This study was concerned with a comparison of the effects of biofeedback and of imagery on finger temperature increase. Twelve volunteer unpaid adults were randomly allocated to one of two groups: finger temperature biofeedback or guided imagery. All subjects received two baseline and five 15 min treatment sessions, during which stimulus presentation was carefully controlled. The biofeedback group rapidly established a superior ability to increase finger temperature; whereas guided imagery was consistently associated with temperature decrease. Some tentative evidence was produced which questioned the alleged importance of cognitive and interpersonal factors in biofeedback training. Implications for theoretical accounts of the mechanism of finger temperature control (via autogenic training or biofeedback), and for developing effective biofeedback procedures are drawn.