The present experiments were designed to examine the characteristics of the temporal patterning present in the spike train of Purkinje cells and to examine the relationship between this patterning and the occurrence of climbing fiber inputs to the same neuron. For this purpose two different methods were used to determine the autocorrelation characterizing the spontaneous simple spike activity of Purkinje cells recorded extracellularly in unanesthetized, decerebrate cats. The first method of calculating the autocorrelation is analogous to that commonly used in central nervous system electrophysiology for stationary stochastic processes and is referred to as the standard autocorrelation. The second method is derived for non-stationary stochastic processes and is referred to as the generalized autocorrelation function (GACF). Two classes of Purkinje cell activity were revealed using the standard autocorrelograms. The type 1 pattern was characterized by a strong positive correlation at short lag times, which decayed to base line over periods greater than 50 ms. The type 2 autocorrelation also exhibited a positive correlation at short lag times. However, the positive correlation in the simple spike activity of this population of cells rapidly decayed to base line within 10-15 ms. These two types of autocorrelations do not appear to be found in two distinct populations of Purkinje cells. Type 1 and type 2 correlograms can be generated by the same cell under different conditions. The use of GACF allowed an examination of the relationship between the occurrence of a climbing fiber input and the patterning present in the simple spike discharge of Purkinje cells. Using this technique, the long-duration positive correlation occurring in the simple spike activity of Purkinje cells possessing a type 1 autocorrelogram was found to occur preferentially in the simple spike activity following a climbing fiber response. This was demonstrated by comparing the GACF of the simple spike activity constructed by triggering on the occurrence of climbing fiber inputs and the GACF of the same simple spike activity constructed by triggering on randomly occurring simple spikes. For those Purkinje cells with a type 2 autocorrelogram, the GACF constructed by triggering on the climbing fiber input did not have a long-duration positive correlation in the simple spike activity following this input. The short-duration positive correlation that was present in these GACFs was similar to that observed in the GACF triggered on the simple spike activity of these cells. In some animals the effects of synchronously activating the climbing fiber input to the cerebellar cortex by olivary stimuli were examined. This procedure usually augmented the long-duration positive correlation occurring in the simple spike activity for type 1 autocorrelations and could convert type 2 autocorrelation patterns into type 1. These alterations could occur independent of the effect of the stimuli on the excitability of the cell as determined by poststimulus time histograms (PSTHs). These results suggest that the simple spike activity following the climbing fiber event possesses a specific temporal pattern, which is responsible for the features observed in the type 1 autocorrelogram found in some Purkinje cells. It is postulated that the occurrence of this patterning produced by the action of climbing fiber afferents may partly determine the responsiveness of Purkinje cells to inputs via the mossy fiber-granule cell system.