The behavioral models of depression proposed by Ferster (1966, 1973), Lewinsohn (1974a, 1974b), and Seligman (1973, 1975, 1978) are summarized and the variables proposed in the models are evaluated. While no one variable has been related conclusively to onset of depression, there is sufficient evidence regarding the effects of these variables to justify further investigation. Most research to date has been directed toward either characterizing depression by correlating various measures of depressive behavior or demonstrating the relation between individual variables and reduction in depression. Correlational research facilitates the specification of the definition of depression. Research on reduction in depression contributes to the development of effective intervention techniques. A theoretical understanding or prediction of depression onset will require research demonstrating antecedents to depression. Problems of such research are discussed and suggestions for future strategies for research are offered.