in 

 figures 18, 19, 20, the field corresponding to (7=8 amperes. 

 Soft iron wire (Z=35 cm , Z>='02! cni ) was here used as above. 



In figure 18 for instance, beginning with the twist —90°, 

 the wire is magnetized twice, giving the temporary and the 

 permanent effect : then untwisted and again magnetized twice 

 giving the two corresponding effects; next twisted +90° and 

 magnetized twice as the figure shows ; finally untwisted again 

 and magnetized twice, etc., for succeeding double cycles. In 

 figure 19 the limits of twist are ±180°; in figure 20, ±360°. 



The general characteristic of these figures is the very large 

 zero displacement, or rather the large return displacements. 

 In other words, the slip phenomenon is much larger on remov- 

 ing twist than on adding it, positively or negatively. The 

 deflections are laid out on a larger scale for increasing twists 

 and they are in general symmetrical. Temporary slopes a are 

 usually negative and even at 360° not strongly positive. Per- 

 manent slopes are positive and about of the above values. The 

 following data may be given. When the lines a and o are 

 constructed either side of 0°, they are designated a' and h' . 



Table III. 

 Twist. a. 



— Magnetostriction of soft iron in longitudinal fields. Current, 8 amperes. 

 Field 800 c. g. s. units 



Displ. Displ. at Displ. 

 Cycle at elonga- non-mag. 

 b. slope. +a'. —a'. +b'. —V. 0°. tion. at 0°. 



+ 90° +2-5 



+ 6-3 +4-5 



— 2-0 —3-0 +60 -i-6-5 5 



4'7 2-0 



180 3'0 



5-4 4-2 



— 1-0 —l'O 5-3 5-5 8. 



5-5 5*2 



360 2-9 



3-9 3-4 



+ 0-5 +0-5 8-0 8 10 



4-0 8*5 



9. The next question at issue is the amount of displacement 

 due to magnetization at the zero of twist, which may be pro- 

 duced by merely twisting the wire to different angles and back 

 again to zero, without magnetizing it when twisted. These 

 results are also inserted in figures 18, 19, 20, and marked £T, ~N'. 

 They increase rapidly as the twist (without magnetization) 

 increases. At 360° it makes little difference whether the 

 twisted wire is magnetized or not, so far as the displacement at 

 the zero of twist is concerned ; in all cases, however, magneti- 

 zation at the elongation increases the slip at 0°. 



Finally a perfectly fresh wire was adjusted, twisted to 360° 

 without magnetization and then freed from twist. The result 

 (amount of displacement due to magnetization at the zero of 

 twist) did not differ essentially from the figure. 



10. In conclusion a few words on the interpretation of the 

 above phenomena may be added. The point of view taken has 

 already been indicated in § 1. In Maxwell's sense any defqr- 



