After evaluating 165 insomniacs, 48 psychophysiologic insomniacs were randomly assigned to one of the following four groups: electromyographic (EMG) feedback, combined EMG and theta feedback, sensorimotor rhythm (SMR) feedback, and no treatment (control). Sleep evaluations by home logs and in the laboratory were done before and after biofeedback and nine months later. No feedback group showed improved sleep significantly more than did the controls. The amount of feedback learning correlated significantly with sleep improvement for the SMR group but not for the other groups. Initial tension of the insomniacs correlated positively with sleep improvement for the EMG group, but negatively with sleep improvement for the SMR group. Those treated with the biofeedback that seemed appropriate for their specific deficiencies showed significant sleep improvements, while those who received inappropriate feedback did not. Appropriate biofeedback methods may be effective for specific types of insomnia, but these procedures offer no panacea for all poor sleep.