In two experiments, the names of six colours were individually and repeatedly presented on a screen in random order, for a five-sec, exposure period, each time followed immediately by the auditory presentation of a different word. Two of the colour names (RED and BLUE) were always paired with UCS words having evaluative meaning. For one group of Ss, BED was paired with words of positive evaluative meaning (for example, beauty, happy, sweet) and BLUE was paired with words of negative evaluative meaning (for example, ugly, bitter, criminal). For another group this was reversed, that is, RED was paired with the negative evaluative words and BLUE with die positive evaluative words. Following the conditioning procedure, for both groups, the evaluative meaning of each of the six colours themselves (not the colour names) was measured on a semantic differential scale. When the data of all Ss were considered, both experiments suggested that the colours had significantly acquired the evaluative meanings of the auditorily presented words that had previously been paired only with the colour names. That is, the verbal conditioning of meaning apparently generalized to the referents of the verbal symbols. The second experiment showed, however, that the significant over-all effects were largely due to Ss who indicated an awareness of the relations among stimuli and the purpose of the experiment.