398 Messrs. Kei Tokibe and Sukeaki Sakai : Effect oj 



Schaefer * and Benton f compared the values of rigidity 

 at ordinary temperature and at — 186° C. : that is, the temper- 

 ature of liquid air. F. Horton J gave us the most trustworthy 

 information regarding the question, in which he showed 

 that the rigidity of copper and steel decreases linearly 

 with the rise of temperature between 10° and 100° ; while 

 in the case of other metals, such as platinum, silver, and 

 aluminium, the diminution of rigidity per degree of rise 

 of temperature increases with temperature. Quite recently, 

 or rather at the same time as our experiment, Mr. 0. Imai § 

 measured the rigidity of steels at different temperatures up 

 to 700° C. with a similar apparatus. He found that the 

 rigidity decreases at first slowly and then somewhat rapidly. 



In the dynamical method of measuring the rigidity of a 

 wire, a weight of a given moment of inertia is suspended 

 from the lower end of the wire and set in torsional oscil- 

 lation. The period and logarithmic decrement of the 

 oscillation being measured, the rigidity of the wire can be 

 calculated. The damping of the oscillation is due to the 

 loss of energy into the surrounding air, as well as to the 

 internal viscosity of the suspended wire, the latter being 

 the principal cause. The effect of temperature on the 

 internal viscosity of the metallic wires has been observed by 

 most of the observers, who have investigated the modulus of 

 rigidity by the dynamical method — notably by Pisati, Tomlin- 

 son, and Horton. Pisati ||, who experimented over a range 

 of temperature from 0° to 300° C, found a marked increase 

 in the internal viscosity of metallic wares at high temper- 

 atures, except in the case of iron, in which the viscosity at 

 first decreased with the rise of temperature and reached a 

 minimum at 100° C, afterwards increasing again and be- 

 coming constant between 200° and 300° 0. Streint 1[ found 

 that the logarithmic decrement of different w 7 ires markedly 

 decreases if the wires be kept for long periods in continual 

 oscillation, and that it increases with the rise of temperature 

 and is less in annealed than in uriannealed wires. Thompson ** 

 came, however, to the conclusion that when the temperature 

 and the amplitude of oscillation are constant, both the period 



* Ann. der Phys. vol. v. p. 220 (1901) ; vol. ix. pp. 665, 1124 (1902). 

 t Phys. Rev. vol. xvi. p. 17 (1903). 

 % Phil. Trans. Roy. Soc. vol. 204. p. 1 (1905). 

 § The paper is not yet published. 

 (I Nuovo Cimento, loc. cit. 



•f[ Sitzb. der Wien. Akad. vol. lxix. Abtk. 2, p. 337 (1874) ; lxxx. 

 Abth. 2, p. 387 (1880). 



** Phys. Review, vol. viii. p. 141 (1889). 



