This study investigated the effect of conflicting nonarbitrary (color) relations on equivalence responding. Three groups of 8 subjects were trained and tested for the formation of three 3-member equivalence classes using nonsense syllables as stimuli. Each subject received two separate exposures. For the No-Color group all stimuli in both training and testing phases were in black lettering. For the All-Color group, all stimuli were in color. Thus, training effectively involved learning to ignore color and, as predicted, during testing color had little or no effect upon performance, in that there was no significant difference in levels of equivalence responding between the no-color and all-color groups. For the Color-Test group, training and testing stimuli were in black and color lettering respectively. During testing, the sample was always differently colored from the "equivalent" comparison, but was the same color as one of the nonequivalent comparisons. These subjects had no history of reinforcement for ignoring color, and thus a possible conflict between arbitrary and nonarbitrary relational control was produced in this condition. Results showed that for this third group, levels of equivalence responding were significantly lower than for either of the other two groups. Furthermore, levels of responding in accordance with color matching were significantly higher for the color-test group than for the all-color group. These data are consistent with Relational Frame Theory.