Three experiments were performed to test Ullman's [ The Interpretation of Visual Motion. MIT Press, Cambridge, Mass. (1979)] independence hypothesis, used in the minimal mapping theory of motion correspondence. Subjects were required to detect the direction of motion (left vs right) of an element in a motion competition display. In control conditions, threshold interelement distances were obtained for this task in the absence of any moving context. In experimental conditions, context elements that moved either to the left or to the right were added to the display. This resulted in changes in thresholds (relative to the control condition) that indicated that a context moving in one direction increased the probability of seeing the competition element move in the same direction. The magnitude of the context effect was shown to be related to the proximity of the context to the competition display, as well as to the number of elements in the context. These results are in conflict with Ullman's independence hypothesis. An alternative model of the motion correspondence process, which uses information about interdependencies between element movements, is briefly discussed.