DIALING HABITS OF TELEPHONE CUSTOMERS 51 



S. Treatment of Cases Where Two or More Calls Were Found to he 

 Waiting on One Trip Circuit Sub-Group 



The oecurrence of several calls \vaitiiif>; on one trip circuit was occasion- 

 ally noted in the analysis. Referring to Fig. 8, a case is shown on trip 

 circuit sub-group 12. At ^9 a line finder is seized. Trip circuit sub-group 

 18 shows that a subscriber is dialing before tone. The appearance of 

 dial pulses on this trip circuit indicates that only one subscriber is 

 demanding service otherwise the dialing would not show. Trip circuit 

 sub-group 12 however appears to have two or more recjuests for service. 

 One of these requests for service began at ti. The start of the second 

 recjuest occurred somewhere between ^4 and t%, perhaps half-way be- 

 tween. At tg, one of the requests was served by a line finder. To sim- 

 plify the handling of such cases, the assumption was made that the 

 first attempt started at ^4 and ended at ^9 and the second request started 



SUBSCRIBER FLASHED 



J ><. 



f 1 ? 



-LINE RELAY OPERATED '^ LINE RELAY RELEASED DIAL TONE -" 



(SUBSCRIBER TOOK (SUBSCRIBER DEPRESSED RECEIVED 



I RECEIVER OFF HOOK) SWITCHHOOK) I 



|<_ 21.6 SECONDS -*H 



Fig. 9 — Example of a customer flashing for dial tone (Tape made October 18). 



at /g. The effect of this is to understate by an indeterminate amount 

 the average value of H for each class of service. This understatement 

 should be noticeable for the higher degrees of overload because the 

 occurrence of several calls on one trip circuit sub-group is then most 

 likely to occur. 



The effect of several calls simultaneously waiting in a trip circuit 

 sub-group and of one or more calls dropping out is to overstate the 

 magnitude of H. For instance if two simultaneous call attempts of five 

 seconds each overlap for one second and both attempts are abandoned, 

 the apparent average waiting period is nine seconds, whereas it should 

 be five seconds. It is believed that the above three rules tend to create 

 understatements which roughly balance this type of overstatement. 



DISTRIBUTION CURVES OF SIMULTANEOUS CALLS 



The detailed analysis of the tapes provided distributions of simul- 

 taneous calls. For each class of service studied these distributions can 

 be compared with theoretical chstributions derived from the generalized 

 trunking formula using the j factors developed in the analysis. Several 



