﻿Oscillations on Iron in a Magnetic Field. 119 



finally whether or not any considerable amount of magnetic 

 energy is available and drawn upon. Allowing mentally for 

 the oscillation energy lost in Joule effects and in radiation, it 

 seems probable that magnetic work is done by the oscilla- 

 tions — whose efforts, therefore, would appear to be attended 

 with hysteresis loss, just as are those of slower variations of 

 magnetic field. 



Table IT. and fig. 7 give, in some degree, support to this 



Fig 7. 



A 



</3 - 



H 











Ql 











7.5 











XO 







■^ 





15 





^ 



vS-^^ 





to 



*^^^^^ 









5 

























; " 







•02.-03 04- -05 06 O? "08 -09 'O 



Spark Length C-ms 



conclusion. The bent (experimental) curve may be taken as 

 representing the magnetic work done by sparks of length 

 shown by the abscissae, other circumstances being the same. 

 The almost straight line represents the potential to which the 

 oscillation-coil was charged at the corresponding spark-length. 

 The relation between the two curves shows that for some 

 portion, at any rate, of the diagram the magnetic work done 

 varies as the square of the charging potential, that is, the 

 electrostatic energy of the initial charge. The bending away 

 of the first curve towards the right is explicable by the con- 

 sideration, that as the applied oscillations get more violent 

 the outermost layers of the iron wire attain, so to speak, a 

 saturated condition; that is, become incapable of being more 

 greatly impressed — with the result that the magnetic effect of 

 the oscillations begins to fall off in amount. It was for this 

 reason that in the experiments of Table I. the spark-gap was 

 kept at the value 0'53 millimetre, well within the part of the 

 curve free from this saturation effect. 



I have to thank my colleague Mr, J. Lister for deducing 

 the gradient curves and drawing all the figures in this paper. 



