1895.] in the Magnetic Permeability of Iron. 225 



A careful examination showed that this explanation was untenable. 

 The increase was found in transformers, the coils of which were 

 thoroughly dry and well insulated, and in which no change had 

 occurred in the separating material between the plates. This was 

 confirmed by removing the iron and trying it with new coils 

 specially insulated also by trying the old coils with other samples 

 of iron which had been separately tested. 



Another explanation suggested was that the effect was a sort of 

 magnetic fatigue that the iron deteriorated somewhat in the same 

 way as with steel springs, which suffer from " fatigue " after a great 

 number of extensions or vibrations. 



The gradual increase in the energy absorbed pointed to a physical 

 change in the iron, which behaved exactly as if it slowly hardened 

 and so became less permeable as if the softening results of the 

 process of annealing were being gradually lost. 



Effect of Annealing. On carefully re-annealing some of the iron by 

 heating it to redness and cooling slowly, it entirely recovered its' 

 original high permeability. 



Having thus found that the iron was not permanently or irre- 

 mediably affected although apart from annealing the effect seems 

 to 'be permanent it remained to ascertain whether the change was 

 due directly to the magnetic action, or whether it was caused by 

 some secondary condition, having the reversals of magnetism as a 

 primary cause. 



Effect of Heat. An investigation was therefore made as to the 

 effect of heat on the iron. 



Under the conditions of working, the temperature of the iron in 

 transformers is raised, both by the energy dissipated in the iron itself 

 by hysteresis and eddy currents, and also' by conduction and radiation 

 from the copper windings forming the primary and secondary con- 

 ductors. This rise of temperature varies according to the conditions 

 of working and the construction of the transformer; usually it is 

 from 20 to 60 C. above the surrounding atmosphere. 



In order to investigate the action of heat, as distinct from magnetic 

 or electric action, some samples of iron were kept for some months 

 in an oven maintained at temperatures varying from about 60 C. to 

 about 75 C. These samples were of iron obtained for transformer 

 construction and were built up in the manner of transformers, having 

 windings which were used simply for testing purposes. 



Method of Measuring. The coils used were of known turns and 

 resistance. The wire was of such a size as to make the C 2 B loss 

 negligible in comparison with the losses in the iron. The power 

 was measured by a very sensitive wattmeter, constructed by 

 Dr. Fleming, the volts by a Cardew voltmeter checked against a 

 Kelvin multicellular electrostatic voltmeter, and the current by a 



