Setting up a Connection. As soon as the key-set lamp lights, in response to such an incoming call, the operator presses a listening button, receives the number from the subscriber, and depresses the corresponding number buttons on that key set, thereby determining the numbers in each of the series of impulses to be sent to the selector and the connector switches to make the desired connection. The operator repeats this number to the calling subscriber as she sets it up, and then presses the starting button, whereupon her work is done so far as that call is concerned. If, upon repeating the call to the subscriber, the operator finds that she is in error, she may change the number set up at any time before she has pressed the starting button.
Building up a Connection. The keys so set up determine the number of impulses that will be transmitted by the impulse-sending machine to the selector and the connector switches. These switches, impelled by these impulses, establish the connection if the line called for is not already connected to. If a party-line station is called for, the proper station on it will be selectively rung as determined by the party-line key depressed by the operator. If the line is found busy, the connector switch refuses to make the connection and places a busy-back signal on the calling line.
Speed in Handling Calls. This necessarily brief outline gives an idea only of the more striking features of the automanual system. A study of the rapidity with which calls may be handled in actual practice shows remarkable results as compared with manual methods of operating. The operators set up the number keys corresponding to a called number with the same rapidity that the keys of a typewriter are pressed in spelling a word. In fact, even greater speed is possible, since it is noticed that the operators frequently will depress all of the keys of a number at once, as by a single striking movement of the fingers. The rapidity with which this is done defies accurate timing by a stop watch in the hands of an expert. It is practically true, therefore, that the time consumed by the operator in handling any one call is that which is taken in getting the number from the subscriber and in repeating it back to him.
TABLE XI
Total Time Consumed by Operator in Handling Calls on Automanual System
| First 100 Calls | ||
| Longest Individual Period | 12.40 | seconds |
| Average five longest Individual Periods | 7.44 | seconds |
| Average ten longest Individual Periods | 6.34 | seconds |
| Shortest Individual Period | 1.60 | seconds |
| Average five shortest Individual Periods | 1.92 | seconds |
| Average ten shortest Individual Periods | 1.96 | seconds |
| Average Entire 100 Calls | 3.396 | seconds |
| Hourly Rate at which calls were being handled | 1060 | |
| Second 100 Calls | ||
| Longest Individual Period | 7.60 | seconds |
| Average five longest Individual Periods | 5.52 | seconds |
| Average ten longest Individual Periods | 5.34 | seconds |
| Shortest Individual Period | 2.00 | seconds |
| Average five shortest Individual Periods | 2.04 | seconds |
| Average ten shortest Individual Periods | 2.18 | seconds |
| Average Entire 100 Calls | 3.374 | seconds |
| Hourly Rate at which calls were being handled | 1067 | |
| Third 100 Calls | ||
| Longest Individual Period | 5.40 | seconds |
| Average five longest Individual Periods | 5.32 | seconds |
| Average ten longest Individual Periods | 4.44 | seconds |
| Shortest Individual Period | 1.60 | seconds |
| Average five shortest Individual Periods | 1.65 | seconds |
| Average ten shortest Individual Periods | 1.80 | seconds |
| Average Entire 100 Calls | 3.160 | seconds |
| Hourly Rate at which calls were being handled | 1139 | |
Owing to the difficulty of securing accurate traffic data by means of a stop watch, an automatic, electrical timing device, capable of registering seconds and hundredths of a second, has been used in studying the performance of this system in regular operation at Ashtabula Harbor. The operators were not informed that the records were being taken, and the data tabulated represents the work of two operators in handling regular subscribers' calls. The figures in Table XI are given by C. H. North as representing the total time consumed by the operator from the time her line lamp was lighted until her work in connection with the call was finished, and it included, therefore, the pressing of the listening button, the receiving of the number from the subscriber, repeating it back to him, setting up the connection on the keys, and pressing the starting key.
It will be seen that the average time for each 100 calls is quite uniform and is slightly over three seconds. The considerable variation in the individual calls, ranging from a maximum of 12.40 seconds down to a minimum of 1.60 seconds, is due almost entirely to the difference between the subscribers in the speed with which they can give their numbers. These figures indicate that, in each of the tests, calls were being handled at the rate of more than one thousand per hour by each operator.
The test of the subscriber's waiting time, i. e., the time that he waited for the operator to answer, for one hundred calls made without the knowledge of the operator, showed the results as given in Table XII, in which a split second stop watch was used in making the observations.