Length of Ferromagnetic Substances hg Magnetization. 393 



but in nickel the contraction was reduced to about two-thirds 

 of its ordinary value. A few years ago one of us* studied 

 the same effect in iron, tungsten-steel, and nickel at tempe- 

 ratures ranging from 18° C. to 100° C. In weak fields the 

 magnetic elongation oi' iron was slightly diminished by 

 heating ; but in strong fields it was increased. In tungsten- 

 steel the elongation was always diminished ; in nickel the 

 contraction was considerably reduced, except in weak fields, 

 in which a minute increase of contraction was sometimes 

 observed. 



In these experiments the range of temperatures was very 

 limited, so that the remarkable effect of high temperatures 

 was not observed. In the present case the experiments were 

 pushed beyond the critical temperature of iron ; in addition 

 to this, the change of length in liquid air was also examined. 



§ 1. Arrangement. 



The apparatus for measuring the change of length by 

 magnetization at high and low temperatures was substantially 

 the same as that used for the study of the efiPect of tension 

 on the magnetic change of length f (see figs. 1 & 2, PL XV.). 

 Repeated experiments showed that this arrangement gave 

 very consistent results, but that it was rather preferable to 

 flatten the surface of the vertically suspended wire in contact 

 with the rotating cylinder. 



To diminish the effect of tension on the magnetic change 

 of length, as well as the yielding- at high temperatures due 

 to tension, thick rods about 1 cm. in diameter and 21 cms. 

 in length were employed. 



In the experiments at high temperatures the suspended 

 weight was generally 1 or 2 kilograms ; its effects were con- 

 sequently almost insensible. To the upper end of the ferro- 

 magnetic rod a copper rod about 1 cm. thick and 25 cms. 

 long was jointed by means of a copper screw and then brazed. 

 The lower end of the rod was likewise attached to a similar 

 copper rod about 20 cms. long. The screwed part in each 

 end of the ferromagnetic rod was about 1*5 mm. This con- 

 nected system hung vertically from a stout support by means 

 of a brass stand with three levelling-screws. The support 

 on which the tripod-stand rested was provided with a brass 

 plate with a hole-slot-plane arrangement. The free end of 

 the bar was connected with a copper wire about 1*5 mm. 



* K. Honda, Jour. Sc. Coll. xiii. p. 83 (1900). 



t K. Honda and S. Shimizu, Jour. Sc. Coll. xvi. Art. 9 (1902) ; Phil. 

 Mag. iv. p. 338 (1902). 



