Several parameters of the 125I-C1q-binding test were investigated: ionic strength, pH, concentration of EDTA, influence of serum C1q and the possibility of interference by polyanions. Lowering the ionic strength of the borate buffer resulted in increased precipitation of 125I-C1q in normal human serum. This increase was dependent on the presence of serum proteins, probably immunoglobulins. When the concentration of the EDTA was decreased, increased precipitation of 125I-C1q in normal human serum was also observed. This was prevented by adding NaCl to the EDTA solution. However at very low concentrations of EDTA (too low to chelate all calcium ions in the serum), increased precipitation of 125I-C1q in normal human serum was observed even in the presence of added NaCl. Addition of purified C1q to sera from patients with very low C1q levels had varying effects on the results of the C1q-binding test: (a) it decreased the C1q-binding activity of some sera, probably by competition with 125I-C1q for binding sites on the immune complexes; (b) it increased the C1q-binding activity of other sera, probably by enhancing the precipitation of immune complexes as a consequence of the cross-linking effect of C1q; or (c) it had no influence, possibly due to the opposite effects of (a) and (b). The addition of dextran sulphate resulted in a dose-dependent increase in the 125I-C1q-binding activity of normal human serum. This effect was dependent on the interaction of dextran sulphate with either C1q or low-density lipoproteins and was prevented by addition of polybrene to the assay. However, addition of polybrene to sera with a high C1q-binding activity scarcely influenced binding activity.