It has long been lore in the electroconvulsive therapy (ECT) field that melancholic depression is particularly responsive to ECT, whereas nonmelancholic depressions are less so. Early large case series in the 1940s and 1950s, while using now outdated diagnostic nosologies, did provide some hints to the particular efficacy of what later on would be considered melancholic or endogenous depression. The first systematic studies of this subject in the 1950s and 1960s seemed to confirm the superior response of endogenous depression to ECT. However, with the advent of the Diagnostic and Statistical Manual of Mental Disorders, Third Edition, and the first criterion set specifically labeled "melancholia," most modern series of studies have failed to find an association of melancholic symptoms with ECT response. Several problems both with the definition of melancholia and the modern methods of diagnosis and outcome assessment are hypothesized to be at the root of the negative findings of modern studies. Suggestions for future ECT/melancholia studies are offered.