Agents known to influence adrenergic function were examined for their mydriatic effects in urethane-anesthetized rats. Both the direct acting adrenergic agonist, clonidine, and the "indirect" acting agonists, cocaine, desipramine and amphetamine elicited mydriatic responses. The polar alpha-2 adrenoceptor agonist, 4-hydroxyclonidine, did not elicit a mydriatic response when administered systemically; however, it did produce a pronounced mydriatic response when administered i.c.v. Inasmuch as the selective lipophilic alpha-2 adrenoceptor antagonist, yohimbine, but not the polar alpha adrenoceptor antagonist, phentolamine, reversed the mydriatic effects of clonidine and cocaine, these data suggest that centrally located alpha-2 adrenoceptors elicit the above mydriatic response. The direct acting alpha-2 adrenoceptor agonists, clonidine (i.v.) and 4-hydroxyclonidine (i.c.v.), were the most efficacious of the agents studied in eliciting the mydriatic response. The indirect acting agents, amphetamine, desipramine and cocaine, were less efficacious. The rank order potency (ED-50) of these drugs was as follows: clonidine greater than desipramine greater than cocaine = amphetamine. The mydriatic effects of cocaine were attenuated by yohimbine and reserpine pretreatments. In addition, the local anesthetic, procaine, and the polar cocaine analog, cocaine methiodide, were significantly less efficacious than cocaine. These results suggest that cocaine elicits mydriasis by indirectly acting at central and postsynaptically located alpha-2 adrenoceptors.