This experiment used a priming paradigm to examine the specificity of mental color codes. On each trial subjects judged whether two color chips were physically identical. On primed trials under the chip condition the prototype color chip was presented for 2 sec prior to the test pair. Under the visual and auditory name conditions the category name was presented prior to the test pair. For same judgments, the chip prime facilitated responses to good members of the color category but inhibited responses to poor members. Under the same conditions the category prime facilitated responses at all goodness levels. It was argued that, under the present conditions, information contained in the mental representation generated to the category name was less specific than that contained in the physical code produced when the prototype chip was perceived.