Three experiments tested whether motion information for nonequiluminant (luminant) and equiluminant dots affects direction judgments when both types of stimuli are moving simultaneously in the same display. The motion directions for the two sets of dots were manipulated to produce four direction differences (0 degrees, 30 degrees, 60 degrees, and 90 degrees). The equiluminant dots were moved in a perfectly correlated fashion, but the percentage of correlated motion for the luminant dots was varied. When subjects judged whether the directions of the equiluminant and luminant dots were the same or different, performance for the conditions with 0 degrees, 60 degrees, and 90 degrees difference improved as the percentage of correlated luminant motion increased. The same result occurred for a control display that contained two sets of luminant dots. However, for the 30 degrees difference, performance was at chance level for the control display, but dropped below chance for the equiluminant-luminant display. When subjects indicated just the direction of the luminant dots, judgments were not affected by equiluminant motion. Judgments for the equiluminant dots also were accurate, except for the conditions with 30 degrees difference; these responses were biased by the luminant motion, indicating some form of motion capture. The interactive effects are discussed in terms of a directionally selective mechanism that combines equiluminant and luminant motion signals.