Recent studies have shown that at least some of the functional effects of serotonin (5-HT) on motoneuron excitability are direct and are mediated via postsynaptic 5-HT receptors on motoneurons. To determine the spatial distribution of direct inputs from the serotonin system on the proximal and distal dendrites of individual motoneurons, we examined identified motoneurons in vivo with a combination of immunohistochemical localization of 5-HT-immunoreactive boutons and intracellular staining with horseradish peroxidase. Seventeen intracellularly stained motoneurons from 12 adult cats were analyzed with light microscopy. Quantitative analysis of 5-HT boutons apposed to dendrites of five representative motoneurons that were entirely reconstructed in three dimensions (each from the lumbosacral spinal cord of a different animal) revealed a total of 7,848 contacts (1,570+/-487 contacts/postsynaptic neuron; mean +/- SD) over the dendrites of these cells. Analysis of contacts on the soma of two of these cells, and on the somas of an additional 12 intracellularly stained motoneurons, revealed a wide range of somatic contacts (11-211 contacts/cell) on motoneuron cell bodies, with an average of 52 contacts/cell. These results indicate that the vast majority of 5-HT-immunoreactive boutons are apposed to dendritic branches rather than to the somatic surface of motoneurons. The spatial distribution of contacts essentially matched the distribution of surface membrane area of the postsynaptic neuron, resulting in a relatively uniform density of contacts (<1/100 microm2) on proximal and distal dendrites. Consequently, the frequency of contacts was higher on the proximal dendritic compartments where available membrane area is greater. There was no preferential distribution of contacts to particular dendrites. Light/electron microscopic correlations were performed on 21 boutons that contacted dendrites (n = 7) of three motoneurons from different animals. At the electron microscope level, most appositions (18/21; 85.7%) selected by our light microscopic criteria were confirmed as direct contacts when the 5-HT boutons were examined through serial sections. Synaptic junctions, generally small and symmetric, were positively identified in only a subset of these cases (n = 6; 28.6%), in part due to the obscuring effects of the peroxidase histochemical precipitate present in both pre- and postsynaptic profiles. A few 5-HT boutons (3/21; 14.3%) selected as contacts by our light microscopic criteria were in fact separated from the adjacent labeled dendrites; in two of these three cases, the separation was due to intrusion of very thin glial lamellae (<0.3 microm in cross section). These results indicate that the bulbospinal serotonergic system(s) provide a significant, direct synaptic input to spinal motoneurons that innervate hindlimb muscles. The nature of the modulatory actions exerted by such widespread synaptic inputs will affect all regions of the somatodendritic membrane and will ultimately depend on the nature of the 5-HT receptors present over different parts of the postsynaptic neuron's dendritic tree.