Doses for definitive prostate irradiation have been derived empirically using low-energy megavoltage equipment without availability of bone density corrections. With their increased availability, higher energy photons are being used more frequently because of their improved depth of penetration. Although inhomogeneity corrections lead to greater accuracy of dose delivery, the clinical utility of corrections in the pelvis is unclear. This study evaluates the effect of bone density on the dose delivered with respect to the photon energy employed. Contours and volumes for 10 patients were taken from computed tomography scans at the center of the prostate gland. Treatment plans for bilateral prostate arc fields were run on the Capintec Treatment Planning System for 4, 6, 10, and 18 MV photon energies. The monitor units needed to deliver 6500 cGy to isocenter without bone correction were used for calculations, both with and without bone correction using the equivalent path length algorithm. The median dose to the isocenter was 6500 cGy for all energies without bone correction. The median doses using the uncorrected monitor units for the 4 MV, 6 MV, 10 MV, and 18 MV photon beams corrected for bone density were 6033, 6062, 6166, and 6228 cGy, respectively. The variance in target doses observed in our patient sample was +/- 2.3%, +/- 2.2%, +/- 1.7%, and +/- 1.4%, respectively, for the 4 MV, 6 MV, 10 MV, and 18 MV beams with bone correction. The increased density of bone in the pelvis does alter the actual dose to the prostate from external beam treatment.(ABSTRACT TRUNCATED AT 250 WORDS)