1. Larval crawling is a bilaterally symmetrical behavior that involves an anterior moving wave of motor activity in the body wall muscles in conjunction with sequential movements of the abdominal prolegs and thoracic legs. The purpose of this study was to determine whether the larval CNS by itself and without phasic sensory feedback was capable of producing patterned activity associated with crawling. To establish the extent of similarity between the output of the isolated nerve cord and crawling, the motor activity produced in isolated larval nerve cords was compared with the motor activity from freely crawling larvae. 2. When exposed to the muscarinic receptor agonist pilocarpine (1.0 mM), isolated larval nerve cords produced long-lasting rhythmic activity in the motor neurons that supply the thoracic leg, abdominal body wall, and abdominal proleg muscles. The rhythmic activity evoked by pilocarpine was abolished reversibly and completely by bath application of the muscarinic-receptor antagonist atropine (0.01 mM) in conjunction with pilocarpine (1.0 mM), suggesting that the response was mediated by muscarinic-like acetylcholine receptors. 3. Similar to crawling in intact animals, the evoked activity in isolated nerve cords involved bilaterally symmetrical motor activity that progressed from the most posterior abdominal segment to the most anterior thoracic segment. The rhythmic activity in thoracic leg, abdominal proleg, and abdominal body wall motor neurons showed intrasegmental and intersegmental cycle-to-cycle coupling. The average cycle period for rhythmic activity in the isolated nerve cord was approximately 2.5 times slower than the cycle period for crawling in intact larvae, but not more variable. 4. Like crawling in intact animals, in isolated nerve cords, bursting activity in the dorsal body wall motor neurons occurred before activity in ventral/lateral body wall motor neurons within an abdominal segment. The evoked bursting activity recorded from the proleg nerve was superimposed on a high level of tonic activity. 5. In isolated nerve cords, bursts of activity in the thoracic leg levator/extensor motor neurons alternated with bursts of activity in the depressor/flexor motor neurons. The burst duration of the levator/extensor activity was brief and remained relatively steady as cycle period increased. The burst duration of the depressor/ flexor activity occupied the majority of an average cycle and increased as cycle period increased. The phase of both levator/extensor motor nerve activity and depressor/flexor motor nerve activity remained relatively stable over the entire range of cycle periods. The timing and patterning of thoracic leg motor neuron activity in isolated nerve cords quantitatively resembled thoracic leg motor activity in freely crawling larvae. 6. The rhythmic motor activity generated by an isolated larval nerve cord resembled a slower version of normal crawling in intact larvae. Because of the many similarities between activity induced in the isolated nerve cord and the muscle activity and movements of thoracic and abdominal segments during crawling, we concluded that central mechanisms can establish the timing and patterning of the crawling motor pattern and that crawling may reflect the output of a central pattern generating network.