After an initial four week baseline period, during which headache activity and medication consumption were monitored, 28 migraineurs were randomly assigned to one of the following groups: (a) the biofeedback temperature warming group, (b) the visual imagery group, (c) the combined treatment group, or (d) the comparison group. All four groups continued to monitor their headache activity and medication consumption during the eight week treatment period and the eight week follow-up period. A two way analysis of variance computed on groups over time indicated a significant decrease in headache activity and medication consumption. During the follow-up period (a) the combined treatment group had significantly fewer headaches than the biofeedback group or the comparison group and (b) the visual imagery group and the combined treatment group had significantly fewer headache hours than the biofeedback group or the comparison group. These results do not appear to be attributable to differences between groups on the amount of time spent in home practice or subjective ratings of relaxation. There was no consistent relationship between increases in finger temperature and headache activity improvement. Decreases in powerful other scores, as measured by the Health Attribution Test, and increases in subjective ratings of internal control were consistent with a reduction in headache activity and medication consumption.