2 ' ART. 6. K. HONDA AND S. SHIMIZU. 



1904, the same measurements were extended so as to include dif- 

 ferent intermediate temperatures between the ordinary and liquid 

 air temperatures ; lastly in the third experiment, which extended 

 from March 10 to May 17, 1904, the magnetizations at high 

 temperatures were measured. 



Our specimens consisted of five ferromagnetic metals, and 

 twelve specimens of nickel steels, kindly placed at our disposal by 

 M. Ch. Ed. Guillaume. They were all examined in the form 

 of ovoids (major axis = 20cm and minor axis = 1 cm). 



In the first experiment, the specimens were first annealed 

 for about 4 hours at 1000°C— 1100°C in charcoal fire, after they 

 were well wrapped in asbestos, and then gradually cooled. These 

 annealed specimens were tested at the ordinary, then at liquid 

 air temperature, and lastly again at ordinary temperature. In the 

 second experiment, the measurement was always commenced with 

 freshly annealed ovoids. Lastly in the third experiment, all 

 specimens were cooled in liquid air for about 15 or 20 minutes. 

 The measurements at the ordinary, and then at higher tempera- 

 tures, were carried out ; the measurements at different descending 

 temperatures down to the ordinary were also made. Thus all the 

 measurements, when set down in order, form a complete cycle 

 with regard to temperatures, whose limits lie between liquid air 

 temperature and 1200°C. The methods and the results of the 

 experiments are given in the following pages. 



I. FIRST EXPERIMENT. 



The apparatus for measuring the change of length was similar 

 to that used in our former experiment above referred to. The 

 ends of the ovoid to be tested were soldered to two short brass 



