The clinical value of lymph node immunohistochemistry was assessed in 343 consecutive patients with apparently node-negative breast cancer using antimucin monoclonal antibodies BC2, BC3 and 3E1.2. Occult metastases were detected in 41 patients (12 per cent). Although most of these were micrometastatic (< 2 mm in diameter), larger or diffuse deposits were found in ten patients. Blind assessment of repeat haematoxylin and eosin stained sections detected metastases in ten of the 41 patients with occult metastases. After a median follow-up of 79 months, patients with occult metastases had a shorter time to disease recurrence (P < 0.05) but not to death. After adjustment for other prognostic factors, the presence of occult metastases in two or more nodes was the most significant predictor of both disease recurrence (P < 0.01) and death (P < 0.01). When the ten patients with positive haematoxylin and eosin sections were excluded from the analysis, the presence of occult metastases in two or more nodes was still associated with a reduced disease-free interval (P < 0.05) and survival (P < 0.05). Lymph node immunohistochemistry is a simple technique affording more accurate definition of nodal involvement than conventional methods.