The paper proposes a model for the timing of repetitive discrete motor responses, specifically focusing on the relationship between interresponse intervals and the timing of key tapping tasks. The authors use a two-process model to account for the observed variance in interresponse intervals, where one process generates trigger pulses (timekeeping) and the other introduces delays (response delays). The model predicts a negative dependency between successive interresponse intervals, which is confirmed by data from a Morse key tapping task. The method involves analyzing the first-order serial correlation between interresponse intervals to distinguish between variance due to the timekeeping process and variance in motor response delays. The results show that the lag one serial correlation is within the range of 0 to -1/2, supporting the two-process model. The authors also examine how the timekeeper's variability changes with the mean interresponse interval, finding that the intercept of the linear relation between timekeeper variance and mean interresponse interval is positive. This suggests that the timekeeper's variability increases with longer interresponse intervals. The paper concludes by discussing the implications of these findings for understanding the mechanisms underlying response timing tasks.The paper proposes a model for the timing of repetitive discrete motor responses, specifically focusing on the relationship between interresponse intervals and the timing of key tapping tasks. The authors use a two-process model to account for the observed variance in interresponse intervals, where one process generates trigger pulses (timekeeping) and the other introduces delays (response delays). The model predicts a negative dependency between successive interresponse intervals, which is confirmed by data from a Morse key tapping task. The method involves analyzing the first-order serial correlation between interresponse intervals to distinguish between variance due to the timekeeping process and variance in motor response delays. The results show that the lag one serial correlation is within the range of 0 to -1/2, supporting the two-process model. The authors also examine how the timekeeper's variability changes with the mean interresponse interval, finding that the intercept of the linear relation between timekeeper variance and mean interresponse interval is positive. This suggests that the timekeeper's variability increases with longer interresponse intervals. The paper concludes by discussing the implications of these findings for understanding the mechanisms underlying response timing tasks.