Factors that reliably predict treatment response in bipolar disorder are much needed, particularly since no available treatment routinely affords complete protection from future illness. The number of pretreatment episodes (PTEs) is a proposed predictor, but its value remains uncertain. We therefore reviewed available research on this topic. Based on a computer-assisted search of the literature, we identified 28 reports providing data on response to lithium treatment and on history of past illness. We evaluated their methods and findings to test the hypothesis that greater PTE count predicts inferior clinical response to lithium. Most studies (68%) found no support for the predicted relationship, and those that did or did not find the hypothesized relationship differed nonsignificantly in ratings of overall study quality and individual factors, including study size, previous lithium use, diagnostic criteria, and outcome measures. The concept that PTE count strongly and consistently predicts inferior clinical response to lithium treatment in manic-depressive disorders is not supported by available research, thus adding to emerging uncertainties about the relationship of past history and later course of these often severe, disabling, and life-threatening illnesses.