1. We describe a group of serotonergic neurons in the pedal ganglia of Aplysia brasiliana and characterize their modulatory effects on the motoneuron input to swimming muscles of the parapodia. Each pedal ganglion contains one cluster of large neurons near its dorsomedial surface that fires in phase with opening (downstroke) of the parapodia; these have been designated parapodial opener-phase (POP) cells. 2. POP cells are large, number 15-20 per ganglion, have peripheral axons in parapodial nerves, have distinctively shaped action potentials, and fire in bursts phasically with motoneurons during the opening, or downstroke portion, of parapodial movement during fictive swimming. Firing individual POP cells with intracellular current indicates that they have no direct detectable effect on muscle, causing neither junction potentials nor contractions. 3. 5,7-Dihydroxytryptamine (5,7-DHT) staining, immunocytochemistry using serotonin (5-HT) antibodies, and direct biochemical measurements revealed that POP cells are serotonergic. Serotonergic nerve endings were also seen in parapodial muscle. 4. Simultaneous intracellular recordings and use of altered divalent concentrations revealed that no detectable direct synaptic interactions exist between POP cells and motor neurons. 5. When POP cells and motoneurons were simultaneously recorded while measuring muscle contractions, it was found that POP cell activity enhances motoneuron-induced tension by 120-900%, averaging around 300%. Variability in the efficacy of individual POP cells suggests that they may influence specific regions or groups of muscle fibers. 6. POP cell activity also significantly increased the rate of relaxation of parapodial muscle contractions, averaging about a 40% reduction in the time required to relax to one-half peak tension. Increased relaxation rate implies a postsynaptic change in muscle behavior. 7. The effectiveness of POP cells to increase contraction tension and relaxation rate was positively correlated with POP cell spike frequency. These effects were slow (seconds) in onset and could persist for a minute or more after cessation of POP firing. Concurrent motoneuron activity is not required for modulation by POP cells. 8. Simultaneous intracellular recording from a POP cell, motoneuron, and muscle fiber revealed that POP cell activity enhanced the amplitude of motoneuron-induced excitatory junction potentials (EJPs). Activity of POP cells did not alter muscle fiber membrane potential, but the experiments left open the possibility that EJP enhancement is presynaptic, postsynaptic, or a combination.(ABSTRACT TRUNCATED AT 400 WORDS)