Epidermal growth factor receptor (EGFr) assays were performed by 3 different methods on 246 human primary breast carcinomas. Scatchard analyses of multipoint binding data on 19 of the first 209 tumours, performed by a displacement method, demonstrated that the majority of tumours exhibited 2 classes of binding site, the high-affinity site with an affinity constant (KD) of mean 2 nmol/l (SD 1.3, range 0.44-7 nmol/l), and a low-affinity site, KD mean 9.5 nmol/l (range 6-15.5 nmol/l). Scatchard analysis of multipoint assays using increasing concentrations of 125I-labelled EGF showed that saturation of the high-affinity site occurred in the majority of saturation of the high-affinity site occurred in the majority of tumours at a concentration of labelled EGF of I nM. Comparison of the KD values of the high-affinity site obtained from displacement assays with those obtained by increasing labelled EGF showed that the KD was significantly higher (p = 0.0002) when measured by the latter method. There was no difference in binding capacity of the high-affinity or low-affinity sites by the 2 methods. A 2-point assay with I nM labelled EGF (specific activity approx. 80-130 microCi/microgram) correlated with quantitative values for the high-affinity site from Scatchard analysis (p less than 0.02). There was a strong inverse relationship between EGFr greater than 10 fmol/mg membrane protein (2-point assay) and ER (dextran-coated charcoal method), Chi-squared 2 X 2 contingency table test = 34.027, p less than 0.0001. Follow-up data from 135 patients revealed a strong inverse relationship between EGFr greater than 10 fmol/mg membrane protein and oestrogen receptor (ER) status and a positive correlation with early recurrence and death. Our data describe a reproducible assay for EGFr and show that a cut-off point at 10 fmol/mg allows clinically useful application.