The relationship of the climbing fiber afferent discharge to the unperturbed and perturbed step cycle was evaluated in the cat. Following a precollicular-premamillary decerebration, cats walked spontaneously on a motorized treadmill. Purkinje cells were recorded extracellularly and simple and complex spikes were discriminated. Right forelimb displacement, biceps and triceps EMG activity, as well as treadmill velocity, were also monitored. In some animals pressure measurements of the contact of the footpad with the treadmill were obtained. Cells were studied during both "normal" and perturbed locomotion. The perturbation consisted of a braking of the treadmill at different phases in the step cycle. Histograms of the simple and complex spike activity, and averages of the right forelimb displacement, biceps, and triceps EMG activity and treadmill velocity were constructed. The complex spike activity of 163 Purkinje cells was averaged through a minimum of 50 sweeps in either normal and/or perturbed locomotion. Statistical analysis revealed that the probability of the climbing fiber afferent discharge in 54% of the cells (36/67) studied during normal locomotion was significantly modulated with the step cycle. For most Purkinje cells the onset of the increase in climbing fiber afferent discharge was coupled to triceps activity and the onset of stance phase. A group of cells exhibited complex spike discharge in association with biceps onset and swing. These observations suggest that complex spike discharge occurs preferentially at the phase transition periods in the step cycle when the trajectory of the forelimb changes from swing to stance or stance to swing. During treadmill braking 51% of the cells exhibited complex spike modulation (70/137). A number of different patterns of climbing fiber afferent modulation occurred. The most common pattern was an increase in complex spike discharge with the resumption of the treadmill movement and locomotion. Analysis of the time of these periods of increased climbing fiber activity suggests that, although in some cells the response is coupled to the treadmill onset, in other cells the modulation occurs at longer latencies. Subsequent analysis aligning the EMG, displacement, and treadmill velocity signals with the times of the climbing fiber afferent discharge suggested some responses were coupled to the reinitiation of the locomotor cycle. The second most common pattern was an increase in climbing fiber afferent discharge at the onset of the perturbation. Also, in some cells, complex spike discharge decreased during the period in which the step cycle was arrested.(ABSTRACT TRUNCATED AT 400 WORDS)