Twenty-four patients with Raynaud's phenomenon and scleroderma were randomly assigned to receive finger temperature biofeedback, frontalis EMG biofeedback, or autogenic training. Only those receiving temperature feedback showed significant increases in finger temperature during training and during a posttraining test of voluntary control, effects not attributable to general relaxation. However, no group demonstrated significant clinical improvement, assessed by symptom reports and by ambulatory monitoring of finger temperature. The need for careful classification of patients with Raynaud's disease and Raynaud's phenomenon in scleroderma is emphasized.