## Info

needs to function. The fact that the peak shifts are frequently (but not always) more than a stop result and less than a reset result is dismissed by these theorists as the result of a probabilistic system of stopping and resetting the clock. If the clock is automatically stopped during a gap and has a small chance of resetting completely each moment during the gap, one would expect a pattern of data similar to that described above. During short gaps, there would only be on average a few trials where the clock reset completely, so the average peak shift on those trials would be approximately a stop result plus a bit, due to a few resets. On long gaps, most gap trials would result in reset, but there might be a few stops in there to lower the average.

The competing hypothesis is that the subject's memory of the duration before the gap decays during the gap. If the stored duration is reduced by 5% during each second of the gap, only a small portion of the stored time would be lost during a

0 0