IN-BAND SINGLE-FKEQUEXCY SIGNALING 



1317 



and 2,000 cycles. The signal-guard characteristics of all of these re- 

 ceivers are shown in Fig. 4. Other parameters used are given in Table III. 



The results of these tests are shown in Fig. 5, where frequency is 

 plotted against signal imitations per 100 calls. The receivers were located 

 at New York and were coiniectcd at ditterent times in trunks to Boston, 

 Toronto, Buffalo, Pittsburgh, Washington and Miami, so as to get some 

 geographical speech distribution. There was' no' detectable geographic 

 effect. 



The tj^pe of speech sound causing the signal imitation is also of in- 

 terest even though we have not as yet been able to put this information 



0.8 

 0.6 



0.08 

 0.06 



0.02 



800 1200 1600 2000 2400 



FREQUENCY IN CYCLES PER SECOND 



Fig. 5 — Results of signal imitation tests. 



