Sixty-nine males from the University of Manitoba and the University of Winnipeg participated in the study. A preliminary session exposed the subjects to written and pictorial depictions of rape and mutually consenting intercourse. Based on the subjects' penile tumescence when observing these depictions, subjects were classified as force-oriented, nonforce-oriented, or unclassifiable. Subsequently, subjects within each of the three groups were randomly assigned to conditions of sexually violent stimuli (SVS), sexually nonviolent stimuli (SNVS), or control stimuli. Those assigned to the SVS condition were exposed to 10 SVS, including feature-length films and written and pictorial depictions, over 4 weeks. Subjects in the SNVS group were exposed to 10 media presentations that depicted only sexually nonviolent activities. Subjects in the control condition were not exposed to any sexual stimuli during the 4-week period. Soon after completing the exposure phase, subjects were exposed to four depictions similar to those used in the preliminary session. Penile tumescence scores and self-reports of sexual arousal were obtained. Force-oriented subjects exposed to either SVS or SNVS were less aroused to the rape depictions in the postexposure session than those in the control condition. A similar pattern occurred with the nonrape depictions for these subjects. No evidence of a similar 'satiation' pattern was observed for either nonforce-oriented or unclassifiable subjects. Findings are discussed in relation to cognitions, personality differences, conditioning processes, stimulus parameters, and response habituation theories. Social and clinical implications are also discussed. Tabular study data and 56 references are provided.