516 



BELL SYSTEM TECHNICAL JOURNAL 



lowing order as regards their insensitivity to flanking channel crossfire: 

 linear on-oflf and two-source without limiter about the same, single sideband 

 with no spacing carrier about the same as with — 6 db spacing carrier, fre- 

 quency-shift. 



It may be observed from Tables II and III that the interchannel crossfire 

 obtained for the linear on-off and two-source arrangements was small, due 

 to the location of the carrier frequencies at the centers of the channel filter 

 bands. The crossfire obtained for the single-sideband and frequency-shift 

 methods given in Tables IV to VI was greater, due to the location of the 

 marking or spacing frequencies near the edges of the channel filter bands. 



Table II 



1955-Cycle Linear On-Off Arrangement, Effect of Operating Adjacent 1785-Cycle 



AND 2125-Cycle On-Off Channels over Line 



Table III 



1785 and 1955-Cycle Two-Source Arrangement without Limiter, Effect of 



Operating Adjacent 1615-Cycle and 2125-Cycle Channels over Line 



The other arrangements previously mentioned were not tested for flanking 

 channel crossfire. The two-source arrangement with limiter and normal 

 band width is thought to be no worse from this standpoint than that without 

 limiter, since a limiter usually helps in discriminating against small spurious 

 currents. From theoretical considerations the one-source two-band ar- 

 rangement with limiter is thought to be better than the two-source arrange- 

 ment without limiter, using the same channel filters and carrier frequencies. 

 The narrow-band two-source arrangement had considerably sharper cut-off 

 filters than the normal-band arrangement, thus causing greater attenuation 

 of frequencies outside the desired band. Consequently the narrow-band 

 two-source arrangement is thought to be no worse than the normal-band 

 arrangement as far as interchannel interference is concerned. The wide- 



