Problems with current hypotheses of catecholamine involvement in antidepressant action are reviewed. The theories examined include those that attribute a key role to increased brain norepinephrine (NE) availability or to the desensitization of brain beta adrenergic receptors. Although these hypotheses are consistent with a great deal of relevant laboratory and clinical data there arc an increasing number of findings that cannot be explained by them. These contradictions include findings on the lag time between pharmacological and therapeutic effects of antidepressants, the actions of propranolol and thyroid hormones on depression, the relationship between cortisol secretion and depression, the net effect of chronic antidepressants on noradrenergic neurotransmission, and the functions of cyclic AMP in the CNS. To reconcile the discordant data in these areas of research, a new formulation, the output hypothesis, which is derived from the foregoing theories, is proposed. The new hypothesis assumes that depression or adverse behavioral effects of stress results when the output of brain cells bearing noradrenergic receptors is too low to meet increased demand resulting from stress or other biologically disruptive events. According to this view antidepressants act by a mechanism akin to adaptation to chronic stress in which there is a prolonged increase in postsynaptic noradrenergic receptor activation in the brain. The prolonged increase is necessary for NE to induce, via the beta adrenoceptor–cAMP system, trophic or long-term metabolic effects that increase the output of catecholamine effector cells to a level commensurate with demand.