The performance of a fluorescent cytochemical oestrogen receptor (ER) assay has been compared with an established dextran-coated charcoal ER assay in 85 cases of primary breast cancer. Both assays were performed on tissue samples from the same tumour, by independent observers in different laboratories. A good qualitative correlation between the two methods was found, with agreement in 92 per cent of cases (P less than 0.0005). Criticisms of the fluorescein-labelled oestrogen methods are discussed. It is concluded that such methods should be validated in each laboratory by comparison with a conventional, quantitative ER assay before being used for clinical decisions.