CHANGE OK LENGTH BY MAGNETIZATION. 



which was stretched by a weight dipping in ;1 vessel of water. 



Heating was effected by means of an electric current. The 



heating coil was wound on a copper tube 40 cm long and 2.o cm in 



diameter well insulated with asbestos 

 paper. The coil Mas wound anti- 

 inductively two turns per centimeter 

 by a wire about 1 mm thick. Two 

 heating coils of the same dimensions 

 were prepared, the one wound with a 

 German silver wire and the other 

 with a nickel wire. The former coil 

 was used in the experiments below 

 700°C and the latter for higher tem- 

 peratures. The melting point of 

 nickel is about 1500°C, while its 

 magnetic property is lost at a tem- 

 perature below 400°C ; hence above 

 this temperature, the presence of the 

 metal does not at all disturbe the 

 magnetic field. By using a nickel 

 wire for experiments at high tem- 

 perature, we may dispense with a 

 costly platinum wire or foil. 



The heating coil was fixed by 

 screws to the upper copper rod ; while 

 in its lower end was a hole, through 

 which the lower copper rod passed 

 without being in contact. The air 

 current which might enter or escape 

 through this narrow opening was 



