We examined the ability of the isolated lumbosacral spinal cord of the neonatal mouse (P0-7) to generate rhythmic motor activity under several different conditions. In the absence of electrical or pharmacological stimulation, we recorded several patterns of spontaneous ventral root depolarization and discharge. Spontaneous, alternating discharge between contralateral ventral roots could occur two to three times over a 10-min interval. We also observed other patterns, including left-right synchrony and rhythmic activity restricted to one side of the cord. Trains of stimuli delivered to the lumbar/coccygeal dorsal roots or the sural nerve reliably evoked episodes of rhythmic activity. During these evoked episodes, rhythmic ventral root discharges could occur on one side of the cord or could alternate from side to side. Bath application of a combination of N-methyl-D,L-aspartate (NMA), serotonin, and dopamine produced rhythmic activity that could last for several hours. Under these conditions, the discharge recorded from the left and right L(1)-L(3) ventral roots alternated. In the L(4)-L(5) segments, the discharge had two peaks in each cycle, coincident with discharge of the ipsilateral and contralateral L(1)-L(3) roots. The L(6) ventral root discharge alternated with that recorded from the ipsilateral L(1)-L(3) roots. We established that the drug-induced rhythm was locomotor-like by recording an alternating pattern of discharge between ipsilateral flexor and extensor hindlimb muscle nerves. In addition, by recording simultaneously from ventral roots and muscle nerves, we established that ankle flexor discharge was in phase with ipsilateral L(1)/L(2) ventral root discharge, while extensor discharge was in phase with ipsilateral L(6) ventral root discharge. Rhythmic patterns of ventral root discharge were preserved following mid-sagittal section of the spinal cord, demonstrating that reciprocal inhibitory connections between the left and right sides of the cord are not essential for rhythmogenesis in the neonatal mouse cord. Blocking N-methyl-D-aspartate receptors, in both the intact and the hemisected preparation, revealed that these receptors contribute to but are not essential for rhythmogenesis. In contrast, the rhythm was abolished following blockade of kainate/AMPA receptors with 6-cyano-7-nitroquinoxalene-2,3-dione. These findings demonstrate that the isolated mouse spinal cord can produce a variety of coordinated activities, including locomotor-like activity. The ability to study these behaviors under a variety of different conditions offers promise for future studies of rhythmogenic mechanisms in this preparation.