Summary The nuclear DNA content of 351 breast carcinomas was determined by flow cytometry from paraffin-embedded tissue to assess the prognostic significance of DNA ploidy, the DNA index (DI) and the S-phase fraction (SPF). The minimum follow-up of the patients was 22 years, and they were all from a defined urban population. DNA ploidy correlated with histological type and grade, mitotic count and nuclear pleomorphism (P<0.0001), and also with axillary nodal status (P=0.0005), tumour necrosis (P=0.001), primary tumour size (P = 0.03), menopausal status (P = 0.004) and the presence of distant metastases at the time of the diagnosis (P = 0.04). Survival corrected for intercurrent deaths of the patients with a diploid tumour was better than that of the patients with a non-diploid tumour (P = 0.0001, 48% vs 28% at 25 years). SPF Leatham & Brooks, 1987; Slamon et al., 1987).The nuclear DNA content determined by flow cytometry has recently emerged as a new prognostic factor in breast cancer. The technique has gained some popularity, since flow cytometric histograms are rapid to produce and often easy to interpret. It has been suggested that breast carcinomas with an abnormal (non-diploid or aneuploid) nuclear DNA content are associated with less favourable prognosis than carcinomas with a normal (diploid) DNA content (Hedley et al., 1984(Hedley et al., , 1987Coulson et al., 1984;Ewers et al., 1984;Thorud et al., 1986;Baildam et al., 1987;Cornelisse et al., 1987;Kallioniemi et al., 1987;Stal et al., 1989). Similarly, the DNA index (DI, the relative DNA content of an aneuploid stemline of cells as compared with diploid cells) and the percentage of S-phase cells in the DNA histogram (S-phase fraction, SPF) have been found by some to be of prognostic significance (Dowle et al., 1987;Hedley et al., 1987;Kallioniemi et al., 1988;Stal et al., 1989;McDivitt et al., 1986;Klintenberg et al., 1986). However, the prognostic value of nuclear DNA content analysis is still unsettled. Some authors report DNA content to be the most important prognostic variable in breast cancer, and to have independent prognostic value in a multivariate analysis , whereas others find it to have prognostic value only in a univariate analysis (Stal et al., 1989;Hedley et al., 1984), and yet a few fail to find any prognostic significance at all when overall survival is concerned (van der Linden et al., 1989;Uyterlinde et al., 1988;Owainati et al., 1987