1. In the isolated stomatogastric nervous system of the crab Cancer borealis (Fig. 1), the muscarinic agonist oxotremorine elicits several distinct gastric mill motor patterns from neurons in the stomatogastric ganglion (STG; Fig. 2). Selection of a particular gastric mill rhythm is determined by activation of distinct projection neurons that influence gastric mill neurons within the STG. In this paper we identify one such neuron, called commissural projection neuron 2 (CPN2), whose rhythmic activity is integral in producing one form of the gastric mill rhythm. 2. There is a CPN2 soma and neuropilar arborization in each commissural ganglion (CoG). The CPN2 axon projects through the superior esophageal nerve (son) and the stomatogastric nerve (stn) to influence neurons in the STG (Figs. 3 and 4A). 3. CPN2 activity influences most of the gastric mill neurons in the STG. Specifically, CPN2 excites gastric mill neurons GM and LG (gastric mill and lateral gastric, respectively) and inhibits the dorsal gastric (DG), anterior median (AM), medial gastric (MG), and inferior cardiac (IC) neurons (Figs. 5 and 6). CPN2 also indirectly inhibits gastric mill neurons Int1 and VD (interneuron 1 and ventricular dilator neuron, respectively) through its activation of LG. The CPN2 excitatory effects are mediated at least partly via discrete excitatory postsynaptic potentials (EPSPs; Fig. 4B), whereas its inhibitory effects are produced via smooth hyperpolarizations. 4. Within the CoG, CPN2 receives excitatory synaptic input from the anterior gastric receptor neuron (AGR), a gastric mill proprioceptive sensory neuron (Fig. 7) and inhibitory synaptic input from the gastric mill interneuron, Int1 (Fig. 8). 5. During one form of the gastric mill rhythm, CPN2 fires rhythmically in time with the gastric mill motor pattern, whereas it is silent or fires weakly during other gastric mill rhythms (Fig. 9). 6. When CPN2 rhythmic activity is suppressed during a CPN2-influenced gastric mill rhythm, the gastric mill rhythm continues, but the pattern is altered (Fig. 10). Moreover, transiently stimulating CPN2 during any ongoing gastric mill motor pattern can reset the timing of that rhythm (Fig. 11). 7. Tonic activity in CPN2 is insufficient to elicit a gastric mill rhythm (Fig. 12). Phasic activity in CPN2 can elicit a gastric mill rhythm only in preparations in which gastric mill neurons are already in an excited state (Figs. 12 and 13). 8. CPN2 recruitment plays a pivotal role in determining the final form of the gastric mill rhythm.(ABSTRACT TRUNCATED AT 400 WORDS)