Gl'IDKD WAVK I'H()I>A(;.\T1()N" TlIHOrCll C.Y KOM VCXKTir MKDIA. Ill 1177 



for the transverse //-field. Hiis may be readily verified hy examining 

 the vectors (E^, + jEy)e~'"' and (H^ + jHy)e~'"', which represent the 

 transverse field vectors in the laboratory system. The transverse fields 

 are elliptically polai'ized at any point and the ratio of minor to major 

 axes of the ellipses are 



I I ^+ I + I ^- I 1 l\H^\ + \H_\\ • 



The ficld.s so far are normalized only hy the choice of a .sinii)lc foiin foi- 

 the function, Eo(r); all components may, of course, be multiplied by the 

 same arbitrary constant. There is some virtue to a normalization })ased 

 upon power flow. The power flow is given in unreduced units by 



f {EX //). dS. 



•'guide 



Using the scaled units of part I with 



r 



'actual ^^ / 



and // replaced by a/ ^ H to gi\'e it the same dimensions as E (e is 



the dielectric constant of the ferrite), the power flow l)ecomes in the 

 present variables 



a/- -^ [ (E^H^ + ^_//-) r dr, 



We shall normalize the E and H fields by dividing the values given by 

 equations (47) by /^'^ This makes the power per (isotropic wavelength 

 in the ferrite)" a constant. In Fig. 9 we show the normalized fields for 

 the TEii-limit and TMu-limit modes as a function of r for the case 

 To = 5.75, 1 p I = 0.6 and several values of a. The beha\ior of the modes 

 as a function of a and p for this rachus is shown in Fig. 14 of Pail T. 

 AVe also show the amplitudes for the isotropic cases, o- = p = 0. It may 

 be recalled from the discussion of cut-off points in Part I that, for the 

 T^I mode at this radius, when cut-off is reached at a = —0.4, the am- 

 plitude of the H^ field is overwhelminglj^ greater than that of the others. 

 Even when normalized to the same power flow, the field amplitudes 

 for a given a are undetermined to a factor of ±1. This factor has been 



