A review of the clinical efficacy of four structurally distinct antidepressant drugs is presented. Their antidepressant activity can be rationalised within current pharmacological hypotheses of drug action, despite markedly different effects on "in vitro" testing. Fluoxetine, a specific serotonin re-uptake inhibitor, has proven safe, effective treatment for depressive illness and may have a role to play in the treatment of obsessive-compulsive disorder and panic attacks. While it has few of the anticholinergic side effects of the tricyclic antidepressants, nausea, tremor, headache, weight loss, nervousness and sweating are side effects most frequently reported. Minaprine, a compound with weak MAO inhibiting properties and effects on serotonergic receptors, has clinical efficacy in the treatment of depression based on several comparative studies. It is claimed that minaprine lacks anticholinergic and sedative properties. Moclobemide, a specific, reversible inhibitor of MAO-A, has been extensively evaluated in depressive illness. The major advantage of this agent over other irreversible, non-specific MAO inhibitors, is the significant attenuation of the so-called "cheese effect" with doses of tyramine likely to be encountered in foodstuffs. Rolipram, a phosphodiesterase inhibitor, represents a new approach to antidepressant treatment. Limited clinical data suggest that the drug may be an effective antidepressant with few side effects. The place of these agents in therapy is yet to be established.