Estimation of duration of an array of time interval and of its separate elements

It is shown that when a human subject estimates verbally the duration of an array consisting of three intervals - two light intervals divided by a dark one - considerable underestimation is observed compared with estimates of the same duration in the control experiment. In case of separate estimation of the three intervals composing the array the sum of the estimates is much more precise even compared with estimates in the control. Perhaps in this case the system for time intervals estimation operates with two "internal clocks", each of which reads the duration of the dark and of the light intervals. In estimating the array of three intervals the subjects estimate the total array and they do not summate their separate estimations. It may be assumed that the system uses only one "clock" for measuring the duration of the total array.