MAGNETIC ALLOYS 113 



range downward from 300° C. to 0° C. or even lower, according 

 to their composition. When cooled below their transformation 

 temperature in some cases the alloys are found to be nearly 

 or quite non-magnetic, whilst in other cases their magnetic 

 quality is greatly improved after passing through a thermal 

 cycle. These contradictory results have been shov^'n to depend 

 upon the manner of cooling. In general, the transformation 

 from the magnetic to the non-magnetic state takes place over a 

 range of about 50° C. The effect of annealing at or just above 

 the upper limit and subsequently cooling at a fairly rapid rate is 

 generally favourable, whilst annealing at or just above the lower 

 limit is exceedingly injurious. The effect of passing the alloy 

 through several such heat cycles is shown by the curves in 

 fig. I ; in each case the specimen was kept at the temperature 

 indicated during several hours and then allowed to cool in 

 air. If quenched from a temperature 50° to 100° above the 

 upper limit, the specimens were non-magnetic. 



As noted, the transformation temperature of the alloys 

 appears to depend on the percentage of copper present. One 

 specimen, containing about 70 per cent, of copper, can be picked 

 up from a cake of ice by a small magnet but cannot be picked up 

 from a table in a room at ordinary temperatures, as the upper 

 limit of its transformation range is about 20° C. and its full 

 magnetic quality is not attained above about 15° C. below zero. 

 This specimen, when tested at 0°, shows a rather low magnetic 

 permeability, becoming saturated at low values of the field ; it 

 exhibits almost no hysteresis loss, all of which phenomena are 

 characteristic of all specimens near their upper transformation 

 limit. 



Microscopic examination of a considerable number of speci- 

 mens has shown that extensive changes in the crystalline 

 structure always accompanies the changes in magnetic quality. 

 Fig. 2 is a micro-photograph of a specimen of good quality 

 when in its best condition. Fig. 3 shows the appearance of the 

 same specimen in the non-magnetic condition. 



No sample in which the bright crystals of fig. 2 were 

 lacking was at all magnetic under any conditions, whilst all 

 specimens showing these crystals were magnetic at tempera- 

 tures below the transformation temperature. Apparently the 

 magnetic quality depends upon some constituent of the bright 

 crystals which may take on or lose its magnetic quality without 



8 



