DNA ploidy patterns were studied by flow cytometry in nuclear suspensions from 149 paraffin-embedded colorectal adenocarcinomas. The DNA ploidy of rectal tumours was not significantly different from that of colonic tumours. Patients with DNA diploid tumours had a significant survival advantage compared with patients with non-diploid tumours, but DNA ploidy did not confer any significant additional prognostic information when tumour site, Dukes's stage, the invasiveness of the tumour, and the number of lymph node metastases were adjusted for in a proportional hazards regression analysis (Cox). It is concluded that DNA ploidy does not contribute significantly to the explanation of why patients with rectal cancer have a poorer prognosis than those with colonic cancer.