CERT.'ilN FACTORS AFFECTING TELEGRAl'll SPEED 327 



When a telcj;raph circuit is worked at a line speed as higli as will 

 be permitted by the availal)le fretiuencN- range, the shape of the 

 received signal will be practically independent of the shape of the 

 transmitted signal, and further, the magnitude of the received signal 

 will be approximately directly proportional to the area included 

 within the impressed voltage wave. 



The area included within the impressed voltage wave being of 

 principal importance so far as the wave received at the distant end 

 is concerned, the areas under the three \oltage waves shown in Fig. 1 

 will next be examined. The areas under waves A and C will be found 

 to be substantially equal while the area under the wave B is only 

 alxjut ().() as great. Consequently, it should be expected that waves 

 A and C will be about equally good from the standpoint of the received 

 signals, while wa\e B will be poorer, producing received signals only 

 about 0.6 as great in magnitude. If the maximum voltage (or power) 

 impressed at the sending end is limited to some given value, the rec- 

 tangular wave is seen to be the optimum, siTce this wave has the 

 maximum area. While the area shown under curve C is approxi- 

 mately equal to that under the rectangular wave, the efTect produced 

 when a number of signal elements of the same polarity and magni- 

 tude are sent in succession is such that the maximum voltage trans- 

 mitted will exceed slightly the corresponding voltage for the case 

 of the unmodified rectangular wave due to overlapping of adjacent 

 signal elements. 



The above comparison of the three waves of Fig. 1 from the stand- 

 point of received signals holds not only for signal elements, but also 

 for complex waves comprising a number of elements. Since for the 

 speeds under consideration the received currents for different shapes 

 of signals applied at the sending end are substantially of the same 

 form, differing, at most, in magnitude, it follows from the principle of 

 >uperposition that any complex signal, whether built up of elements 

 of one shape or another at the sending end, will produce substantially 

 the same wa%e form at the receiving end, the differences in the shapes 

 of the elements at the sending end producing differences principally 

 in magnitude of the received waves. 



Consideration will next be given to the relative interference which 

 the different wave forms of Fig. 1 will produce in the frequency range 

 assigned to other circuits. Since interference into other circuits 

 results from having the telegraph signal elements contain frequencies 

 which spread into the ranges assigned to other circuits, it is evident 

 that the wave will be the best from the standpoint of interference 

 which contains the least amount of these outside frequencies. By 



