CHAXGE OF VOLUME AXD OF LENGTH. 63 



have been exceedingly minute as the magnetizing field was very 

 small. Dr. Knott^^ has published several papers on the change 

 of internal volume of ferromagnetic tubes, showing that iron, 

 nickel, and cobalt are subject to the change by magnetization. 

 As our former result regarding the same question was somewhat 

 different, especially in the case of nickel, we have thought it 

 advisable to settle the discrepancy by fresh experiments. 



9. Iron and Steel (Fig. 2). — Preliminary experiments on soft 

 iron and steel ovoid showed that considerable increase in the volume 

 change takes place as the ovoids are annealed. The increase 

 becomes more significant as the field is made stronger. In steel, 

 the effect of annealing is greater than in iron. In strong fields, 

 the volume change of the annealed steel ovoid is nearly twice as 

 great as in the unannealed state. AYolfram steel is very little 

 affected by annealing as regards the volume change, but the 

 change itself is much greater than in nickel or iron. The 

 motion of the capillary meniscus in the dilatometer can be easily 

 followed by the naked eye. The curves in Fig. 2 have been 

 plotted from measurements made on annealed ovoids. 



10 Nickel (Fig. 2). — As specimens of nickel almost always 

 contain traces of iron, the change of volume will probably depend 

 on the chemical nature. In addition to this, the mechanical 

 process which the metal had to undergo before it could be brought 

 to a form suitable for experiment, nmst have substantially altered 

 its elastic behaviour. 



The nickel rod, which we used in the former experiment, 

 was hammered from a nickel plate to a prism of square cross- 

 section. It contained 1.7Ö <^/o of iron, besides traces of man- 

 ganese and carbon. The ovoids used in the present experiment 



1) Knott, Trans. Koy. Soc. Edinb. 38, 527, 1896; 39, 457, 1898. 



