The many different recognized functions of C3 are dependent upon the ability of the activated C3 molecule both to bind covalently to protein and carbohydrate surfaces and to provide binding sites for as many as eleven different proteins. The location of the binding sites for six of these different proteins (factors B and H, complement receptors CR(1), CR(2) and CR(3) and conglutinin) was examined in the naturally occurring C3-fragments generated by C3 activation (C3b) and degradation by Factor I (iC3b, C3c, C3d,g) and trypsin (C3d). Evidence was obtained for at least four distinct binding sites in C3 for these six different C3 ligands. One binding site for B was detectable only in C3b, whereas a second binding site for H and CR(1) was detectable in both C3b and iC3b. The affinity of the binding site for H and CR(1) was charge dependent and considerably reduced in iC3b as compared to C3b. H binding to iC3b-coated sheep erythrocytes (EC3bi) was measurable only in low ionic strength buffer (4 mS). The finding that C3c-coated microspheres bound to CR(1), indicated that this second binding site was still intact in the C3c fragment. However, H binding to C3c was not examined. A third binding site in C3 for CR(2) was exposed in the d region by factor I cleavage of C3b into iC3b, and the activity of this site was unaffected by the further I cleavage of iC3b into C3d,g. Removal of the 8,000-dalton C3g fragment from C3d,g with trypsin forming C3d, resulted in reduced CR2 activity. However, because saturating amounts of monoclonal anti-C3g did not block the CR(2)-binding activity of EC3d,g, it appears unlikely that the g region of C3d,g or iC3b forms a part of the CR(2)-binding site. In addition, detergent-solubilized EC3d (C3d-OR) inhibited the CR(2)-binding activity of EC3d,g. Monocytes and neutrophils, that had been previously thought to lack CR(2) because of their inability to form EC3d rosettes, did bind EC3d,g containing greater than 5 × 10(4) C3d,g molecules per E. The finding that monocyte and neutrophil rosettes with EC3d,g were inhibited by C3d-OR, suggested that these phagocytic cells might indeed express very low numbers of CR(2), and that these CR(2) were detectable with EC3d,g and not with EC3d because C3d,g had a higher affinity for CR2 than did C3d. A fourth C3 binding site for CR(3) and conglutinin (K) was restricted to the iC3b fragment. Because of simultaneous attachment of iC3b to phagocyte CR3 and CR(3), the characteristics of iC3b binding to CR3 could only be examined with phagocytes on which the CR(1) had been blocked with anti-CR(1). Inhibition studies with EDTA and N-acetyl-D-glucosamine demonstrated a requirement for both calcium cations and carbohydrate in the binding of EC3bi to CR3 and to K. However, CR(3) differed from K in that magnesium cations were required in addition to calcium for maximum CR(3) binding activity, and NADG produced less inhibition of CR(3) activity than of K activity.