This paper is the report of a dosimetric study of 41 rectal and rectosigmoid complications after radiotherapeutic treatment (1974-1978) of 287 cervical uterine tumors. Treatment consisted of external irradiation (25 MeV linear accelerator) and intracavitary irradiation (Fletcher-Suit applicator) at different doses depending on tumor stage. Dosimetric measurements were expressed as the maximum rectal dose and mean rectal dose on the anterior surface of the rectum, as proposed by the Groupe Européen de Curiethérapie. Rectal doses were also studied as a function of intracavitary irradiation and intracavitary + external irradiation (maximum rectal and mean cumulative doses for each). The results show a significant difference in the state of the patients with and without complications, based on the dose reaching the rectum. The maximum and the mean cumulative rectal doses serve as one of the primary indicators for predicting complications. These values should therefore be determined before placement of intracavitary sources or, at the latest, before the second intracavitary application. We have shown that there is no fixed threshold dose, but that it varies from one region to another, depending on level of external irradiation. Our results argue in favor of adapting individual patient therapy based on simple precautions, which are adjustable to all treatment modalities. This method could lead to complete elimination of late rectal and rectosigmoid complications arising from radiotherapeutic treatment of cervical uterine cancer.