150 BELL SYSTEM TECHNICAL JOURNAL 



Discussion of Certain Trends wiih Respect to Frequency 



It may be helpful, in recapitulating, to consider the different trends 

 which ultra-short wave transmission shows with respect to frequency, 

 and to mention their relationships to the phenomena of the ground 

 wave at lower frequencies. 



The simplest trend is that to be found in free space, that is, in cases 

 for which the effect of the ground is negligible. Changes, if any, in 

 transmission efficiency with frequency are then due to the air itself. 

 Such evidence as there is on this point indicates that the assumption 

 of absorption by the air is unnecessary within the range of our experi- 

 ments, and in fact this is to be expected on theoretical grounds. The 

 "free space" trend therefore gives merely a horizontal line. In Figs. 

 13 and 14 the high frequency portion of the curve oscillates about this 

 line and would approach it if reflections from the earth were decreased 

 in strength. 



In general, however, the effect of the earth will alter this trend in 

 such a way as to give a variation with respect to frequency. Perhaps 

 the most familiar variation is the loss of efficiency in going to high 

 frequencies when vertical polarization is used. The decrease is due to 

 conduction losses in the ground. This "conductivity trend" appears, 

 for example, in the work of Sommerfeld and of Zenneck. Many 

 experimental observations have been made of it at broadcast fre- 

 quencies for distances up to a few hundred miles over paths which are 

 obviously "non-optical." We have seen it here in the optical path 

 tests made over sea-water. Fig. 3. It appears also in the low frequency 

 end of the "vertical polarization over water" curve in Fig. 14. 



In several cases we have noted that 69 megacycles was more 

 efificiently transmitted over land than 17 megacycles. This is opposite 

 to the conductivity trend and appears to have a very different cause. 

 For both optical and non-optical paths it is believed to be associated 

 with a phase change at reflection of 180°, and the effect is most pro- 

 nounced when reflection occurs without appreciable loss of amplitude. 

 This "negative reflection" trend is exemplified on the one hand by 

 the very poor transmission with very low frequencies when horizontal 

 polarization is used, and, on the other, by the excellent transmission 

 at 75 mc. (4 meters) between Beer's Hill and Lebanon (Fig. 13). In 

 the latter case the difference in path lengths of direct and reflected 

 waves was not negligible compared with a half wave-length, and it 

 is a phase shift due to this cause which apparently prevents destructive 

 interference. This negative reflection trend also appeared in the non- 

 optical paths over level land (Fig. 8). It is affected not only by the 

 negative reflection in the neighborhood of the antennas (negative 



