Immunohistochemical staining for oestrogen receptor (ER) has been carried out using antibody ER ID5 on 170 women who received first-line tamoxifen treatment for evaluable metastatic breast cancer. ER status had been determined some years previously, using a ligand-binding cytosol assay. The adequacy of the tissue used for the cytosol assay was always checked by histology on an adjacent block and was deemed to be typical of the tumour overall as was the block used for immunohistochemistry. Six different methods were used to assess the degree of staining and comparisons were made to determine which method gave the most clinically relevant results. Clinical outcome was assessed both in terms of duration of response to tamoxifen determined by log-rank analysis and type of response using the chi-squared test. The ER immunohistochemical assay gave superior results compared with the cytosol assay, with all of the subjective methods of assessment of staining giving statistically significant correlations with clinical outcome. The additional contribution of progesterone receptor (PR) staining with antibody NCL PGR was also studied.