Interhemispheric connections were studied in tree shrews (Tupaia belangeri) after multiple injections of horseradish peroxidase or horseradish peroxidase conjugated to wheat germ agglutinin into the cortex of one cerebral hemisphere. After an appropriate survival period, the areal pattern of connections was revealed by flattening the other hemisphere, cutting sections parallel to the cortical surface, and staining with tetramethylbenzidine. Architectonic boundaries were identified by using sections stained for myelinated fibers. Labeled cells and axon terminations formed largely overlapping distributions that covaried in density, although labeled cells appeared to be more evenly distributed than labeled terminations. Connections were concentrated along the border of area 17 (V-I) with area 18 (V-II). However, connections also extended as far as 2 mm into area 17 to include cortex representing parts of the visual field 10 degrees or more from the zero vertical meridian. Clusters of dense connections spanned the width of area 18, where they alternated with regions of fewer connections. These clusters roughly corresponded in location to regions with heavier myelination. In the visually responsive temporal cortex, connections were also unevenly distributed. The organization of most of this cortex is not understood, but one subdivision, the temporal dorsal area (TD), has been identified on the basis of reciprocal connections with area 17. The central part of the TD had few interhemispheric connections, while most of the outer border had dense connections. The auditory cortex had dense and patchy connections throughout. The pattern in the primary somatosensory cortex (S-I) varied according to the representation of body parts, so that the cortex related to the forepaw had sparse connections, while connections were dense but uneven over much of the representation of the face, nose, and mouth. A focus of connections was found at the border of the forepaw and face representations, where the myelination of S-I cortex is interrupted. Dense, uneven connections also characterized the second somatosensory area, S-II. The motor cortex was densely connected, with only slightly fewer terminations rostral to the forepaw region of S-I. Other parts of frontal cortex had dense connections. The distribution of cortical connections varied with depth for at least some areas, so that clusters of cells and terminations were found in supragranular layers in S-I, S-II, and TD, while infragranular labeled cells were more evenly distributed.(ABSTRACT TRUNCATED AT 400 WORDS)