216 BHLL SYSTEM TECHNICAL JOURNAL 



tliis explanation is the necessity for rather high conductivity to account 

 for the propagation of \va\cs to great distances without large ab- 

 sorption. 



In 1912 there appeared an article by Kccles ' in which the bending 

 of waves around the surface of the earth was explained on the basis 

 of ions in the upper atmosphere which became more numerous as 

 the vertical height increased and thereby decreased the effectixe 

 dielectric constant which is a measure of the velocity of propagation 

 of the wave. In this case the velocities at higher levels will be sliglith- 

 greater than the velocities at lower levels, which will result in a bend- 

 ing downward of the wave normal and a consequent curvature of the 

 vva\e path to conform to the curvature of the earth. In order to 

 produce this effect without absorption the ions must be relatively 

 free. If they suffer many collisions during the period of a wave, 

 energy will be absorbed from the wave and pass into the thermal 

 agitation of the molecules. Thus absorption of the wa\'e can be 

 computed provided the nature of the mechanism is luuierstood 

 thoroughh'. 



Sommcrfeld and others luivc worked nut the ctTcct of the iiniK'rfecl 

 conductivity of the grouml uijon the wa\e front and such computa- 

 tions lead to a prediction that the electric vector in the wa\'e near 

 the ground will be tilted forward and thus have a horizontal com- 

 ]ionent. This effect of imperfect conductivity is usualh- given as 

 the cause of the large electromotive force which is induced in the 

 so-called "wave antenna." This effect, however, apparenth' does 

 not lead to an e.vplanation of the bending of waves around the earth. 



There has recently appeared an article by Larmor - in which the 

 idea of a density gradient of ions or electrons is developed further to 

 explain the bending of waves around the earth without a large absorp- 

 tion. This paper, as well as that of Hccles, leads to the conclusion 

 that long radio waves will be bent around the earth, and that the 

 effect increases as the scjuare of the wave length, becoming vanishingK- 

 small for very short waves. 



The large amount of data now available from both qualitative 

 and quantitative observations of radio transmission shows that the 

 |)hcnomena may be more complicated than would be indicated !)>• 

 these theories. It is found that very long waves possess a considerable 

 degree of stability and freedom from fading and that as the wave 

 length decreases the attenuation antl the magnitude of fluctuations 

 increases until for a wave length of (he orcjcr of two or three hundred 



' IVoc. Roy. Soc., June, 1912. 

 > Phil. Mag., Dec., 1924. 



