On a repetitive tapping task, the within-hand variability of intertap intervals is reduced when participants tap with two hands as compared to one-hand tapping. Because this bimanual advantage can be attributed to timer variance (Wing-Kristofferson model, 1973a, b), separate timers have been proposed for each hand, whose outputs are then averaged (Helmuth & Ivry, 1996). An alternative notion is that action timing is based on its sensory reafferences (Aschersleben & Prinz, 1995; Prinz, 1990). The bimanual advantage is then due to increased sensory reafference. We studied bimanual tapping with the continuation paradigm. Participants first synchronized their taps with a metronome and then continued without the pacing signal. Experiment 1 replicated the bimanual advantage. Experiment 2 examined the influence of additional sensory reafferences. Results showed a reduction of timer variance for both uni- and bimanual tapping when auditory feedback was added to each tap. Experiment 3 showed that the bimanual advantage decreased when auditory feedback was removed from taps with the left hand. Results indicate that the sensory reafferences of both hands are used and integrated into timing. This is consistent with the assumption that the bimanual advantage is at least partly due to the increase in sensory reafference. A reformulation of the Wing-Kristofferson model is proposed to explain these results, in which the timer provides action goals in terms of sensory reafferences.