With the immunofluorescence technique (IFT) using Crithidia luciliae as a substrate, 14,417 sera sent to our laboratory for routine anti-dsDNA determination, were screened for the presence of antibodies to dsDNA. The 1,260 sera that were found IFT positive were then assayed with the Farr radioimmunoassay, in which 3H-labelled PM2-DNA is used as antigen. Only 470 sera (37%) were found to be Farr positive. This discrepancy is, at least partially, caused by the fact that the Farr assay does not detect anti-DNA of low avidity, whereas the Crithidia-IFT does. Sixty-eight percent of the IFT-positive/Farr negative sera were found positive with the PEG assay, a radioimmunoassay that also employs double stranded PM2-DNA as antigen, and that also detects anti-dsDNA of low avidity. The IFT performed on IFT positive/Farr negative sera was found to be rather irreproducible. It was shown that this was due to local increases of the salt concentration resulting from the way the assay was performed. The problem could be overcome by careful control of the assay conditions, i.e. never letting Crithidia slides dry up after washing with PBS. In the PEG assay, these sera sometimes showed a DNA binding that decreased with time. It could be shown that this is caused by a parallel increase in pH during the incubation as a result of CO2 evaporation from the serum.