1. Using intracellular recording and voltage-clamp techniques, we examined the biophysical properties of a Ca(2+)-activated slow inward current and its physiological role in plateau potential generation in the dorsal gastric (DG) motor neuron of the stomatogastric ganglion in the crab, Cancer borealis. 2. As shown in the accompanying paper, a brief puff of serotonin (5-HT) evoked a plateau potential in the DG neuron. Intracellular loading of the Ca2+ chelator ethylene glycol-bis (beta-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA) prevented 5-HT from evoking a plateau potential. On the contrary, rapid increase of intracellular Ca2+ by photolysis of caged-Ca2+ (bound to DM-nitrophen) evoked a plateau response in DG bathed in normal saline. 3. Extracellular tetrodotoxin (TTX), tetraethylammonium (TEA), 4-aminopyridine (4-AP), and Cs+ and intracellular iontophoresis of Cs+ were used to block voltage-dependent INa, IK, and Ih. Under these conditions we voltage clamped DG using two electrodes and isolated a long-lasting tail current after a short depolarization of the cell. 4. The reversal potential of the slow tail current was extrapolated to be -27 +/- 3.5 (SE) mV. Na+ substitutions with choline+, tris(hydroxymethyl)aminomethane+ (Tris+) or n-methyl-glucamine+ (NMG+) did not significantly affect the reversal potential or the amplitude. 5. The slow tail current was Ca2+ dependent. It was reduced or abolished by the Ca2+ channel blocker Co2+, intracellular injection of EGTA, and by Ba2+ replacement of Ca2+ as the charge carrier. The activation and deactivation of this current do not show an apparent dependence on voltage. 6. When the voltage-dependent Na+, K+, and Ca2+ channels were blocked, a brief puff of caffeine evoked a slow depolarization. In voltage clamp, caffeine evoked a slow inward current with an apparent conductance increase. This current was reduced by intracellular EGTA. The current-voltage (I-V) relationship of the caffeine-evoked current was linear with a reversal potential of -25 +/- 4.8 mV. This was not statistically different from the reversal potential of the depolarization-evoked tail current. 7. 5-HT enhanced the depolarization-evoked slow tail current but had no effect on the caffeine-evoked slow inward current. 8. We conclude that the slow tail current is a Ca(2+)-activated nonselective current, similar to the Ca(2+)-activated nonspecific cation currents described in other preparations. This current appears to play an important role in plateau generation and maintenance in DG. 5-HT has no direct effect on the properties of this current, but it indirectly enhances the current through an increase of voltage-dependent Ca2+ current.