372 BELL SYSTEM TECHNICAL JOURNAL 



has been said, however, regarding the results of this work, which was carried 

 out entirely on loaded cables, because there was no progress in commercial 

 utility to report. Nevertheless progress had been made, but of a somewhat 

 negative character. 



Much had been learned regarding the inherent limitations of the repeater . 

 element itself. 



The loaded circuits of that era which were sufficiently regular for good 

 transmission without repeaters were being found not to be sufficiently 

 regular when used with repeaters. Moreover, similar types of loaded cir- 

 cuits in the same cable were found to have large differences in their impe- 

 dance characteristics. 



In all of the experiments the 21-type repeater circuit was used. In this 

 circuit, the line sections between which the repeater works must be closely 

 equal in impedance, in order to avoid transmission distortion by interaction 

 of the input and output currents, which may be very disastrous to intelli- 

 gibility, especially when attempts are made to operate two or more repeaters 

 in tandem in the same circuit. 



It thus happened that the limitations of the mechanical repeater element, 

 the repeater circuit, and the lines, combined to accentuate each other's 

 effect in piling up the practical difficulties. 



A Plan Evolves 



The November-December 1910 analyses of what had been done and what 

 remained to be accomplished resulted in an engineering decision to renew the 

 attack on the repeater problems on an "all-out" basis, according to a plan 

 which would be designated as a "four-prong" offensive in the military lan- 

 guage of today. The following statement of this fundamental plan is 

 quoted from the Work Order No. 7655, "General Repeater Study," prepared 

 by Jewett, and which was officially approved by Carty on April 1, 1911. 



"Nature of Work 



"A general study to determine the proper characteristics for the best telephone 

 repeater, its circuit, and the general terminal and line conditions that must be 

 fulfilled to make this repeater available for both loaded and non-loaded lines. 

 This study will include — • 



"(1) A complete study of the characteristics of the existing receiver-transmitter 

 (Shreeve) type of repeater with a view to determining whether the action of this 

 repeater cannot be improved upon and whether modifications in the repeater ele- 

 ment, its circuit or in the line conditions will make it suitable for general use on 

 loaded lines. 



"(2) A study of other possible repeater ideas, particularly in the domain of 

 molecular physics. Certain characteristics of discharge of electricity through gases 

 and vapors seem to offer considerable possibility of obtaining a telephone amplifier 



