338 H. R. Hninley and A. L. Rossiter : 



the annealed " Stalloy." In one series of experiments in which 

 the readings were extended to an induction of about 20,000 the 

 analysis showed that all the (//-'s suddenly tended towards a zero 

 value but as the higher harmonics were so pronounced and the 

 wave form so distorted these results are nut included in the 

 tables or graphs. That these angles tend to zero value at the 

 high inductions may be seen by plotting the readings for Table 



III. 



Comparison with iron show that for extreme values, i.e., for 



very low and very high inductions the values of tlie ^'s are about 

 the same. But between these limits the " Stalloy " gradient is 

 less steep than that for iron especially for the higher frequencies, 

 so that in general for a given induction \p is greater for " Stalloy " 

 than for iron. Annealing as can be seen in Fg. 2a has the eifect 

 of increasing \^ for all inductions, this increase being more marked 

 in the higher frequencies. 



In Fig. 2b are given the corresponding curves for /i^ and b^. 

 These curves apparently start from the origin, rise rapidly until 

 Bi = 1000 and then steadily and gradually increase. The shapes 

 of the curves are similar to those obtained for iron, and the 

 values obtained about the same, a very slight increase being 

 noticeable. Annealing makes only a slight difference in these 

 characteristics for the low frequency, for the curves intersect more 

 than once, and what variation there is might almost be due to 

 instability. On plotting the figures for the higher frequencies, 

 however, we notice a gradual decrease as was the case for iron in 

 the values of b^ and b^ as we increase the frequency, as well as a 

 marked difference between the values for the unannealed and 

 annealed " Stalloy," which amounts in the highest speed taken 

 to about 10 per cent, of the latter. 



From a practical point of view Fig. 3 is perhaps the most 

 interesting. In this figure the A'alues of the total, hysteresis, 

 and eddy-current losses are given for the annealed ring. These 

 curves show^ an almost proportional increase of I and E with 

 increase of frequency. That this should be so for E is evident, 

 but on examination we find that I is given approximately in 

 terms of n and ^ for values of induction up to 11,000 by 

 the Steinmetz analogue proposed by Lyle for iron. Namely 

 for the annealed rinsr. 



