Placebo-controlled trials are used widely in the development of new pharmacological treatments. They have sometimes been challenged as being unethical, in clinical situations where patients can receive an existing effective and acceptable treatment. It has been argued that studies of potential antidepressants should employ only a comparator-controlled design, whereby new compounds have to be found at least as efficacious as existing treatments. By contrast, others have argued that sole use of comparator-controlled trials is itself unethical, as more patients will be exposed to potentially unhelpful treatments. This article reviews the rationale for conducting placebo-controlled treatment studies in depressed patients, examines the underlying ethical issues, and describes the provisions that should be applied when investigating the efficacy and tolerability of potentially valuable new antidepressant treatments. Placebo-controlled studies of new antidepressants are justified, both scientifically and ethically; restrictions on placebo-controlled investigations will hinder the arrival of more efficacious and better tolerated antidepressants.