488 BELL SYSTEM TECHNICAL JOURNAL 



the conductivity. An equally logical assumption leading to a dif- 

 ferent result would be that the functional relationship between An 

 and pn be independent of the conductivity. Both are undoubtedly 

 incorrect but the error introduced may not be large for the better con- 

 ductivities. 



In obtaining curve B of Fig. 6 the values of a„ given in footnote 21 

 were used so that Eckersley's solution would be consistent with 

 Watson's solution for the perfect conductivity case. The values of 

 An were calculated on the assumption that the functional relationship 

 between An and pn be independent of the conductivity. If the mag- 

 nitude of ^ „ were assumed independent of the conductivity, curve B 

 would be raised approximately 7 db. 



