A method has been developed for determining the mean volume of breast cancer in women at the time of the involvement of the first, second, third,... nth axillary lymph nodes. It has been found that the proportion of patients with axillary involvement as well as the number of involved nodes increase progressively with tumour size. This orderly involvement of axillary nodes is observed in all patient subsets despite a wide spread of tumour volume at the time of invasion of the axillary nodes. This makes it possible to compute for each patient or subset of patients the size of the tumour at the time of the first node involvement, a parameter which characterises the propensity for nodal involvement. A strong correlation was demonstrated between the propensity to lymphatic involvement and the probability of distant dissemination. During tumour progression the capacity for lymphatic spread is on average acquired much earlier than the capacity for haematogenous spread. For tumours of the outer quadrants, the volume at first axillary involvement is smaller than for tumours located in the inner quadrants, whereas the tumour volumes at the time of distant metastatic initiation are equal for the two tumour sites. The discrepancy between these two observations shows that axillary involvement, while being a good index of the propensity of the tumour cells to acquire the capacity for distant spread, is not the cause of this spread. From a clinical point of view, these data show that the prognostic significance of axillary involvement can be further increased by taking into account the size of the tumour. The data suggest that there is a continuum from slow growing disease with late axillary involvement and late distant dissemination to the most aggressive subtype.