By radioimmunoassay we measured the amount of endogenous C1q that was precipitated by polyethylene glycol (PEG) under the conditions of the 125I-C1q-binding test (C1q-BT). We found a linear correlation between the percentage endogenous C1q that was precipitated and the 125I-C1q-binding activity (C1q-BA). We concluded that the 125I-C1q behaves like the endogenous C1q. To detect circulating immune complexes (CIC) which had already bound C1q, human sera were added to tubes coated with anti-C1q. Under the conditions used, no C1q-bearing CIC were detected. In addition, 7 sera from patients with high C1q-BA were analyzed by sucrose-gradient ultracentrifugation. No C1q was found in the fast sedimenting fractions, although C1q-BA was detected in these fractions. With IgG-coated tubes we observed that PEG enhanced the binding of 125I-C1q as well as endogenous C1q to aggregated and monomeric IgG. PEG also enhanced the binding of CIC to C1q-coated tubes. The results suggest that CIC detected by the C1q-BT do not bear C1q in significant amounts in the circulation and that these CIC become detectable only in the presence of PEG.