236 Mr. J. M. Baldwin on the Behaviour of 



4. The arrangement o£ apparatus was practically identical 

 with that already described by Lyle *, the iron ring and the 

 coils on it being replaced by a long solenoid, through which 

 the specimen (a long straight rod or bundle of wires) passed. 

 Hound the centre of the specimen a small secondary coil 

 was wound, by means of which the flux was determined. 

 The speed was regulated by altering the number o£ cells in 

 the armature circuit of the rotary converter and by altering 

 slightly the resistance in the field- magnet circuit. The 

 reducing factor of the galvanometer was determined by con- 

 necting a Clark cell to the galvanometer, with a megohm in 

 series, the current balance being used only to adjust the 

 current to the required value. 



The procedure in any one experiment has also been de- 

 scribed. In every case a full wave was taken, the ordinates 

 being 12° in phase apart. In the reductions, no notice was 

 taken of the seventh harmonic, which was always small and 

 could not have affected the results, its influence on the 

 amplitude being quite negligible, and its influence on the 

 phase being practically eliminated by taking the phase of 

 the first harmonic from four ordinates on each side of the 

 zero ordinate of the curve. In every case the third com- 

 ponent t was plotted, and from it the ninth harmonic was 

 taken, thus leaving the third harmonic. 



The wave forms having been analysed, the results were 

 reduced to absolute measure by applying to them the proper 

 factors % to reduce them to magnetic intensity and induction 

 respectively in the form 



11 = 11! [sin cot + h 3 sm s (cot — 4> 3 )-\-h 5 sin 5 (cot— fa}'] 



B = B : [sin(W — ^ 1 )+& 3 sin3(©^ — 3 ) + b~ & sin o(a>t — 5 )] 



the harmonics of order higher than the fifth being neglected. 

 From these the value of the iron loss per c.c. per cycle 



1=5*5.1 { S i n <9 1+ 3A 3 6 3 sin 3(0 3 -<£ 3 ) + 57iA sin 5 (0 B -£«)}. 



To obtain the amount of this energy loss which is due to 

 hysteresis, the eddy-current loss has been calculated from 



* T. R. Lyle, Phil. Mag. [6] vol. ix. p. 104 (1905). 

 t T. R. Lyle, Proc. Roy. Soc. Victoria, vol. xvii. pt. '2, p. 391 (Feb. 

 1905) ; Phil. Mag. [6] vol. xi. p. 25 (1906). 



\ T. R. Lyle, Phil. Mag. [6] vol. vi. p. 549 (1903). 



is 



