Axillary nodal status was analysed in 960 consecutive cases of primary invasive breast cancer operated in two Danish hospitals. After stratification according to the number of nodes removed, the rate of node positivity in each subgroup was calculated. We found that the probability of finding at least one metastatic node increased continuously up to about 10 removed nodes. In lymph node negative patients, who did not receive any adjuvant treatment, there was a significant association between the recurrence-free survival and the number of nodes removed, provided that less than 8-10 nodes had been collected. The percentage of node positivity in each subgroup seemed to level off above 10 nodes at about 64%, suggesting that this represents the true rate of node positivity at the time of primary surgery. We conclude that about 10 axillary nodes should be removed in order to minimize the risk of erroneous classification of the axilla.