660 BELL SYSTEM TECHNICAL JOURNAL 



of the sphere, with a consequent saving of half of the radiated energy. 

 If instead of saving this energy we added it to the shaded side, the 

 energy available at B would be doubled. This is a simple explanation 

 of the effect of directivity in the transmitting case. The receiving case 

 is quite similar. 



If the transmitter is at B, the energy available at A is diminished by 

 reradiation losses. If we avoid reradiation through the unshaded half 

 of the sphere, the radiation equivalent resistance is halved and the load 

 energy will be doubled, after rematching the load to the antenna im- 

 pedance. 



With this knowledge of the usefulness of sharpened directivity, the 

 designer is tempted to carry it to an extreme. The degree of directivity 

 that may be beneficially attempted is, of course, limited by the varia- 

 tion in the apparent direction of wave arrival. For transatlantic, 16- 

 meter signals over a daylight path, the horizontal plane angular varia- 

 tion, at New York has been * measured, by observing phase differences 

 between spaced antennas, to be some 5 degrees or less, but apparently 

 random throughout this range. Over a combination path of darkness 

 and daylight, a horizontal angle variation considerably greater than 

 this magnitude is frequently observed. 



In the vertical plane, the variations in the apparent directions of 

 arrival are considerable and also random. On rare occasions, angles as 

 high as sixty degrees from the horizontal have been recorded. A sharp 

 low angle antenna may well be expected to decrease in output as the 

 angle of the wave direction becomes high. 



Knowing that the interpretation of wave directions, by means of 

 observed phase differences between spaced antennas, might be compli- 

 cated if multiple waves of varying angles were present, two vertically 

 polarized test antennas were built having optimum response at 27 de- 

 grees and at 6 degrees from the horizontal, respectively as show^n in Fig. 

 3-A. These angles were experimentally obtained from airplane meas- 

 urements. Fig. 3-B, which has been smoothed out for publication, is 

 characteristic of about 80 per cent of the comparative data obtained 

 on these two antennas, as measured by automatic signal recorders. 

 Examination will show that, very frequently, the high angle antenna 

 increases in output as the low angle antenna loses, or vice versa, indicat- 

 ing that the waves are varying in their vertical angle. Similar methods 

 have also cross-checked the horizontal plane movements previously 

 mentioned. 



Where it is planned to design a single fixed antenna for a particular 



* H. T. Friis, "Direction of propagation and fading of short waves," Proc. I. R. E., 

 May, 1928. 



