The estimated probability of false negative nodal staging was analysed in breast cancer patients who had lower axillary dissection in conjunction with mastectomy and who were initially characterized as node negative. The series comprised 3128 consecutive female patients with invasive mammary carcinoma treated by surgery alone according to a nationwide DBCG protocol, 77-1a (Danish Breast Cancer Co-operative Group). The median number of removed lymph nodes was four, range 0-30. Median followup period was 6.5 years, quartiles 5.5 and 8.0 years. Ipsilateral axillary recurrences appeared in 178 patients (5.7%) during the followup period. The estimated 5-year probability for developing axillary relapse, calculated on the basis of lifetable analyses, was 19% in patients with no lymph nodes removed, 10% with 1-2 removed and negative nodes, 5% with 3-4, 3% with 5-10, and 3% with more than 10 removed and negative lymph nodes. No detectable increase in estimated probability could be traced in patients with five or more nodes removed during followup beyond 5 years, while in patients with less than five nodes removed the probability of axillary recurrence ascended continuously. The study also revealed a significantly decreased survival in groups of patients with less than five nodes removed and negative compared with groups of patients with five or more nodes removed. It is concluded that lower axillary dissection is associated with a certain risk of overlooking lymph nodes metastases, whereby the patients may be deprived of optimal adjuvant therapy. Misjudgment of the qualitative axillary nodal staging is modest, provided that at least five lower lymph nodes are removed.