SynopsisThe development of knowledge about the mechanism of action of tricyclic and the so-called ‘atypical’ antidepressants (AD) is reviewed. The discovery of clinically active antidepressants with little or no effect on noradrenaline or serotonin uptake has disproved the widely accepted concept that inhibition of monoamine uptake is a prerequisite for antidepressant activity. Another serious objection to this hypothesis is that blockade of monoamine uptake occurs in a matter of minutes after administration while 2–3 weeks of repeated treatment are necessary for the clinical AD effect. Nevertheless, the effect of repeated treatment with AD is compatible with the hypothesis that changes in central monoamine transmission are involved in the clinical activity of these drugs. Major changes in monoamine function after repeated treatment with AD include: desensitization and reduced density of noradrenaline receptors coupled to the adenylcyclase system, opposite changes in the sensitivity of α1 (increased) and α2-adrenoreceptors (decreased), down regulation of serotonin2 receptors and complex changes in the behavioural and electrophysiological responsiveness to serotonin agonists, subsensitivity of presynaptic dopamine receptors and enhanced activity of the mesolimbic dopamine system, decreased and increased density of GABA-A and GABA-B receptors respectively and down regulation of [3H]benzodiazepine binding.It remains to be clarified whether some of these changes have larger roles than others or whether they all contribute to the AD activity. An important role of dopamine in the activity of AD drugs is suggested by findings in the forced swimming test, whereas both catecholamines seem to be involved in the attenuation of escape deficit provoked by inescapable shock (learned helplessness). No clear evidence for a role of serotonin (with the possible exception of serotonin1A receptors) or GABA has been obtained in these experimental models of depression. The general validity of these findings obviously rests on the assumption that these models represent significant aspects of human depression.