The selectivity, peripheral vs. central actions, of the antidopaminergic agent L-646,462 was assessed in two ways. First, elevation of prolactin in serum (peripheral) and homovanillic acid in the striatum were measured in rats. L-646,462 was found to have a central/peripheral activity ratio of 143, whereas comparable values derived for haloperidol, metoclopramide and domperidone were 1.4, 9.4 and 1305, respectively. Second, the ID50 values required to block apomorphine-induced emesis in beagles (peripheral receptor-mediated response) were compared with those required to block apomorphine-induced stereotypy (central receptor-mediated response) in rats. Central/peripheral ID50 ratios of 234, 9.2, 129 and 7040 were obtained, respectively, for L-646,462, haloperidol, metoclopramide and domperidone. The selectivity of L-646,462 for peripheral serotonin (5-HT) receptors in rats was determined by measuring its effectiveness in blocking 5-HT-induced paw edema (peripheral response) and 5-hydroxytryptophan-induced head twitch (central response); a ratio of 114 was obtained. This value agrees nicely with the ratio of 143 derived in the rat ( vide supra) for peripheral selectivity for dopamine receptors. L-646,462 is, therefore, selective in vivo, preferentially blocking dopamine and 5-HT receptors located outside the blood-brain barrier. With regard to dopamine-receptors, L-646,462 was about equipotent and more selective than metoclopramide, while being less potent and less selective than domperidone. Unlike metoclopramide or domperidone, L-646,462 also possessed a reasonably potent 5-HT receptor antagonist effect in vivo.(ABSTRACT TRUNCATED AT 250 WORDS)