8 AEÏ. 4. — K. HONDA AND T. TERADA. 



by applying a magnetic field to the specimen and observing the 

 deflection of the mirror corresponding to different fields. 



Our procedure in the experiment was usually as follows : — 

 The wire to be tested was first annealed for about 4 hours in a 

 charcoal fire, and then gradually cooled. To avoid surface oxida- 

 tion during the process of annealing, the wire was buried in a 

 fine powder of charcoal, and sealed in an iron tube, which was 

 placed in a furnace. In this w^ay, a small quantity of carbon 

 might enter into the wire, but the oxidation was quite inappre- 

 ciable. Variation in the diameter of the wire caused by the 

 oxidation was avoided in the above manner. 



To either end of the wire thus annealed, was brazed the 

 brass rod before mentioned, and the whole was suspended in 

 position as shown in the foregoing figure. Then the mirror and 

 pulley-systems were placed in position. For avoiding a conti- 

 nuous slight vibration of the mirror, as well as for the rapid 

 damping of the vibration due to loading or unloading the wire, 

 it was necessary that a glass rod fixed to a stand should come 

 lightly in contact with the bifilar cords near the friction-wheel 

 q. The addition or removal of a weight from the pan was 

 occasionally observed to cause a lateral displacement of the image 

 in the mirror ; this was, however, easily effaced by adjusting the 

 clamp of the carriage to the rod. 



The magnetic change of length under constant tension was 

 first measured. Since the hysteresis effect of tension on the 

 -length change was found to be considerable, the tension was 

 varied cyclically between zero and its maximum value before be- 

 ginning the experiment, in order to remove any initial effects. 



According to the results of our experiments, the ascending 

 branch in the curve of elongation due to tension, lies a little 



