282 THE BELL SYSTEM TECHNICAL JOURNAL, MARCH, 1954 



taking the sum of the loads carried bj^ a stated number of trunks and 

 subtracting it from the total offered load, the amount of load "lost" or 

 overflowed can be obtained. It should be noted here that the load carried 

 by a given number of trunks under the pressure of a constant value of 

 offered load, and provided lost calls are cleared, will always be the same 

 whether or not the load is presented to said number of trunks in consecu- 

 tive order. Thus, seven trunks to which 240 CC8 are presented \\ ill carry 

 185 CCS (Fig. 3, Curve B) whereas six trunks will carry only 165 CCS, all 

 values to the nearest whole CCS. Therefore, it may be said that the 

 "last" trunk in a group of seven carries 20 CCS under the Erlang B 

 concept when the offered load is 240 CCS, since by adding one trunk to 

 a group of six, the capacity of the group to carry load is increased by 

 20 CCS. The significance of this notation will appear presently. 



Let us leave the subject of direct high usage trunk group loading for a 

 moment and consider the loading of final trunk groups such as AT and 

 TN, in Fig. 2. We said earlier that an office such as A originates traffic 

 of such small volume to each of many other offices in the city that direct 

 routing would be economically unsound. Such items in the aggregate 

 constitute a load of considerable size and all are routed via a tandem 

 office for distribution to their respective destinations together with simi- 

 lar small parcels originating at B, C, etc. As stated earlier trunks to and 

 from tandem were customarily provided on a P. 01 delay basis and due 

 to the size of the aggregate loads mentioned above such groups varied 

 in size from about 20 trunks upward. Any traflfic overflowing from direct 

 trunks to an alternate route via tandem would, therefore, require the 

 addition of trunks to these rather large groups already established for 

 the handling of traffic for Avhich the tandem route was the first and only 

 route. It now becomes pertinent to inquire into the capacity of trunks 

 added to the tandem groups to carry rerouted load arriving from the 

 direct high usage groups. From the P.Ol trunk capacity table, it can be 

 shown that adding a trunk to a group of 20 trunks increases the capacity 

 of the group by 27 CCS, to the nearest whole CCS. Similarly adding one 

 trunk to a 40 trunk group increases the capacity of the group by 29 

 CCS. It was decided in the New York City trials to use a constant 

 average value of 28 CCS as the capacity of any trunks added to groups 

 in an alternate route via tandem in order to accommodate rerouted 

 traffic and still keep the groups on a P.Ol basis for all traffic. 



We are now in a position to compare the efficiency of a triuik added 

 to a final route (two legs constitute one route) with the efficiency of a 

 trunk added to the direct high usage group. The efficiency of the former 

 is, for practical purposes, a constant, whereas the efficiency of the latter 



