This paper describes an improved immunohistochemical method for demonstrating oestrogen receptor (OR) protein in paraffin-embedded sections of tissue fixed for 1.5 h in formalin. Thirty-two cases of infiltrating ductal breast carcinoma were stained with a monoclonal anti-OR antibody (H222), using a standard streptavidin-biotin method, following pretreatment with pronase. OR counts in paraffin sections were compared with those of frozen sections and with cytosolic values determined by a dextran-coated charcoal method. Twenty-seven of the carcinomas were OR-positive in paraffin sections. There was concordance between the paraffin section and the frozen section-determined receptor status in 30 cases (94 per cent) and a strong correlation was observed (r = 0.76; P less than 0.0001). Similarly, OR counts in paraffin sections correlated with cytosolic OR values (r = 0.60; P less than 0.001) and there was concordance in 97 per cent of cases. The percentage of positively-stained tumour cells in paraffin sections ranged from 0 to 94 per cent with staining intensities comparable to those seen in frozen sections. Staining of paraffin sections identified more OR-positive tumours than either frozen section staining or cytosolic assay. This study validates immunohistochemical OR analysis in formalin-fixed, paraffin-embedded breast carcinomas using a commercial anti-OR antibody.