C. Bancs — Apparent Hysteresis. 101 



The curves given correspond to mean currents. The scale 

 of the phenomenon increases with the strength of the current. 

 For 1 ampere, figure 5, the decrements of rigidity due to the 

 initial magnetizations line aa are still of marked value ; the 

 subsequent effects (line bb) lie within the errors of observation. 

 For 6*5 amperes all decrements have been greatly magnified as 

 seen in figure 6. 



Finding that a more systematic series of results would facili- 

 tate the discussion below, I have added figures 3' to 6', for the 

 same sample of (submerged) wire and a circular field (current 

 6'5 amperes), and for twists increasing from 45° to 360°, 

 applied to each wire. These were 35 centims. long and *024 

 centims. in diameter. The cycles were many times repeated 

 as the figures show at once. The striking feature of all these 

 experiments is the remarkable constancy (for the same wire 

 and current) of the slopes of the cycles, i. e., of lines connect- 

 ing the means of initial and subsequent deflections. This 

 slope shows no certain variation between 45° and 360° of 

 twist. The marked contrast between the temporary slopes, a, 

 due to the first magnetization and the permanent slopes, 5, due 

 to the ensuing magnetizations is sustained. All slopes are 

 negative. 



The displacements of the fiducial zeros or slips observed at 

 the end of the cycles do not increase as fast as the twists. 

 They show their chief increase between 45° and 90°. To 

 better exhibit the displacement of the zero I have given the 

 successive scale readings in the zigzag curves c, attached to 

 each figure. The position of the spot is marked off vertically 

 to scale, and the successive observations are charted at equal 

 distances apart horizontally, for distinction, and joined by 

 straight lines. The field is alternately made and broken begin- 

 ning with no field. When the sign of the deflection (positive, 

 up) and that of the twist (given by the attached sign) agree, 

 rigidity is increased. The curves show that rigidity is dimin- 

 ished on closing the current and increased on breaking it. 

 Moreover, after the first closing, the diminution of rigidity is 

 enormously greater than on subsequently closing the current 

 for the same twist. This is particularly true for small twists. 



5. The endeavor must now be made to elucidate these phe- 

 nomena. With regard to the relatively small decrement of 

 rigidity obtained in the circular magnetizations succeeding the 

 first (i. e., the permanent decrements), I have already shown 

 (1. c.) that they may be reproduced in brass wire by the same 

 means ; and until specially refined experiments are made, they 

 may be regarded as the mere result of the accession of heat. 

 The phenomena are thermal and elastic in their nature. So 

 much, therefore, for the slopes of the lines, b. 



