Stalloy. 339 



I = (.00106 + .0000226«)§'''-' 

 and for the unannealed ring 



I = (.00184 +.0000331«)^i '• 



These forms agree fairly well with experiment except in a 

 few isolated cases occurring chiefly in the unannealed sample. 

 The value of the exponent for each case is remarkable, for 

 it seldom falls below 1.57 for iron. 



A fair average equation for transformer iron was found to be 

 I = (.00175 + .000027«)^i •'' 



The results for eddy-current losses found by us fully sub- 

 stantiate the claims made by the makers of " Stalloy." Tlie 

 value of the specific resistance was found slightly higher than 

 that given by other experimenters but 50,000 may be taken as 

 a fair working value. • These discs, which are much thicker 

 (.075 cm q.p.) than those usually used in iron work, show an 

 eddy-current loss much less than that calculated for iron of only 

 half the thickness, the value being about one quarter of the 

 total loss. Seeing that for " Stalloy " I is itself less than for 

 iron, the former must be considered admirably suitable for 

 transformer work. The values of E increase rather considerably 

 with increase of frequency the relative increase being greater 

 than for I. This fact will be seen from Fig. 3 which gives 

 all the losses to the same scale. 



Perhaps the most remarkable results are those obtained for 

 the hysteresis loss U. The values for the middle frequency (.035) 

 only are plotted in the figures ; but by examining the Tables I. and 

 III. it will be found that the values for the other frequencies prac- 

 tically coincide with those from which the graph is plotted. Results 

 for ordinary iron show an increase of ~ U with frequency which 

 although small is still much greater than the maximum variation 

 for "Stalloy " wliich occurs between the frequencies .035 and .02. 

 The values of U for frequencies .07 and .035 are practically identi- 

 cal. For " Stalloy " therefore U seems to be independent of 

 the frequency and only about half the magnitude of that ob- 

 tained for ordinary iron. 



The values of the " kinetic hysteresis " I-U-E are given in 

 the tables. On plotting, these results will be found to form 

 smooth graphs which reveal an increase of magnitude with 

 increase of frequency. Here again the values are less than 



