Bodily gestures play an important role in the communication of expressive intentions between humans. Music ensemble performance, as an outstanding example of nonverbal human communication, offers an exemplary context to study and understand the gestural control and communication of these expressive intentions. An important mechanism in music ensemble performance is the anticipation and control of interpersonal timing. When performing, musicians are involved in a complex system of mutual adaptation which is not completely understood so far. In this study, we investigated the role of performers' gestures in the mediation process of interpersonal timing in a dyad performance. Therefore, we designed an experiment in which we controlled for the use of hand and arm movements in a musical task, in which dyads were asked to synchronously tap out a melody. Next to their comfortable/natural way of tapping, we instructed participants to either perform pronounced expressive hand and arm gestures in between successive taps, or to restrict from any overt body movement. In addition, we looked at effects of visual contact (yes/no) and tempo (slow: 50 beats per minute; fast: 100 beats per minute). The results show that performers' gestures improve interpersonal musical timing, in terms of the consistency and accuracy of onset asynchronies, and of the variability of produced inter-onset intervals. Interestingly, we found that the use of expressive gestures, in regard to comfortable/natural movements, add to these positive timing effects, but only when there is visual contact and at the slow tempo. In addition, we found that the type of gestures employed by musicians may modulate leader-follower dynamics. Together, these findings are explained by human anticipation mechanisms facilitated by gesturing, shedding new light on the principles underlying human communication of expressive intentions, through music.