When two identical targets move toward one another from opposite sides of a display and continue past one another along collinear trajectories, they can be perceived to either stream past or bounce off of one another. Streaming is the dominant perception in motion displays free of additional transients, while bouncing predominates when a transient (eg auditory or visual) is presented at the point of coincidence. We investigated whether the auditory induced bias towards bouncing would persist as the probability of a motion reversal was reduced by introducing a spatial offset either vertically in a 2-D display or in depth in a 3-D display. Offset conditions were combined with two auditory conditions (tone or no-tone at the point of coincidence) in the presence or absence of a central occluder. In conditions with no sound, streaming was reported on a clear majority of trials, regardless of spatial offset. When a transient tone was presented, reported motion reversals dominated and persisted for increasing verbal offsets up to 17.9 min of arc and for 3-D trajectory offsets up to 25.6 min of arc. The bounce-promoting effect of an auditory tone at the point of coincidence in stream/bounce displays persists in spite of rendering the visual motion sequence unambiguous and more consistent with streaming.