﻿Oscillations on Iron in a Magnetic Field. 117 



simple function of the gradient of the cyclic curve. The 

 curves show, moreover, that the maximum effect of the spark 

 is greater in the cycles of larger amplitude; and not only 

 because the hysteresis curves are steeper in large cycles than 

 in small, but also because in small cycles the effect o£ the 

 oscillations is, for some unknown reason, much smaller than 

 the diminished gradient would lead one to expect. 



It must be mentioned here that during the above measure- 

 ments the effect of changing the sign of the initial charge 

 given to the oscillation-coil was repeatedly tried. On no 

 occasion was there any perceptible difference in the magneto- 

 meter deflexion. The damping of the oscillations must 

 therefore have been very trifling. 



A matter quite different, but one of considerable interest, 

 is examined in the following short table of the effect on the 

 magnetometer deflexion of varying the length of the spark- 

 gap from which the oscillation-coil was excited. The obser- 

 vations here recorded were all made at that point of the 

 0*200 ampere cycle where the magnetic field is zero ; obser- 

 vations made at other points of this or other cycles gave 

 results the same in character. 



Table II. 





Length of Spark-gap 





Magnetometer 



in centimetres. 





Deflexions. 



0-024 





0-8 



039 





0-8 



053 





1-1 



070 





2-6 



087 





4-4 



This table is plotted in fig. 7. On the same diagram 

 appears the curve of Paschen's* observations of potential 

 difference and spark-length. 



The most important question to which the results set forth 

 in this paper can be applied, is that bearing on the real nature 

 of the phenomena observed. A number of observers have 

 summed up their views by the phrase " hysteresis is annulled 

 by oscillations. " The present experiments at first sight bear 

 out that conclusion anew. The effect of the oscillations at every 

 point of the cyclic curve is such as to carry the representative 

 point towards the curve that would be yielded by a substance 



* Table in J. J. Thomson's ' Eecent Researches,' p. 78. 



