Estrogen receptor determinations have been performed on 241 cytosols from 160 breast cancer tumors using both radioactive ligands ([3H])-estradiol, [3H]R2858) and monoclonal antibodies (Abbott ER-EIA Kit) in order to compare the two methods and to evaluate the clinical usefulness of the new immunological, simplified assay. Intra- and interassay reproducibility of the enzyme immunoassay (EIA) method was studied during a 6-month period on 35 standard curves with 4 different batches of monoclonal antibodies. Intraassay coefficients of variation studied on duplicates were smaller than 5% in most cases. Interassay reproducibility of the curves showed coefficients of variation lower than 10% except for standard 0 and 5 fmol/ml. Seven different control specimens provided by Abbott Laboratories were assayed with the EIA method, with interassay coefficients of variation from 1.7% [233.4 +/- 4 (SD) fmol/ml] to 18.2% [18.5 +/- 3.3 fmol/ml]. Pooled cytosols used as control for the dextran coated charcoal method had interassay variation coefficients between 3.8 and 11.4%. Reproducibility has been studied on clinical specimens assayed twice at two different periods with either EIA or dextran coated charcoal methods. Slopes obtained were 1.05 and 0.96, respectively. A good stability of EIA results was obtained with protein concentrations in the range 4-0.15 mg/ml cytosol. No significant effects of dithiothreitol or monothioglycerol (1 mM) on EIA and dextran coated charcoal assay were observed. Eighty breast cancer cytosols were assayed with both EIA and Scatchard analysis. The slope of the regression curve obtained was 1.04 (r = 0.963). Cytosols were assayed by EIA and by a saturating concentration of tritiated ligand (5 nM). With 153 cytosols the EIA/5 nM slope was 1.34 (r = 0.978). This slope can be compared with the slope Scatchard/5 nM obtained with 90 cytosols: 1.29 (r = 0.985). Absence of cross-reactivity of monoclonal ER antibodies with progesterone receptor was observed.