1. The issue of isochronicity of olivocerebellar fibre conduction time as a basis for synchronizing complex spike activity in cerebellar Purkinje cells has been addressed by latency measurement, multiple-electrode recording and Phaseolus vulgaris leucoagglutinin (PHA-L) tracing of climbing fibres in the adult rat. 2. The conduction time of the olivocerebellar fibres was measured by recording Purkinje cell complex spike (CS) responses from various areas of the cerebellum. The CSs were evoked by stimulating the olivocerebellar fibres near the inferior olive. In spite of a difference in length, as determined directly by light microscopy, the conduction times of different climbing fibres were quite uniform, 3.98 +/- 0.36 ms (mean +/- S.D., n = 660). 3. Multiple-electrode recording of spontaneous Purkinje cell CS activity was employed to study the spatial extent of CS synchronicity in the cerebellar cortex. Recordings of CS were obtained from Purkinje cells located on the surface and along the walls of lobule crus 2a. The rostrocaudal band-like distribution of simultaneous (within 1 ms) CS activity in Purkinje cells extended down the sides of the cerebellar folia to the deepest areas recorded (1.6-2.6 mm deep). As shown in previous experiments, the distribution of simultaneous CS activity did not extend significantly (500 microns) in the mediolateral axis of the cerebellar cortex. 4. In two animals a detailed determination of the length of the olivocerebellar fibre bundles was performed by staining the fibres with PHA-L injected into the contralateral inferior olive. This measurement included fibre bundles terminating in twenty-six different areas, ranging from the tops of the various folia to the bottoms of the fissures in both the hemisphere and the vermis. There was a 47.5% difference between the length of the longest measured fibre bundle (15.8 mm, terminating in lobule 6b, zone A) and the length of the shortest measured fibre bundle (8.3 mm, terminating in the cortex at the base of the primary fissure, zone D), after correction for tissue shrinkage. To attain an isochronous conduction time the conduction velocities for these two fibre bundles were calculated to be 4.22 m/s and 2.37 m/s, respectively. 5. By interpolating between measured points a simple formula was derived to estimate the average length of olivocerebellar fibres terminating in any given area of the cerebellar cortex, excluding the paraflocculus, the flocculus and the most lateral regions of the hemisphere. 6. We investigated the most likely mechanisms by which conduction velocity variations with length could result in global isochronicity.(ABSTRACT TRUNCATED AT 400 WORDS)