EFFECT OF SIGNAL DISTORTION IN MORSE TELEGRAPHY 281 



value of distortion lower than that with any other type or combination 

 of distortion causes a circuit to be unsatisfactory, according to the 

 opinion of the operators. Moreover, the opinion of the operators dis- 

 agrees considerably with the effect which the distortion has upon their 

 accuracy of reception, as shown by the curve. 



A summary of the effect of negative characteristic distortion as 

 given on the curve of Fig. 5 is as follows. The accuracy of reception 

 remained nearly constant up to about 45 per cent distortion and then 

 commenced to decrease rapidly. The operators remarked on the 

 distortion at about 15 per cent and considered the circuit unsatisfactory 

 at about 30 per cent. 



The nature of the errors made by operators for this type of distortion 

 was analyzed for values of distortion from 15 per cent to 70 per cent. 

 Consistent misinterpretations occurred for distortions greater than 40 

 per cent and were similar to those obtained with positive characteristic 

 distortion. The most common errors were as follows: J interpreted 

 as K or M ; N as T; G as M ; and K as M. In addition, there were the 

 usual miscellaneous interpretations, together with a large number of 

 letters added and a smaller number of letters omitted. 



4. Effect of Fortuitous Distortion 



The curve of Fig. 6 which is an average of three test curves, illustrates 

 the effect of fortuitous distortion upon the accuracy of reception. The 

 distortions plotted are the maximum values which were equalled or 

 exceeded about three or four times a minute. It appears from the 

 curve that large amounts of fortuitous distortion have very little 

 effect upon the accuracy of reception. In fact the accuracy is prac- 

 tically unchanged over the range of distortion from zero to 75 per cent. 

 This is not very surprising in view of the fact that such distortions 

 occur only about three or four times per minute, and may affect signals 

 without destroying their identity. Even with fortuitous effects of such 

 large values as to cause the breaking up of signals, only a relatively 

 small decrease in the accuracy of reception was produced in certain 

 tests. 



The errors made by the operators for this type of distortion were 

 largely of a miscellaneous nature. There were very few misinterpreta- 

 tions which were repeated consistently, although most of the errors 

 indicated a shortening of dashes to dots and a dropping out of dots and 

 spaces. There were also a considerable number of letters omitted and 

 a few letters added. 



Briefly summarized, the efifects of fortuitous distortion upon recep- 

 tion are as follows. The accuracy of reception remained practically 

 19 



