90 
Proceedings of the Royal Society of Edinburgh. [Sess. 
both while the temperature was that of liquid air and after the normal 
temperature had been resumed. The specimen was enclosed in a glass 
tube which fitted into a cylinder of asbestos placed inside the magnetising 
solenoid. One end of the glass tube was closed ; the other was open and 
bent up so that the tube could be kept filled with liquid air. The curve 
(fig. 2, curve 2) obtained when the specimen was thus cooled to 190° C. 
showed an increase of about 25 per cent, in the saturation value of I, but 
otherwise did not differ materially from that obtained at the room tempera- 
Quenching I EST. 
ture (fig. 2, curve 1).* On retesting the specimen when it had warmed up 
to normal temperature it was found that the improvement had been only 
temporary. 
The specimen was now heated to 400° C. and quenched by plunging 
vertically into cold water. The magnetic quality was somewhat destroyed 
by this treatment, and there was a decided diminution in hysteresis (fig. 2, 
curve 3). [With steel, nickel, and cobalt similar effects have been obtained 
by the author, but in a diminishing ratio.] Immersion of the quenched 
alloy in liquid air produced a temporary increase of over 30 per cent, in the 
* The effect on the magnetometer due to the liquid air being magnetic was quite 
negligible. 
