2 ART. 9.— K. HONDA AXD S. SHIMIZU: CHANGE IN LENGTH 



increase with tension. Cobalt is practically unaffected by tension. 

 B. Brackett,'^ G. Klingenberg'^^ and K. TangP^ also investigated the 

 same subject and obtained results similar to those of Bidwell. 



In Bidwell's experiment, whieli is generally regarded as the 

 most reliable, the wire to be tested carried the magnetizing coil with 

 it, so that even the smallest tension was greater than 3 kilograms j)er 

 square millimeter. Hence the effect of small loading, Avhich is 

 remarkable in nickel, was not well studied. The reason which led 

 him to adopt such an arrangement, was, according to his statement, 

 to avoid the disturbance due to the electromagnetic action between 

 the wire and the magnetizing coil. Moreover the sensibility of his 

 apparatus could no longer be considered to be sufficiently delicate. 

 It was, therefore, desirable to repeat his experiment with an arrange- 

 ment giving higher accuracy. 



A few month ago, Professor Nagaoka and one of us measured 

 the mag-netic elonu'ation of the nickel steel, kindlv placed at oui- 

 disposal by Dr. Ch. Ed. Guillaume. It showed a remarkable anomaly 

 Avith reofard to the mafjnetic elono'ation. Much interested bv the 



coo »J 



result, we proceeded to examine the effect of loading on the magnetic 

 elonjration of the allov as well as other ferromag^netic metals. 



2. The apparatus used in the present experiment is, in 

 principle, tlie same as that used by Professor Xagaoka. The chief 

 difference consists in using a rotating cylinder*^ to cause a reflecting 

 mirror to turn through a minute angle, instead of the three pivots 

 system. 



1) Brackett, Phy. Rev. [5] 5, 257, 1897. 



2) Klingenberir, Inaug.-Diss., Berlin, 1897; Beibl. 21. 897, 1897; Inautc.-DLss., Rostock, 

 p. 34 ; Beibl., 23. 270, 1899. 



3) K. Tangl, Drud. Ann. 6. 34, 1901. 



4) H. Hertz. Tnstrumentenkundte, 3, 17, 1883 ; Gesammelte Werke. 1. p. 227. 



