We examine how various sources of information contribute to the percept of motion in depth. Subjects were presented with targets moving in depth, and were asked to judge their velocities and final positions. On each presentation, the target's position relative to the two eyes (target vergence), the size of the target's retinal image and the difference in this image's position relative to that of the background in the two eyes (relative disparity), each either changed as they normally would for a target moving at a fixed speed towards the observer, or did not change at all. Subjects' judgements for various such combinations show that all three sources of information influence both the perceived velocity of motion in depth and the final perceived position, but in different ways. This is not too surprising, because the assumptions that the use of each source of information are based on, are different for the two tasks. We propose that the way the different sources are combined is governed by the likelihood of the assumptions, that are required to use that information, being true under the given circumstances.