Recordings were made from single units in the medullary reticular formation (MRF) between AP-4.2 and AP-12.9 and from the midline to 3.7 mm lateral in chronically prepared, unrestrained cats walking on a treadmill. Recordings were made with rigid microelectrodes held in a microdrive, and reticulospinal neurons were identified by antidromic stimulation of their axons through microwires chronically implanted into the spinal cord at the L2 level. Electromyograms (EMGs) were recorded from flexor and extensor muscles of the fore- and hindlimbs as well as from back and neck muscles. In total, 295 cells were recorded from 40 penetrations in 4 cats; 252 of these cells were recorded from the more medial regions of the reticular formation encompassing the gigantocellular, magnocellular, and lateral tegmental fields; 38.5% of these (97/252) were antidromically identified from the spinal cord. The remaining 43 neurons (43/295) were recorded from a more lateral and ventral position. These medial and ventrolateral groups of neurons differed not only in position but also in aspects of their discharge during locomotion. Rank-ordered raster displays, triggered from the onset of each recorded muscle, were used to correlate neuronal and muscular activity. The discharge rate of 31% of the reticulospinal neurons (30/97) was modulated once or twice in each step cycle and was strictly related to one or more of the recorded EMGs (EMG-related neurons) on the basis of the pattern of discharge. The discharge of 33/97 (34%) of the neurons was modulated at the periodicity of the locomotor rhythm but could not be correlated with any of the recorded EMGs (locomotor-related cells), whereas the remaining 34/97 neurons (35%) were either silent, fired tonically, or were not related to the locomotor pattern (unrelated cells). Of the EMG-related neurons 27% were related to flexor muscles and the remaining 63% to extensor muscle activity. The discharge pattern of all except two of the flexor-related neurons was correlated with hindlimb muscle activity, whereas that of the extensor-related neurons was correlated almost equally with fore- and hindlimb muscles. Correlations were found with muscles lying both ipsilaterally and contralaterally to the site of the recordings. Although the locomotor-related neurons showed no preferential relation with any of the recorded EMGs, a comparison of the depth of modulation of their discharge measured from postevent histograms suggested that more of these cells were related to the forelimb than to the hindlimb.(ABSTRACT TRUNCATED AT 400 WORDS)