A series of experiments was performed to evaluate the mechanism(s) by which chronic administration of desipramine (DMI) facilitates the locomotor stimulant action of d-amphetamine, a response thought to be dependent on the mesolimbic dopaminergic system. Prior lesions of central noradrenergic or serotonergic neurons, induced by neonatal treatment with 6-hydroxydopamine (6-OHDA) or intraventricular injections of 5,7-dihydroxytryptamine (5,7-DHT) respectively, failed to block DMI-induced facilitation of amphetamine hypermotility. Chronic administration of DMI did not significantly influence specific [3H]spiperone binding in the striatum or the nucleus accumbens. In other experiments it was found that chronic administration of some (amitryptyline, imipramine, mianserin, iprindole) but not all (zimelidine, nomifensine, fluoxetine) antidepressants enhanced the locomotor response to d-amphetamine. The weak anticholinergic effects of the latter compounds suggest that the positive results obtained with the former drugs may be related to their anticholinergic properties. This hypothesis is consistent with the observation that chronic administration of scopolamine also increased the locomotor response to d-amphetamine. The results suggest that the facilitation by chronic DMI of amphetamine-induced locomotor activity is not mediated by primary actions of this tricyclic antidepressant on central noradrenergic or serotonergic systems. In addition, the results argue against an effect of DMI on dopamine receptors as measured by [3H]spiperone binding. Instead, the facilitation of the amphetamine response by some of these antidepressant compounds may be related to their anticholinergic effects. The results are viewed as being consistent with a dopaminergic-cholinergic hypothesis of affective illness and suggest that the anticholinergic properties of some antidepressant compounds may contribute to their therapeutic effects.