188 BELL SYSTEM TECHNICAL JOURNAL 



New York and London-New York for the year 1930. Again the 

 mid-range slope as shown by the extended broken lines is about the 

 same. The horizontal separation of roughly 25 db illustrates the 

 much more favorable noise-to-signal conditions on the low latitude 

 circuit, since the transmitter power and antenna gains were substan- 

 tially the same in the two cases. 



In Fig. 4 (b) are shown 6-A noise distribution cur\cs for the Hono- 

 lulu-San Francisco, Buenos Aires-New York and London-New ^'()rk 

 circuits representing conditions during 1932 and 1934. The mid- 

 range slope is remarkably similar for all six curves. A comparison of 

 these curves illustrates the high degree of reliability obtained on the 

 circuit to Honolulu. 



In Figs. 4 (a) and (b) the data for the low latitude paths cover about 

 9 hours of daylight operation as compared with 24-hour full time 

 operation for the transatlantic case. Although data are not available 

 for a 24-hour comparison some available experience indicates that such 

 a comparison would not have altered the separation between the curves 

 shown very appreciably. After all, the interest here is mainly in the 

 slopes of these curves, and there is no reason to believe that the slopes 

 A\()uld be aflfected. 



Field Intensity Distribution 



So far the discussion has shown that over the dependable portion 

 of the 6-A noise distribution curves representative of both different 

 circuits and different years the slopes appear to be nearly the same. 

 If this is the case one curve may be constructed to represent approxi- 

 mately the effect of transmission improvement or degradation upon 

 the performance of any long range short-wave circuit. The useful 

 range of this curve is, however, limited by the dependable range of 

 the 6-A noise measurements. To extend the useful range of the curve 

 in order to estimate the effect of large changes in transmission im- 

 provement it is necessary to resort to a correction for the bend at the 

 high noise end. Correction at the low noise end would concern the 

 less important case of transmission degradation. Although it is pos- 

 sible to apply an approximate correction for the above-mentioned 

 effect of the automatic gain control upon the bend at the high noise 

 end it is probably more accurate to consider the distribution of field 

 intensity data at these times of high noise-to-signal ratio. 



The limiting conditions on short waves are predominantly those 

 accompanying magnetic disturbances when the signal fields drop to 

 \ery low values. The indications are that the atmospheric noise 

 fields also decrease to a less noticeable extent during these disturb- 



