AIMS--(i) To assess the validity of an immunocytochemical technique for detecting pS2 protein in paraffin wax embedded tissue; (ii) to provide further data on the relation between pS2 protein and oestrogen receptor (ER) and progesterone receptor (PgR). METHODS--Breast cancer excision biopsy specimens were obtained from 35 previously untreated patients. An immunoradiometric assay was compared with an immunohistochemical method for measuring pS2 protein. ER and PgR were measured in cytosol fractions by enzyme immunoassay and the relation between the presence of these receptors and pS2 protein was assessed before and after subdivision of the women into groups of over or under 50 years of age. RESULTS--A good correlation was seen between the immunoradiometric and immunohistochemical methods for pS2 protein measurement (r = 0.84; p = 0.0001). Log-transformed data showed a significant correlation between increasing values of ER and pS2 protein (r = 0.45; p = 0.006) and to a lesser extent between pS2 protein and PgR (r = 0.38; p = 0.03). Correlations were also shown between pS2+ and PgR+ status (p = 0.01), and between ER and PgR positivity (p = 0.05; Fisher's exact test). pS2+ protein status was only associated with ER+ status in patients aged 50 years or less. CONCLUSIONS--The two methods for pS2 analysis are virtually interchangeable. This provides strong support for using immunohistochemistry for pS2 in paraffin wax embedded tissue. The association with ER positivity and pS2+ protein status only in the premenopausal patients may be due to the higher levels of oestrogenic stimuli in that group.