25cS Bl'lLL SYSTEM TECHNICAL JOURNAL 



the record on three days is caused by the transmitting station shifting 

 to a different frequency in an attempt to improve conditions. As to 

 commercial transmission results over this channel during this period: 

 the first two days were fair, the third day poor, the 27th, the 28th, 

 29th, and 3()th very poor, and July 1 still rather poor. 



The Problem ok the Transmitting Medium 



These adverse effects in short-wave transmission are ascribed to 

 the nature of the medium through which the propagation of the waves 

 takes place. The short-wave signals which reach a distant point are 

 carried by waves which have traveled in the upper regions of the 

 atmosphere, where a condition of ionization exists which causes the 

 waves to move in a curved path and, finally, to arrive again at the 

 earth's surface. The ionization in the upper part of the atmosphere 

 varies with atmospheric conditions and hence its action on the waves 

 which are passing through it varies from day to night, from season to 

 season, in a more or less regular manner, on which are superposed 

 fortuitous variations due to other conditions. The conditions in the 

 upper atmosphere may be such that two or more waves arrive at a 

 distant point from the same source after having traversed different 

 paths. If the length of one of the paths is changing, the resulting 

 signal from the two waves will pass through a series of maxima and 

 minima in time, which process is known as fading. This complicated 

 path condition is present at practically all times, since it is only on 

 very rare occasions that short-wave signals do not fade in and out. 

 Furthermore, there appear to be different kinds of fading corresponding 

 to different transmission paths. For example, the fading on the North 

 Atlantic short-wave circuits is of a deep slow variety as compared 

 with the faster and more choppy type of fading experienced on the 

 north and south circuit between New York and Buenos Aires. 



To some extent this fading can be overcome by means of auto- 

 matic gain control in the radio receiver which causes a steady signal 

 to be delivered to the listener. However, this does not correct for 

 the distortion which may be produced by interference between two 

 transmission components. This distortion may result from a selective 

 fading of the various frequencies in the voice band and an oscillo- 

 gram showing this condition is given in Fig. 10 which is taken from 

 a paper by R. K. Potter.^ These are records of transmission across 

 the North Atlantic of the voice band occupied by 10 suitably spaced 

 tones of equal amplitude at the transmitting end. There is shown in 

 the vertical columns a succession of snapshots which are separated 



■' See bibliography 19. 



