6 BF.LL SYSTEM TECHNICAL JOURNAL 



balance between the circuits eniplo\ed for subscribers' loops and 

 those employed in intenjftice trunk lines. The larger the wire, the 

 belter will be the talk. But it will also be more expensi\c. The 

 first step in solving this i^roblem is to decide how good the trans- 

 mission must be to afford satisfactory service to the telephone using 

 public. Our present standartls are a matter of growth; the accumu- 

 lated results of long and extensive experience. They are live, working 

 standards constantly being intelligently scrutinized and, when neces- 

 sary, modified. A discussion of the values of the standards employed 

 would uiuluK' prolong this |)aper. Therefore, let it suffice, at this 

 time, to slate that the telephone offices in a large city, including its 

 einirons, ma>' be divided into metropolitan offices and suburban 

 offices; that is, the central business offices separated from the subur- 

 ban residential offices. Between subscribers in difTen-nt districts 

 suitable standards of transmission are decided iipdii. 



Before descril)ing this stud\' further, reference must be made to the 

 practical necessity for the standardization of construction materials. 

 Subscribers' loops run in length from a few hundred feet to 3, 4 or r> 

 miles. If we tried theoretically to make all talks exactly equal in 

 loudness, we should have as many different sizes of wire in our cables 

 as there are different lengths of loop. To reduce the complexit>, our 

 cable conductors are of certain standard sizes, which experience has 

 shown are sufficiently close together to meet the needs of the busi- 

 ness. These standard sizes, in American Wire Gauge, are Nos. '24, 

 22, 19, 16, 13 and 10; the three latter not being used in subscribers' 



lonps. 



Having adopted st.mdards of transmission and standards of cabU' 

 conductor sizes, our problem is to obtain the standards of transmis- 

 sion with the standards of cable contluctors in the most economical 

 nianiuT. 



The nn'ihdd of doing this, in brief, is to tigiiie out the annual costs 

 \\lii( h would lie iiu lured in doing it a number of different ways and 

 to select the wa\ that gi\es the lowest annual cost. In this kind of 

 a stud\', which we call a "loop anil trunk" study, it has been con- 

 venient to designate the subscribers' loops b\- their maximum circuit 

 resisUuice. .Adojiting this form of designation, it may be assunuil. 

 first, ih.it all of the subscribers' loops will ha\e an average transmit - 

 tini; ami receiving efficiency as good or better than a 350-ohm looj); 

 as a second assumption, that they will be as good or better than a 

 H)()-ohm loop; and, as third and fourth assumptions, 450 and oOO-olim 

 loo[)s, respecti\eK-. In assuming, for example, a 3.50-ohm loo[) in 



