The reliability of an immunohistological method, applied to paraffin wax sections, was assessed for determination of oestrogen receptor content of biochemically oestrogen receptor negative breast carcinomata. Sixty consecutive tumours with oestrogen receptor concentrations of less than 10 fmol/mg cytosol protein, as estimated by dextran-coated charcoal biochemical assay, were examined. Paraffin wax sections were treated with DNAse before applying a peroxidase-anti-peroxidase method using ER-ICA monoclonal antibodies. Fifty one cases (85%) were negative, six (10%) weakly positive, and three (5%) were moderately positive. No strongly positive cases were seen. It is suggested that cases with weakly positive staining, especially when localised to a small area, should be regarded as negative. On the other hand, as the three moderately stained cases included two small tubular carcinomas and an invasive ductal carcinoma with high progesterone receptor concentrations, it is more likely that the biochemical assay in these cases represented false negative results due to sampling error or inclusion of fibrous or other non-neoplastic tissue in the assayed samples. It is concluded that the immunohistological method used here is fairly reliable and would be especially valuable for determination of oestrogen receptor content in small, mammographically detected tumours from which no tissue would be available for biochemical assay or frozen section examination.