526 PROCEEDINGS OF THE AMERICAN ACADEMY. 



generally very different if this current has the same direction as its prede- 

 cessor or the opposite direction : that is, if the magnetism of the core 

 follows the hysteresis diagram in a direction corresponding to D P in 

 Figure 4 or in the direction D Q. An easy way of testing whether the 

 hysteresis curve of a large electromagnet has an insignificant area is to 

 obtain large oscillograph records of the building-up curves of direct and 

 reversed currents and to compare the two. I took, therefore, a series of 

 building-up curves for a gap-width of 35 mm., using currents of 2.75 am- 

 peres and 5.60 amperes, and found that in both cases the curves were 

 wholly indistinguishable 16 even when enlarged and superposed on the 

 screen, whether the current in question had the same direction as its 

 predecessor or the opposite direction. With this gap-width, therefore, the 

 magnet is an example of a circuit "containing iron 1 ' with an induction 

 flux for steady currents almost exactly proportional to the strengths of 

 these currents, and in this sense with a fixed inductance. If there were 

 no eddy currents, and no time-lag in the magnetization of the core, the 

 growth of the current in the coil should follow the law 



C= E {l-e- rtlL ), 

 r 



where L is this fixed inductance. Figure 10 shows the actual oscillograph 

 records for 2.75 amperes and 5.60 amperes in full line, and the theoretical 

 curve in dotted line for E = 80. In this case, where the " statical 

 effective " value of /x is independent of the current, but where eddy cur- 

 rents and what we may term time-lag in the taking up of the magnetism 

 by the iron may enter, it is interesting to see that in spite of the retarda- 

 tion due to eddy currents, the current in the main circuit builds up more 

 quickly than would correspond to the statical value of ^ when the current 

 is 2.75 amperes and that it starts to do so when the current is 5.6 amperes. 

 The building-up curves shown in this paper are careful reproductions 

 of oscillograph records, of which I have several hundreds. Some of 

 these were obtained with the aid of a Duddell Double Oscillograph, the 

 drum of which could be turned either by an electric motor from the 

 alternating street circuit or by clock-work, but most of them I got with 

 the help of two single instruments made by Mr. J. Coulson, who helped 

 me to take the photographs, and these served their purpose admirably. 

 One of them, which was used in measuring comparatively small induction 

 currents and needed to be very sensitive, was not quite aperiodic when 



16 For a similar case, see Professor T. Gray, Phil. Trans., Vol. 184. 



