C receptors 1 (CR1) of human E are involved in the transport of C3b-coated immune complexes (IC) in the circulation. Many studies have suggested that the binding of IC to E is multivalent. This would require CR1 to be clustered on the cell membrane, but no direct evidence for such clustering is available. We studied the distribution of CR1 on human E by immunofluorescence and shadow-casting immuno-electron microscopy techniques with the use of a monoclonal anti-CR1 antibody followed by FITC- or gold-conjugated second antibodies, respectively. By immunofluorescence, CR1 appeared as small dots (clusters) on fixed and unfixed E prepared either at 4 degrees C or at 37 degrees C. In the same donor, the number of clusters varied extensively from cell to cell (e.g., 1 to 43 clusters/E for a donor with 520 CR1/cell), but the mean number of clusters per cell correlated significantly with the mean number of CR1/cell. These images contrasted with those obtained for Rhesus D (RhD) Ag used as controls (RhD Ag are known to be evenly distributed): only a faint uniform fluorescence was seen despite the presence of 10,000 antigenic sites. As determined by immunocytochemical method, more than 65% of the total gold particles were organized in clusters (2 to 15 gold particles/cluster) whether cells were prefixed or not. Quantitative determinations suggested that each gold particle corresponded to one CR1. The fraction of gold particles grouped into clusters of three or more receptors, the mean size of the clusters, and the maximal size of clusters correlated with the mean number of CR1 per cell. By contrast, RhD Ag were distributed homogeneously (less than 2% gold particles in clusters). These data are the first to demonstrate the preclustered nature of CR1 on E. Such distribution could explain the high binding efficiency of C3b-coated IC to E despite the low number of CR1 per cell.