60 THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1952 



effect is that the data obtained by the observer are generally displaced 

 outward along the time axis by about 0.8 second. 



5. The message rate data were for individual, PBX and two-party 

 subscribers and the flat rate data were for individual and PBX sub- 

 scribers. Furthermore, certain of the flat rate subscribers had auxihary 

 message rate service. It seems likely that different characteristics would 

 be obtained for the individual, PBX and two-party subscribers since 

 there appear to be reasons for expecting significant differences in their 

 dialing habits. The PBX operator is in a position to "shop" for tele- 

 phone service. If she fails to get dial tone on one outgoing line, she can 

 try any other free hue. This can also be done by subscribers with multi- 

 line service. This "shopping" for service tends to produce a large volume 

 of disconnections when dial tone is slow. The individual and the two- 

 party subscribers cannot do this and hence they can be expected to 

 show fewer disconnections. 



6. The results are in terms of intervals of time from the instant the 

 sender is seized. It would, of course, be preferable to have these results 

 in terms of time from receiver off hook. Since on the average the sender 

 is seized in a time interval of about the same magnitude as that of the 

 reaction time of the observer, Figs. 13 to 15 can be read approximately 

 correctly when the abscissas are redesignated "time in seconds from 

 receiver off hook." The foregoing results have been presented to furnish 

 an increased understanding of subscriber dialing habits. In the next 

 section additional results based on individual line records taken on the 

 tapes are presented. 



SUBSCRIBER DIALING HABITS OBSERVED BY INDIVIDUAL LINE RECORDS 



As indicated in the previous section, the results obtained by means 

 of the sender monitor tests were subject to certain shortcomings, hence 

 data taken on individual lines with the 100-pen recorder have been 

 analyzed to augment the information concerning the dialing habits of 

 subscribers. 



As noted heretofore, several of the pens on the 100-pen recorder were 

 available for taking observations on subscribers lines. Two pens were 

 used per subscriber line, one recorded the operation of the subscriber's 

 hue relay while the other pen marked whenever the subscriber's line 

 was busy. On an originating call both pens started marking when the 

 subscriber initiated a call. When a line finder was obtained, the line 

 relay pen ceased marking and it was presumed that the subscriber 

 obtained dial tone at that instant. Dialing, hang-up and flashing by a 



