Stalloy. 337 



4500. The /x,,, graph for " Stalloy " is therefore steeper for the 

 smaller inductions before the maximum value is reached, after 

 this the downward gradient is more gradual than for iron and at 

 the same time more constant. Tn general, for a given induction 

 [jlq is considerably greater for " Stalloy " than for iron. 



The effect of annealing is to increase the value of the char- 

 acteristic very largely. The magnitude of this increase will be 

 best seen in Fig. 1, the corresponding curves for the other 

 periods of alteration being very similar in appearance. 



The total loss (I.) in the ''Stalloy" shows a distinct im- 

 provement over that for iron. For all inductions and for all 

 frequencies the total loss is found to be less for "Stalloy " than 

 for ordinary iron of the same thickness. A very fair impres- 

 sion of the ■■ Stalloy " total losses will be obtained by taking 

 two-thirds of the corresponding values for ordinary iron at all 

 inductions. It is also important to notice that these losses 

 are greatly reduced by annealing. This reduction is not very 

 appreciable for inductions below 5000, but above that there is 

 a marked difference. It will be noticed however that the curves 

 intersect where the induction is 15,500 ; for inductions above 

 this the annealed losses increase very rapidly. In comparing 

 these figaires with those obtained for iron by Lyie it should be 

 remembered that we have included the effects of the higher har- 

 monics which in the unannealed ring were considerable. The 

 conclusions arrived at for these total- losses, namely that 

 " Stalloy " is superior to iron in this respect and that anneal- 

 ing very considerably reduces such losses, although discussed 

 for one speed only are fully borne out by the results for the 

 other speeds as will be readily seen by plotting curves for the 

 total loss from the other tables given. 



Fig. 2a shows the effect of annealing upon the phase lag of 

 of the third and tifth harmonics of the induction behind the first. 



Both i/fo and i/^- fall rapidly from a maxiniuni value for each as 

 the value of Bj is increased from zero, they then seem to reach 

 small limiting values when B, is increased to high values. Here 

 again as in the statical experiments the necessity of careful 

 handling was proved imperative. We obtain further evidence 

 of the instability of the unannealed " Stalloy," for in general the 

 experimental vari:.«tion from the graph is wider for it than for 



