Collagenase dissociation, performed on 40 human breast cancers, yielded between 1 million and 50 million cells from less than 1 g of tissue from each tumor. Approximately 60% of cells (mean) was considered viable as judged by trypan blue exclusion and phase microscopy. On subsequent flow cytometric analysis, 20 cancers (50%) were considered diploid, three were tetraploid, and the remainder, hyperdiploid. Thymidine labeling (TLI) and flow cytometry following mechanical dissociation also were performed on 23 of these 40 tumors. Among this group of 23 cases, the median percentage of S-phase cells obtained by collagenase dissociation was 5.4, by TLI was 5.7, and by mechanical dissociation was 9.7. There was excellent correlation between the percentage of S-phase cells obtained by collagenase and TLI (r = 0.847, p = 0.0001) but only fair correlation between the percentage of S-phase cells obtained by mechanical dissociation and TLI (r = 0.597, p = 0.0027). The percentage of S-phase cells obtained by either collagenase or mechanical dissociation predicted whether a tumor was above or below median TLI in 19 of 23 cases (p = 0.0018). Estrogen receptor positivity or negativity did not predict whether a tumor was above or below median TLI (r = 0.283, p = 0.130) or above or below median S-phase fraction following collagenase dissociation (r = 0.218, p = 0.182), nor did quantitative estrogen receptor correlate significantly with TLI (r = 0.283, p = 0.13) or S-phase fraction (r = 0.218, p = 0.18).