The results of antinuclear antibody tests using the indirect immunofluorescence technique may be reported as a description of the pattern and the intensity of fluorescence obtained at a certain dilution. If quantitative results are required titration is necessary. Such titrations may vary greatly between different laboratories. The present study involving 26 laboratories shows an improvement of interlaboratory comparability for the homogeneous fluorescence pattern when a common reference serum is used. Cultured cells as substrate appear to give better quantitative agreement than rat liver sections. National reference sera should be standardised in items of the appropriate WHO reference preparation.