1. The discharge patterns of 91 identified pyramidal tract neurons (PTNs), located within the forelimb region of area 4 of the cat motor cortex, were recorded during the voluntary modifications of gait needed to step over obstacles attached to a moving treadmill belt. Recordings were made simultaneously from flexor and extensor muscles acting around the shoulder, elbow, wrist, and digits of the forelimb contralateral to the recording site. 2. Analysis of the changes in electromyographic (EMG) activity during the gait modification showed increases in the activity of most flexor muscles of the shoulder and elbow, as well as in the wrist and digit dorsiflexors, when the contralateral forelimb was the first to pass over the obstacle. This period of augmented activity could be subdivided into two parts: one associated with the initial flexion of the limb that was needed to bring it above and over the obstacle (phase I), and the second associated with increased wrist dorsiflexor muscle activity before foot contact (phase II). 3. The discharge frequency of a total of 57/91 (63%) of the recorded PTNs was significantly increased during the gait modification when the limb contralateral to the recording site was the first to step over the obstacle; six of these neurons also showed a significant decrease in their discharge in a different part of the step cycle. In a further 21/91 (23%) neurons, discharge frequency was only decreased, whereas the remaining 13/91 (14%) PTNs showed similar patterns of activity both during control walking and during the gait modifications. 4. Most of those neurons (47/57) in which significant increases in firing frequency were observed, discharged maximally during the period of increased activity of the physiological flexor muscles. Twenty-three of these cells (23/47) discharged maximally in phase I, and 12 (12/47) in phase II. A third population of PTNS (12/47) started to increase their discharge in the stance phase of the step cycle immediately preceding the modified cycle. Seven (7/57) PTNs increased their discharge during the stance phase of the modified cycle, and the remaining three could not be classified as being preferentially related to any one part of the step cycle. 5. The frequency modulation of 41/57 PTNs was less when the leg contralateral to the recording site was the second to encounter the obstacle. In many neurons there was also an appreciable change in the time in the step cycle that peak discharge occurred. These changes in amplitude and timing paralleled the changes observed in the temporal relationships of the muscles.(ABSTRACT TRUNCATED AT 400 WORDS)