The cerebellum receives input from visual cortical areas via a relay in the pontine nuclei. We have compared the location in the pontine nuclei of terminal fields of fibres from visual areas 18 and 20, and the posteromedial lateral suprasylvian visual area. Due to individual variations in the precise location of terminal fields, comparisons were performed in individual animals. Horseradish peroxidase-wheat germ agglutinin conjugate was used as an anterograde tracer in combination with the Fink and Heimer method for visualization of anterograde degeneration. Most of the terminal fields of area 20 are widely separated from those of area 18. Fibres from the posteromedial lateral suprasylvian visual area and area 20 terminate close to each other but overlap of terminal fields is limited. Area 18 and the posteromedial area have in some places completely overlapping terminal fields; in other places, however, there is only partial overlap or complete separation. Generally, segregation of terminal fields from different areas is most pronounced in the caudal part of the recipient zone of the pontine nuclei. The terminal fields of fibres from the three cortical areas studied appear as numerous patches arranged in a complicated mosaic that tend to form concentric lamellae around the ventromedial aspect of the peduncle. Within these lamellae, area 18 projects mainly to the innermost one, area 20 to the outermost, and the posteromedial area to an intermediate lamella. Whether terminal fibres from different areas are segregated (non-overlapping) or overlapping in the pontine nuclei is of relevance for the functional organization of the cerebrocerebellar pathway. Segregation of terminal fields from different areas would mean that the areas in question influence different sets of pontocerebellar neurons and thereby relay information to the cerebellum in separate channels. Overlap of terminal fields from different areas could mean that convergence on the same pontocerebellar neurons occurs (although convergence cannot be proved with the techniques employed in this study). This study indicates that information from visual areas is relayed at least in part in separate channels from the cortex to the cerebellum.