Radical perineal prostatectomy, relative to retropubic prostatectomy, has become an increasingly used surgical technique for prostate cancer, following advances in laparoscopic methods for pelvic lymph node dissection. Recent protocols of risk stratification may even obviate the need for lymph node dissection. Section of the rectourethralis muscle (RUM) is necessary for access to the retroprostatic space, however, during this procedure rectal injuries may be produced. In this work, we studied the topography and morphology of the RUM, which, despite its importance in perineal surgery, has not been univocally described in the literature. After in situ formalin fixation, the pelvic viscera were removed from 16 male cadavers (age: 54-72 years) and from 4 full-term infants (gestational age: 37-38 weeks). Serial macrosections of the bladder base, prostate gland, and lower rectum cut in horizontal (6 adults and 2 infants) and sagittal (6 adults and 2 infants) planes underwent histological (hematoxylin and eosin, azan-Mallory, and Weigert's staining) and immunohistochemical (anti-smooth muscle actin and anti-sarcomeric actin) study. The remaining 4 adult specimens were cut in horizontal and sagittal planes and plastinated using the epoxy resin E12 sheet procedure. RUM was identified in 10 of 12 (83%) adult specimens and in 4 of 4 (100%) infant specimens. In both sagittal and transverse sections, it showed a triangular-shaped configuration. In all cases, at the level of its posterior portion, fibers continuing with the longitudinal muscular layer of the rectum were visible. In the majority of adult and infant cases, attachment of muscle fibers into the anterior wall of the anal canal was also observed. Anteriorly, the mean (+/-SD) distance between the RUM and the membranous urethra was 5.3 (+/-1.25) mm in adults and 1.0 (+/-0.41) mm in infants. Location of RUM in the prerectal space and the absence of urethral attachment makes the original name of this muscle, "prerectal," by Henle, more correct. In 7 of 10 (70%) adult cases and in 1 of 4 (25%) infant cases, muscle fibers were densely packed along the lateral portions of the RUM, while in its central portion connective tissue was prevalent, with sparse numbers of smooth muscle fibers. Immunohistochemical staining showed that this muscle consists almost entirely of smooth fibers. In all the infant specimens, the RUM was clearly separated from the levator ani, while in 8 of 10 (80%) adult cases, striated fibers of the levator ani and smooth fibers of the RUM intermingled. These structural associations suggest a functional cooperation between the two muscles, particularly in determining the anorectal flexure.