482 



BELL SYSTEM TECHNICAL JOURNAL 



however, whether the precision of the experiment would justify dis- 

 tinguishing between these two values. It will be noted that the effect 

 of refraction is appreciable and that the agreement between experi- 

 ment and theory is greatly improved by taking the effect of refraction 

 into account. 



As an indication of the effect of the finite conductivity of ocean 

 water, the theoretical curve for propagation over imperfectly conduct- 

 ing plane earth has been added in each case. Curve 4 for imperfectly 

 conducting plane earth is substantially the same as that for a perfectly 

 conducting plane for 0.8 mc. (Fig. 3), indicating that the effect of the 



50 



500 1000 



DISTANCE IN KILOMETERS 



5000 



Fig. 3 — Comparison between theory and experiment on 0.8 mc. Experimental 

 points show received field strength from S. S. America, March 1922, taken from 

 Fig. 16 of "Radio Transmission Measurements" by R. Bown, C. R. Englund and 

 H. T. Friis, Proc. LR.E. 11, 115-152, April 1923. 



Curve 1 — Theoretical neglecting refraction. 



Curve 2 — Theoretical assuming average refraction from meteorological data. 



Curve 3 — Theoretical assuming refraction to give best fit with experimental points. 



Curve 4 — Theoretical for plane earth taking finite conductivity into account. 



imperfect conductivity is negligible on this frequency. For 4 mc. 

 Fig. 4, curve 4, the effect of the imperfect conductivity while not 

 negligible is small compared to the effect of the earth's curvature. 



If an attempt be made to take into account the imperfect con- 

 ductivity of the earth by applying Eckersley's extension of Watson's 

 solution, curve 3 for 7^ = 1.45 of Fig. 4 would be moved almost l^ack 

 to curve 1 for K = \. There are, however, several reasons for cjues- 

 tioning this extension of Watson's work that will be discussed later. 



