When stressed, toadfish become ureotelic and excrete almost all of their nitrogenous waste in 1-3 daily pulses of urea-N across the gills. Intravascular injections of 5-hydroxytyptamine (5-HT; serotonin) and analogues also elicit marked excretory pulses of urea-N from toadfish in vivo, suggesting that 5-HT release is the proximate trigger for spontaneous pulses. However it is unclear whether 5-HT is acting on the gills directly or elsewhere to cause the effect indirectly. A perfused whole gill preparation which maintained normal pressure relationships and stable vascular resistance was employed to address this question. Bolus injections into the ventral aortic perfusate of either 5-HT (1, 10 μmol kg(-1)) or the specific 5-HT(2) receptor agonist α-methyl 5-HT (1, 10 μmol kg(-1)) elicited rapid urea-N pulses from perfused toadfish gills. The effective doses, the post-injection delays (5.5 ± 1.3 min, range=2-22), the percent occurrences (57-85%), and the magnitude of the induced urea-N pulses (615.4 ± 131.3 μmol-N kg(-1), range 66.0-2634.0), were all similar to those previously reported when these agents were injected in vivo. Bolus injections of 5-HT and α-methyl 5-HT also elicited a biphasic response in ventral aortic pressure, reflecting an initial rapid short-lived vasodilation and a subsequent longer-lasting vasoconstriction. These events were similar to those which have been recorded to occur at a greater frequency during spontaneous urea-N pulsing in vivo. Neither the urea-N pulsing nor the cardiovascular responses to 5-HT were inhibited by the 5-HT(2A) receptor subtype blocker, ketanserin (pre-injection with 10 μmol kg(-1) plus 33 μmol L(-1) in the perfusate). Overall, these results provide strong support for the idea that the proximate stimulus for natural urea pulsing in vivo is 5-HT mobilization, acting directly in the gills.