1891.] upon the Magnetisation of Iron and other Substances. 115 



trace of polarity, and care was taken to prevent new polarities being 

 given to it by accidental contact with the needle. On heating the 

 magnetite in the wire loop, as in experiment (a>, the influence of 

 temperature was more marked than with the iron, as the needle 

 receded towards the magnetic meridian before the magnetite was 

 visibly red-hot, and advanced again very readily when the source of 

 heat was removed. 



(d.) A small rectangular prism of nickel, 0*2 inch long and 0'05 inch 

 across the sides, was submitted to the magnetometer as in the previous 

 experiments, when the increase of temperature in diminishing the 

 magnetic power of the nickel was most pronounced at the temperature 

 of melted tin, and the metal became quite insensible to the needle at 

 a point much below the red heat. 



(e.) A rectangular prism of pure cobalt, of the same dimensions 

 as that used in the previous experiment, was submitted to the action 

 of the magnetometer, when, contrary to expectation, the needle 

 advanced towards the cobalt before it became visibly red-hot, and 

 remained stationary when the temperature was raised to redness ; 

 thereby showing an increase of magnetic power- of the cobalt with 

 increase of temperature. 



The magnetic behaviour of the cobalt was so remarkable as to 

 induce me to make further experiments upon the metal with more 

 powerful magnetising forces. 



A cube of pure cobalt from the Chemical Museum of the Owens 

 College was kindly placed at my disposal for these experiments by 

 Professor H. B. Dixon, F.B/.S. The cube was 0'3 inch across 

 the sides, and a short piece of platinum was screwed into the centre 

 of one of its faces for suspension from the balance over the large 

 electro-magnet. The temperature of the cube, below the red heat, 

 was determined by the fusion of small fragments of zinc and tin, 

 placed in a conical recess drilled into the upper face of the cube. 



Similar cubes of nickel and malleable iron were prepared for com- 

 parison with the results obtained with the cobalt. 



The method of experimenting was as follows : The cube of 

 magnetic metal was suspended over the electro-magnet, excited by a 

 current of 20 amperes, and while in this position was heated by an 

 oxyhydrogen flame until the requisite temperature was attained. 

 The cube was then quickly brought into contact with the pole of the 

 electro-magnet, without any intermission of the heating blast, and 

 the magnetic force was measured by the weight required to detach 

 the cube from the electro-magnet. The cube was reduced to the 

 temperature of 76 C. by immersing it in a bath of solid carbonic 

 acid and ether, whilst suspended over the electro-magnet. 



The results of these experiments with the magnetic metals are 

 given below : 



i 2 



